framework,version,device,op_name,kernel_source,gemm_dtype,m,n,k,latency
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,1024,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,5120,0.0185244447655148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,6144,0.009775111244784461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,4096,0.017799999978807237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,1536,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,64,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,2560,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,5120,0.08883644474877252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,7168,0.05837066968282064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,10240,0.17656177944607207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,7168,0.011773332953453064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,2560,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,7168,0.014293332894643148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,128,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,5120,0.15373333295186362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,12288,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,5120,1.6371101803249781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,10240,0.0124195557501581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,2560,0.024215110474162634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,65536,0.3284240033891466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,512,0.002593777866827117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,2048,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,64,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,51200,0.09691555632485284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,1536,0.0038977778620190094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,3072,0.17439556121826172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,3072,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,3072,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,65536,0.20092888673146567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,64,0.0063528890411059065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,51200,0.34208356009589297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,8192,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,65536,0.0770933363172743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,8192,0.048750221729278564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,3584,0.027120888233184814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,768,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,128,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,7168,0.0747493306795756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,51200,0.05124533176422119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,3072,0.019952888290087383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,64,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,1024,0.024430221981472437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,2560,0.13554044564565024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,16384,0.02950577934583028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,12288,0.6094462076822916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,4096,0.05458933446142408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,51200,0.31493867768181694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,3072,0.008411555654472774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,1536,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,2048,0.011511111425028907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,3072,0.0484808882077535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,32,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,256,0.010046222143703038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,4096,0.011957333319716983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,51200,0.08672800328996445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,2560,0.03668177790111966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,2560,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,512,0.04185600082079569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,16384,2.4796693589952254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,128,0.0358586675590939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,128,0.019318222999572754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,7168,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,768,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,4096,0.11281511518690322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,7168,0.009891555541091496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,2048,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,65536,2.6165608300103083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,7168,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,3072,0.021195555726687115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,8192,0.07233688566419813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,8192,0.5504213439093696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,7168,0.06264888577991061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,2048,0.010209778116808997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,10240,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,16384,0.02347555591000451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,7168,0.015039111177126566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,1024,0.011703111231327057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,64,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,256,0.009400000174840292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,6144,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,2560,0.03756711218092177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,3584,0.08309955729378594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,1024,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,128,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,2048,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,1024,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,1536,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,12288,0.19695288605160186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,1536,0.05220266514354282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,3072,0.021640888518757288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,2560,0.026187555657492742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,12288,0.034272889296213783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,3072,0.034987555609809026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,6144,0.018107554978794523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,10240,0.02803022331661648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,4096,0.016326222154829238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,51200,1.101928922865126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,12288,0.027727110518349543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,12288,0.10646222035090129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,7168,0.03815111186769273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,65536,0.1031751102871365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,12288,0.08112800121307373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,8192,0.03152622116936578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,6144,0.03050311075316535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,65536,1.8917484283447266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,4096,0.04098844528198242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,7168,0.027588443623648748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,256,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,3072,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,3072,0.04480977853139242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,8192,0.019515555765893724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,1024,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,768,0.03353066576851739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,4096,0.016214221715927124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,3584,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,64,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,64,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,4096,0.047332445780436196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,768,0.03068888849682278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,256,0.009820444716347588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,1024,0.0284204449918535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,3072,0.06284444199668036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,12288,1.6394400066799586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,128,0.017500445246696472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,4096,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,8192,0.06390488810009427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,8192,0.233857790629069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,5120,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,32,0.0018888889915413326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,4096,0.03965866565704346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,2560,0.04351466563012865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,768,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,1536,0.03358755509058634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,3072,0.015784889459609985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,7168,0.045997334851158984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,768,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,1024,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,10240,0.04529244369930691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,10240,0.0884702205657959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,2048,0.008173333273993598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,128,0.0034426665968365143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,12288,0.05906755394405789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,10240,0.022914666268560622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,2560,0.04931111137072245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,3584,0.06346488661236234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,3584,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,3584,0.021927111678653296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,3584,0.009741333623727163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,3072,0.008058666355080074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,51200,0.35005511177910703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,65536,0.06342577934265137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,3072,0.01685066686736213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,2048,0.0069004446268081665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,12288,0.013740444348918067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,51200,0.12824889024098715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,12288,0.0862542192141215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,3584,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,128,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,256,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,4096,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,1024,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,10240,0.08582844336827596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,512,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,65536,5.009850819905599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,512,0.013807999591032663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,3584,0.08298222223917644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,5120,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,12288,0.1932542191611396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,65536,0.09601600302590264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,3584,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,1024,0.025995555851194594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,51200,0.7074844572279188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,2048,0.02230577833122677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,51200,0.05301600032382541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,2560,0.00942844483587477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,16384,0.2965928978390164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,768,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,16384,0.022431999444961548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,1536,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,1024,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,5120,0.1455306609471639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,1536,0.02860800094074673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,1024,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,2048,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,7168,0.022095110681321885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,768,0.9915333853827583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,768,0.01089599976936976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,6144,0.20516444577111137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,256,0.012359999948077731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,768,0.17617421680026582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,51200,0.1200595564312405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,768,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,512,0.015491555134455362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,1024,0.017697778013017442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,512,0.07428710990481906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,2048,0.04573600159751045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,65536,0.29859556092156303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,10240,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,1024,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,2560,0.007164444360468123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,3072,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,65536,0.1887084510591295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,32,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,256,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,512,0.008672888908121321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,32,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,16384,0.5851413408915201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,12288,0.026004445221689012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,5120,1.6237707138061523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,6144,0.03223199976815118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,4096,0.12284621927473281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,256,0.08455644713507758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,8192,0.07655022541681926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,768,0.025472000241279602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,256,0.002201777779393726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,16384,0.41840089692009824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,10240,0.6353608767191569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,3072,0.3369137710995144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,1024,0.006684444430801604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,6144,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,6144,0.05110133356518216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,10240,0.057922667927212186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,2048,0.0399271117316352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,4096,0.003913777983850903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,3072,0.340864896774292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,768,0.015280000037617154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,12288,0.8373191091749402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,512,0.009726222190592024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,5120,0.04009600149260627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,6144,0.0047866664826869965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,768,0.0028880000528362063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,2048,0.02440711193614536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,12288,0.5852257940504286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,16384,0.012803555362754397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,768,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,16384,1.1757129033406575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,768,0.027170666389995154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,2560,0.12241421805487739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,3584,0.015127999915017022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,7168,0.06697244114345975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,7168,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,65536,1.2234275605943468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,2560,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,65536,0.07546666595670912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,3072,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,3584,0.014441778262456259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,64,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,5120,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,1024,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,3072,0.09170311027103001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,5120,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,51200,0.10014222065607707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,65536,0.08351555797788833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,1536,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,1024,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,512,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,7168,0.021944888763957556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,8192,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,6144,0.05140977766778734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,6144,0.05077866713205973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,16384,5.181349436442058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,7168,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,4096,0.010882666541470421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,128,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,512,0.008813333180215623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,256,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,1024,0.007101332975758447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,2048,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,32,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,3072,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,10240,8.55331336127387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,7168,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,7168,0.016869333055284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,7168,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,256,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,3072,0.020676443974177044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,1024,0.03341066506173875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,32,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,51200,0.09738755226135254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,2048,0.09550578064388698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,10240,0.03242666522661845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,4096,0.02555911077393426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,4096,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,2048,0.024749333659807842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,512,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,1024,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,6144,0.055177778005599976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,7168,0.02193333374129401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,1024,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,256,0.01752355529202355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,7168,0.4844951099819607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,64,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,3584,0.03758222195837233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,32,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,64,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,64,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,16384,0.044171555174721606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,8192,0.04565688967704773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,1536,0.08234755860434638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,12288,0.03388977713055081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,1024,0.021665778425004747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,3584,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,3584,0.05539200041029188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,16384,0.18993066416846383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,1536,0.003569777641031477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,16384,0.2345520125495063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,3584,0.0122097780307134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,128,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,10240,0.5660044352213541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,3584,0.05234666665395101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,1024,0.009752000371615091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,5120,0.02598222262329525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,10240,0.1240711079703437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,51200,1.4723368750678167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,10240,1.8167794545491536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,16384,0.0322728885544671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,1536,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,65536,0.7809226777818469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,1536,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,10240,0.8664346800910102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,512,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,65536,0.06640533606211345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,3584,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,5120,0.10567022032207912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,256,0.2890391084882948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,10240,0.22344711091783312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,2560,0.3284657796223958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,65536,0.07698844538794623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,256,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,3584,0.015302222636010913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,12288,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,128,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,51200,0.046557333734300405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,256,0.020976000361972388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,8192,0.4438897768656413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,128,0.015909334023793537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,1024,0.020222221811612446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,8192,0.1020124422179328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,12288,1.0924782223171658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,12288,0.4464106559753418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,7168,0.1293404499689738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,2048,0.04208177659246656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,8192,0.6778542200724283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,51200,0.5512622197469076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,65536,0.12000266710917155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,768,0.01962844365172916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,10240,0.1981511116027832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,128,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,1536,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,7168,0.030415111117892798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,8192,0.03858933183881972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,3072,0.033944000800450645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,12288,0.08451200193829006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,51200,0.016542222764756944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,10240,0.08803200059466892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,128,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,4096,0.011538666983445486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,7168,0.21961332692040336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,1536,0.11166221565670437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,2048,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,8192,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,5120,0.09656177626715766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,7168,0.018965333700180054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,4096,0.016590222716331482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,128,0.06982666916317411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,65536,0.07245333327187432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,16384,0.10394577847586738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,7168,0.013656888571050433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,6144,0.048231111632453076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,768,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,10240,0.07228533426920573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,3584,0.20028089152442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,65536,0.08986044592327541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,51200,0.14686044057210287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,12288,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,65536,0.1702080037858751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,1024,0.03722488880157471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,8192,0.04657422171698677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,6144,0.017261332935757108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,7168,0.03286755416128371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,2560,0.03683111071586609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,768,0.002599999929467837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,3072,0.03252799974547492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,10240,0.030180444320042927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,7168,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,65536,0.7876142395867242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,768,0.06379822227689955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,768,0.008649778034951951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,12288,2.30477163526747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,256,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,6144,0.024010666542583044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,4096,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,128,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,1024,0.021828444467650518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,7168,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,3072,0.059749331739213735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,256,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,10240,0.21858667002783883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,4096,0.02052533295419481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,6144,0.03447377681732178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,7168,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,3072,0.025754666990704004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,1024,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,10240,0.012461333639091916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,51200,0.02642755541536543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,5120,0.020807999703619216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,64,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,7168,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,64,1.1678684022691515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,32,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,1024,0.05150844322310554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,4096,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,1024,0.0063982220987478895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,1536,0.00360622215602133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,4096,0.014712888333532544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,8192,0.04559555649757385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,128,0.3235982259114583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,6144,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,8192,0.013905778527259827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,1024,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,10240,0.015829333000712924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,12288,0.03172889020707872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,768,0.10718933078977798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,2048,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,3072,0.17789244651794434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,16384,0.058021333482530385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,51200,0.12215110990736221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,3584,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,256,0.013827555709415011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,2048,0.01348622226052814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,768,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,12288,0.15305777390797934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,3072,0.0792542232407464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,512,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,6144,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,1536,0.021661332911915247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,65536,1.742846171061198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,3072,0.1993537743886312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,16384,0.7282506624857584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,3072,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,2560,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,768,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,3072,0.01349066694577535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,5120,0.008714666797055138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,768,0.04725955592261421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,16384,0.3168293370140923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,8192,0.023327999644809302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,2560,0.1321964396370782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,51200,0.04940533306863573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,768,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,65536,0.06976711087756686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,2048,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,10240,0.049457778533299766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,3072,0.01625422139962514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,5120,0.015678221980730694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,65536,0.0670915577146742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,65536,0.10819911294513279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,1536,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,1536,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,128,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,128,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,51200,0.11645689275529648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,51200,0.19828266567654082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,1024,0.018941332896550495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,5120,0.0764568911658393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,10240,0.034627556800842285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,6144,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,4096,0.007970666719807519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,64,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,128,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,5120,0.1477715571721395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,5120,0.009054222040706212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,128,0.1738791068394979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,2560,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,12288,0.09102755784988403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,2048,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,256,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,7168,0.037667555941475764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,16384,0.029535111453798082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,128,0.014910222755538093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,2048,0.020115555988417733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,12288,0.06847822003894381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,3072,0.00906044410334693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,8192,0.5028497907850478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,12288,0.062005334430270724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,51200,0.1803884506225586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,64,0.0028044444819291434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,7168,0.010649777948856354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,768,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,51200,1.2702453401353624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,8192,0.022794667217466567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,2560,0.02515999972820282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,51200,0.8573670917087131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,65536,0.06481244166692098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,1024,0.026363554928037856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,16384,0.2632399929894341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,8192,0.050607111718919545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,5120,0.03488444288571676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,16384,0.1125297811296251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,51200,0.14399911297692194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,256,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,3072,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,512,0.00720888872941335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,12288,0.02106399999724494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,8192,0.011010666688283285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,256,0.006533333410819371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,2048,0.01824977828396691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,3072,0.09218044413460626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,512,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,12288,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,51200,4.582279205322266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,10240,0.0748222205373976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,5120,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,3584,0.010658666491508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,128,0.015263110399246216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,512,0.012237333589129977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,8192,0.02274755636850993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,16384,0.15678844186994764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,65536,0.8305902481079102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,6144,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,2048,0.008099555969238281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,512,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,51200,0.3703600035773383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,256,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,65536,0.9522684945000542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,1536,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,768,0.034985777404573225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,12288,0.027079110344250996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,5120,0.04162933429082235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,51200,0.05237333310974968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,16384,0.3216008875105116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,1536,0.0035866668654812705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,12288,0.07511022355821398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,512,0.01648977730009291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,64,0.00851288851764467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,7168,0.02237777743074629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,2048,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,16384,0.14250044027964273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,1536,0.0503768887784746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,16384,0.687197314368354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,5120,0.0383200016286638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,8192,0.037187556425730385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,1536,0.043790221214294434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,1024,0.06196888950135973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,3072,0.12532711029052734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,32,0.010518222219414182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,7168,0.010328888893127441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,65536,0.10820888810687596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,128,0.011471110913488599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,16384,0.02718577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,5120,0.017660443981488545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,32,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,12288,0.1704853375752767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,6144,0.030625777112113103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,51200,0.21598488754696318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,16384,0.19874578052096895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,256,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,16384,0.05190222130881416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,65536,0.5320071114434136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,16384,0.021309332715140447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,64,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,7168,0.010066666536860997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,2560,0.010241777532630498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,4096,0.028573334217071533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,3072,0.008070222205585903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,2560,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,4096,0.1061484416325887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,4096,0.008409777449236976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,8192,0.1052924460834927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,1536,0.009540444446934594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,256,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,6144,0.01868799991077847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,12288,0.6339279810587565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,768,0.01963200006220076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,1024,0.011757333245542316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,64,0.018989332848125033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,1536,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,3072,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,3584,0.010642666783597736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,4096,0.023738667368888855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,51200,0.37619911299811465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,12288,0.03819110989570618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,51200,0.5434880256652832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,128,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,10240,0.09351378017001682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,10240,0.016102222932709586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,32,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,12288,0.01702577703528934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,65536,0.9849546220567492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,2048,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,51200,0.06670488913853963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,128,0.01807733376820882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,51200,1.5365208519829645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,2048,0.09538488917880589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,512,0.0099653336736891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,2560,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,8192,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,2560,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,10240,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,7168,0.011439999772442712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,5120,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,3072,0.16734488805135092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,3072,0.02367288867632548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,65536,0.05676266882154676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,7168,0.039022223816977605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,128,0.006360888895061281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,128,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,4096,0.09049510955810547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,4096,0.025379555092917547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,51200,0.21719555060068765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,51200,0.35184889369540745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,7168,0.040708445840411715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,16384,0.07922577857971191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,8192,0.02110311057832506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,2560,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,256,0.0041759999261962045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,2560,0.4406729009416368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,5120,0.04605866803063286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,16384,0.016593777471118502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,32,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,1024,0.04498222139146593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,8192,0.1484044392903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,2048,0.007810666329330868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,1536,0.045485334263907544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,2048,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,51200,0.23847821023729113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,2560,0.04708622230423821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,4096,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,2048,0.00980711148844825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,2560,0.06880355543560453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,1024,0.029005332125557795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,12288,0.08908622132407294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,64,0.17028000619676378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,7168,0.011241777903503843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,16384,0.0491502218776279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,7168,0.021731555461883545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,4096,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,32,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,1024,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,512,0.006020444548792309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,512,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,2560,0.005807999935415056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,16384,0.12493599785698785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,51200,9.151159498426649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,65536,0.06689244508743286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,8192,0.016162667009565566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,5120,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,6144,0.02458577851454417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,10240,0.18156088723076713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,4096,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,1536,0.015092444088723926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,768,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,7168,0.013044444223244986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,6144,0.083423998620775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,4096,0.023532445232073467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,12288,0.053824888335333936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,2560,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,4096,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,32,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,16384,0.2561039924621582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,1024,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,5120,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,128,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,256,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,6144,0.1211964421802097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,768,0.11534844504462348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,512,0.05564888980653551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,3072,0.06039200226465861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,768,0.04952889018588596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,65536,0.0768702228864034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,5120,1.1434026294284396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,512,0.02167822255028619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,12288,0.028078221612506445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,768,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,51200,0.07107555866241455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,16384,0.02162933349609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,5120,0.14767644140455458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,1536,0.03930400146378411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,768,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,256,0.02351288828584883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,10240,0.08444977468914455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,64,0.003559999995761447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,51200,0.40481776661343044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,512,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,2560,0.024853333830833435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,128,0.014854222536087036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,256,0.0031351111829280853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,4096,0.021669333179791767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,3072,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,12288,0.013423999978436364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,32,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,768,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,10240,0.04349777764744229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,2048,0.016467556357383728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,6144,0.013917333549923368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,51200,0.09383200274573432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,1536,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,768,0.14322488837771946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,8192,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,8192,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,2560,0.04698666599061754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,3072,0.022477333744366963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,65536,4.121330685085721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,7168,1.2521475685967338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,256,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,2048,0.026389333936903212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,2560,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,65536,0.7329022089640299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,64,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,1024,0.018426666657129925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,4096,0.01741422216097514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,512,0.024903999434577093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,6144,0.017883555756674874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,3072,0.024409777588314478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,128,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,1024,0.02180888917711046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,8192,0.05750222338570488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,256,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,128,0.014326221413082547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,16384,0.04144088758362664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,1024,0.0199644449684355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,1024,0.045217778947618276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,3072,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,256,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,65536,0.05801333321465386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,1024,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,128,0.03626755542225308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,51200,1.1375172932942708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,51200,31.47066582573785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,7168,0.5487608909606934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,256,0.027665777338875666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,128,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,512,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,3584,0.004580444345871608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,65536,0.14680088890923396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,128,0.015115555789735583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,12288,0.12468443976508246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,4096,0.18497067027621797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,768,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,3072,0.09777422083748712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,51200,0.3584551016489665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,5120,0.6250649028354222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,64,0.18278933895958793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,7168,0.05139644609557258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,65536,1.5162649154663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,1024,0.03594222333696153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,2048,0.05045244428846571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,2560,0.1321528885099623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,512,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,1536,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,1024,0.01591022147072686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,1536,0.7908248901367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,4096,0.013030221892727746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,7168,0.027942223681343928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,768,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,1536,0.00489688871635331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,64,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,512,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,32,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,768,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,5120,0.06084444125493368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,51200,0.07626844114727445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,4096,0.029676444000667993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,2048,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,5120,0.07959555255042182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,7168,0.02546666728125678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,512,0.044473777214686074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,128,0.035113778379228376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,256,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,6144,0.42952622307671445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,12288,0.2500435511271159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,16384,0.06761244270536634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,32,0.021595555875036452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,8192,0.672452449798584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,1024,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,16384,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,256,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,2048,0.04244977898067898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,5120,0.029264890485339697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,7168,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,7168,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,3072,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,8192,0.033430222007963396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,512,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,2560,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,8192,0.13858400450812444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,64,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,128,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,3072,0.03172533379660712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,12288,0.018552000323931377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,256,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,8192,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,1024,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,2048,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,2048,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,51200,0.08305955595440336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,256,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,3072,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,51200,0.04799911048677233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,4096,0.11803377999199761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,2560,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,512,0.19609688387976754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,4096,0.07645955350663927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,6144,0.014030221435758801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,2048,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,51200,0.6913386450873481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,512,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,51200,0.7289990848965116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,5120,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,1024,0.020605333977275424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,64,0.006028444402747684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,2048,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,1024,0.03979644510481093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,51200,0.12114844057295059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,4096,0.03811288873354594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,512,0.00906577789121204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,16384,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,8192,0.07731911208894518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,1024,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,4096,0.014927999840842353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,4096,0.010168000227875179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,10240,0.17984622054629854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,768,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,16384,0.25569155481126576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,12288,0.17570932706197104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,512,0.0063564446237352155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,6144,0.09667111105389065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,32,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,2048,0.047931555244657725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,6144,0.019259555472267997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,16384,0.05339644352595011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,128,0.008445333275530074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,128,0.0029502221279674103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,4096,0.011180444724029966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,3072,0.042225778102874756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,8192,0.15153243806627062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,51200,11.697132534450956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,4096,0.013904000322024027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,64,0.01998222205373976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,128,0.004582222137186262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,4096,0.05083910955323113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,10240,0.028374221589830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,16384,0.023611555496851604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,2048,0.02569244470861223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,64,0.0017946666727463405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,6144,0.5731217596266005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,4096,0.04786933461825053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,32,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,128,0.014920888675583733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,512,0.022836445106400385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,1536,0.013602666556835175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,2560,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,8192,0.056799113750457764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,51200,0.5081982082790798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,128,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,8192,0.04018844498528375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,10240,0.025411556164423626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,512,0.008388444781303406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,32,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,7168,0.04968177941110399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,1024,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,65536,0.7100897894965278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,512,0.057612445619371205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,768,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,2560,0.013657777673668332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,3584,0.013468444347381592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,128,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,3072,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,3584,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,2048,0.20446756150987414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,7168,0.022056000100241765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,65536,1.258628421359592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,2560,0.02569777766863505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,512,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,7168,0.022277333670192297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,3584,0.017859554953045316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,512,0.0045422220395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,51200,0.09821866618262397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,51200,0.04294844468434652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,7168,0.1309831142425537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,128,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,2048,0.007088888850477006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,7168,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,51200,0.07887022362815009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,12288,0.05413155423270332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,8192,0.09611911243862575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,1024,0.014173333843549093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,4096,0.012136889000733694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,16384,0.03979111048910353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,7168,0.030014223522610132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,2048,0.0292577776643965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,256,0.015999999311235216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,51200,0.25894755787319607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,768,0.00852177788813909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,5120,0.011232888533009423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,1536,0.010290666586822933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,2560,0.04696711235576206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,12288,0.03973333372010125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,12288,0.03685422076119317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,768,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,10240,0.03660888804329766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,6144,0.04265244470702278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,65536,0.05567022164662679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,8192,0.011033777561452655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,6144,0.03395377927356296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,2048,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,256,0.008608889248636033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,1536,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,7168,0.013832000394662222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,6144,0.135169784228007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,5120,0.055982225471072726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,3072,0.23424355189005533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,51200,0.22251288096110025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,512,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,4096,0.06686400042639838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,7168,0.011137777732478248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,3072,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,1536,0.005731555736727185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,8192,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,128,0.0907084412044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,3072,0.029680887858072918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,1536,0.22752266459994844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,16384,0.29034754965040416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,5120,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,8192,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,3072,0.022462222311231825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,65536,0.37684088283114964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,12288,0.049545778168572314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,10240,0.02625688910484314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,128,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,768,0.017471111483044095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,3072,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,128,0.012479999827014076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,4096,0.025434666209750708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,10240,0.27620087729560006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,6144,0.03183555603027344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,65536,0.19732444816165498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,1536,0.029436445898479883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,10240,0.332023991478814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,512,0.00871288859181934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,2048,0.025404444999165002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,2560,0.054413331879509814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,10240,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,64,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,12288,0.022026666336589392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,12288,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,51200,0.026134222745895386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,2048,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,12288,0.04170044594340854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,768,0.0029893333299292457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,8192,0.07490489217970106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,5120,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,51200,0.051339556773503624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,16384,0.019665777683258057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,16384,0.2846711211734348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,8192,0.031168000565634832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,12288,0.0760293338033888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,512,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,128,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,64,0.006034666465388403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,6144,0.05463822351561653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,2560,0.46885331471761066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,51200,4.79301495022244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,5120,0.023728888895776536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,1536,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,16384,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,16384,0.04515733321507772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,5120,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,16384,0.016505777835845947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,12288,0.20638488398657906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,2560,0.007700444095664554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,16384,0.010696888797812991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,1024,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,256,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,64,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,5120,0.032608889871173434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,768,0.020424000091022916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,7168,0.009974222216341231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,6144,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,1024,0.032283554474512734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,4096,0.014231999715169271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,4096,0.04077600108252631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,16384,0.049025777313444346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,512,0.00563466673096021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,3584,0.01424088908566369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,16384,0.3403555552164714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,65536,0.08329688840442233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,7168,0.06452355782190959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,512,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,512,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,128,0.016551110479566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,2048,0.006795555353164673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,3072,0.05464533302519056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,1536,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,3072,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,1024,0.24040177133348253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,6144,0.018977777825461496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,768,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,2560,0.03905511233541701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,10240,0.5474426481458876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,1536,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,10240,0.08389066987567478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,5120,0.012315555579132505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,3584,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,256,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,6144,0.056114666991763644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,4096,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,32,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,10240,0.032990223831600614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,7168,0.02825688984658983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,64,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,51200,0.05315022336112129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,16384,0.5310551325480143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,4096,0.00812177773978975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,4096,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,8192,0.04789333211051094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,2048,0.007510222494602203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,2048,0.008084444536103142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,10240,0.4339040120442708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,32,0.0025759999536805684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,8192,0.03437511126200358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,256,0.007810666329330868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,5120,0.04227022329966227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,512,0.029993779129452173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,512,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,2560,0.04870311088032193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,5120,0.043637331989076406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,12288,0.034824000464545354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,12288,0.7299448649088541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,65536,4.140668233235677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,256,0.026540444956885442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,51200,0.07591111130184598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,1536,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,256,0.006177777631415262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,64,0.9303351508246528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,7168,1.7499556011623805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,5120,0.02968977888425191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,1024,0.24243378639221191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,16384,0.0915635559293959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,256,0.0028844444702068963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,64,0.05787644121381971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,128,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,256,0.01370044466522005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,8192,0.2902159955766466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,1024,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,1024,0.00369777778784434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,2048,0.014987554815080432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,5120,0.08250755733913846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,16384,0.038922667503356934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,12288,0.009913778139485253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,6144,0.3679493268330892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,7168,0.2609688970777723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,1024,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,2560,0.025439111722840205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,128,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,10240,0.10815911160575019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,51200,0.04355022311210632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,10240,0.0891031093067593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,8192,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,2560,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,512,0.11409866809844971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,32,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,768,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,256,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,64,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,51200,0.599471992916531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,8192,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,65536,0.064301331837972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,512,0.020735111501481798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,8192,0.04144088758362664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,128,0.012290666500727335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,6144,0.01166311071978675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,4096,0.03151199883884854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,51200,0.012440889245933957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,7168,0.587073802947998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,1024,0.040307554933759905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,1536,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,16384,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,65536,0.07370044125450982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,12288,0.02015911042690277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,8192,0.02995022137959798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,4096,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,1536,0.006760889043410619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,4096,0.12059022320641412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,768,0.022184888521830242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,65536,2.464228524102105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,128,0.01959733333852556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,3072,0.016714667280515034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,5120,0.09127200312084621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,2560,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,3584,0.02158755560715993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,16384,0.09942489200168186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,1536,0.043177776866488986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,5120,0.010427555276287926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,6144,0.01110666659143236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,10240,0.07963555388980441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,4096,0.14238222440083823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,256,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,65536,0.18350666099124482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,1536,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,3072,0.017092444830470614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,10240,0.02130311065249973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,6144,0.07247466511196561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,7168,0.01292622254954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,5120,0.10225422514809503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,6144,0.34371288617451984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,32,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,8192,0.023018666439586218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,51200,0.05526222123040093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,2048,0.006798222247097228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,16384,0.27043644587198895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,1024,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,3584,0.20895822842915854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,12288,0.24322578642103407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,65536,0.12184800042046441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,16384,1.8975306616889105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,6144,0.012055111428101858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,10240,0.04099555479155646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,5120,0.012464889221721225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,2560,0.030589332183202107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,10240,0.1532613303926256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,768,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,6144,0.021176889538764954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,6144,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,7168,0.01053422192732493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,10240,0.014192000031471252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,256,0.0037848887344201407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,256,0.01717866626050737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,1536,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,6144,0.02966133422321743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,64,0.0487937761677636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,10240,0.015824000040690105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,6144,0.009381333159075843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,3072,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,5120,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,2048,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,64,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,65536,0.07750133011076185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,4096,0.008023999631404877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,3584,0.03550666570663452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,6144,0.025248888466093276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,5120,0.8313475714789497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,32,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,6144,0.05528266562355889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,16384,0.13085156016879612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,3072,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,1024,0.050112889872656934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,768,0.014190221826235453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,32,0.0053635554181204895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,6144,0.058895111083984375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,3584,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,10240,0.05819822020000882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,16384,0.7801671028137207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,2560,0.038675556580225624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,8192,0.020260444945759244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,8192,0.03202844328350491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,768,0.01677955521477593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,51200,0.3205893304612902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,1024,0.04401333464516533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,1024,0.009582222335868413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,3584,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,65536,176.3758002387153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,10240,0.04244977898067898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,16384,0.2814799944559733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,12288,0.014582221706708273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,5120,0.5132115681966146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,1536,0.015545777148670621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,6144,0.034169776572121516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,64,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,3584,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,256,0.13668355676862928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,2048,0.007135999699433644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,2048,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,128,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,768,0.031778666708204485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,65536,0.6769030888875326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,8192,0.03155288762516446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,3072,0.0070062221752272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,3584,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,3584,0.027024888330035742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,3584,0.013555555707878537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,3072,0.03319999906751845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,128,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,768,0.01998844411638048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,5120,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,8192,0.0392231113380856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,64,0.012207110722859701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,2560,0.03792799843682183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,7168,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,6144,0.019811555743217468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,6144,0.14743378427293566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,1024,0.04072177741262648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,3072,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,10240,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,768,0.016894222961531747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,1536,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,768,0.016541333662139047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,12288,0.12482933203379314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,1024,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,51200,0.060554669962988965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,2048,0.017151110702090792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,65536,0.0641288889778985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,65536,0.468140443166097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,51200,0.0804684427049425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,4096,0.013019555144839816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,512,0.00601333338353369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,6144,0.07418755690256755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,12288,0.11191199885474311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,8192,0.12399910555945502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,768,0.03315111001332601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,1536,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,1536,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,256,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,7168,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,10240,0.012402666939629449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,128,0.016567111015319824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,65536,0.049386666880713574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,4096,0.027127999398443434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,2048,0.003921777837806278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,16384,0.02251822253068288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,16384,0.2388355467054579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,3072,0.02052177819940779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,768,0.00666933341158761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,16384,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,256,0.01018311083316803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,2560,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,12288,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,1536,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,32,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,10240,0.01313244468635983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,65536,0.07725066608852811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,51200,0.0580915543768141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,2560,0.03217511044608222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,51200,0.05063644382688734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,256,0.012230222423871359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,4096,0.014537778165605334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,5120,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,7168,0.1416391134262085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,2560,0.011113777756690979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,32,0.0015084444441729123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,10240,0.0566613343026903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,10240,4.308999803331163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,2048,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,2048,0.028079999817742243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,256,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,6144,0.04077422287729051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,64,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,256,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,768,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,512,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,256,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,65536,1.1860862308078342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,256,0.002307555534773403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,256,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,5120,0.02472355630662706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,2560,0.03587377733654446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,12288,0.02126311096880171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,65536,0.2511235607994927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,5120,0.024725332856178284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,10240,0.04298399885495504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,256,0.002206222257680363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,1536,0.012734221915404001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,768,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,2560,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,51200,0.05865244732962715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,6144,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,2560,0.009389333426952362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,6144,0.043996443351109825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,768,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,51200,0.05843377775616116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,128,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,12288,0.030396444929970637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,6144,0.12754310501946345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,12288,0.4236737887064616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,3072,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,2560,0.07861422167883979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,1024,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,2048,0.013095999757448832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,256,0.03161511156294081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,10240,0.1314693291982015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,51200,0.4426604376898871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,512,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,8192,0.12033243974049886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,12288,0.29685600598653156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,512,0.014292443792025248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,16384,0.029417776399188574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,64,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,128,0.01920711166328854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,2560,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,3584,0.007785778078767989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,512,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,10240,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,10240,0.08402933014763726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,8192,0.34995823436313206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,10240,0.013080889152155982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,64,0.03870933254559835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,768,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,2560,0.07605333460701837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,2560,0.008833777573373582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,768,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,6144,0.2572302288479275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,10240,0.019201777047581144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,6144,0.012162666353914471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,5120,0.02330577704641554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,3584,0.019256888164414298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,6144,0.014290667242474027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,1024,0.0030213333666324615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,7168,0.10391200251049465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,12288,0.036842667394214206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,512,0.014588443769348992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,2048,0.019916445016860962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,12288,0.07757066355811225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,7168,0.05332977904213799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,7168,0.14103378189934626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,3584,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,12288,0.12804178396860758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,5120,0.012891555825869242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,32,0.004208000169859992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,768,0.007880888879299164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,3072,0.06892089049021403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,5120,0.027849776877297297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,2560,0.06744177473915948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,3072,0.03183555603027344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,10240,0.027263111538357202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,5120,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,32,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,1024,0.09277777539359199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,512,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,2560,0.0337413317627377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,768,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,512,0.04616444309552511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,51200,0.14829777346716985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,2560,0.02332533399264018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,51200,0.33392710155910915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,32,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,8192,0.005422222117582957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,256,0.002982222164670626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,1536,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,32,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,4096,0.01423733267519209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,32,0.017848889032999676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,3584,0.008393777741326226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,8192,0.010688888529936472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,3072,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,512,0.09421510828865899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,768,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,64,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,10240,0.15672443972693548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,256,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,3584,0.02632355524433984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,3072,0.01739288866519928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,16384,0.543892436557346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,2048,0.06738133562935723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,8192,0.07522400220235188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,51200,0.06322666671541002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,16384,0.0362026658323076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,10240,0.036321779092152916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,5120,0.4320755534701877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,128,0.003867555409669876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,768,0.015240889456537036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,256,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,8192,0.02402399977048238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,10240,0.2560204399956597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,8192,0.152383115556505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,10240,0.017254221770498488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,3584,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,65536,0.06469066937764485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,6144,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,1536,0.009413333402739631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,6144,0.03916977842648824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,10240,0.5814604229397243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,12288,0.05316977699597677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,7168,0.034276445706685386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,10240,0.04756266540951199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,16384,0.28876977496676975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,768,0.1840746667650011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,1024,0.06525778108172946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,8192,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,64,0.005489777773618698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,8192,0.02875822120242649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,3072,0.04187822341918945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,1536,0.019367999500698514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,6144,0.03549066517088149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,16384,0.020769778225156996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,2560,0.2629760106404622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,1024,0.0317848887708452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,4096,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,10240,0.03534844517707825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,64,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,8192,0.14601511425442165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,3072,0.010429333481523724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,16384,0.11596443918016221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,3072,0.004607999904288186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,6144,0.011956444217099084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,64,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,2560,0.019541333119074505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,65536,6.420788235134548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,4096,0.051232887638939746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,2048,0.006735999965005451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,65536,0.24403821097479927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,7168,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,8192,0.02499466637770335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,5120,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,1536,0.016778666112158034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,5120,0.01817244456874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,10240,0.02583644456333584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,2048,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,3584,0.011258666714032492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,256,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,16384,0.05794755617777506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,32,0.0025928889711697898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,1536,0.2241955598195394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,2048,0.007363555332024892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,32,0.03601244423124526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,512,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,3072,0.004876444323195351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,128,0.014109333356221518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,128,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,6144,0.06864177518420748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,51200,1.82116084628635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,16384,0.02351466649108463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,51200,0.15013244416978624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,51200,0.5547048780653212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,12288,0.013938667045699226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,51200,3.4786133236355252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,256,0.017521777086787753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,12288,0.2670088873969184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,6144,0.09848355584674412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,6144,0.010022222167915767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,10240,0.06998844279183282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,1536,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,7168,0.3957502312130398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,3072,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,51200,0.04931288957595825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,8192,0.016452444924248587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,8192,0.040694223509894475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,2560,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,12288,0.013753777576817406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,16384,0.03366311060057746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,7168,0.03589244352446662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,3072,0.008195555872387355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,1024,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,6144,0.026329777306980554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,32,0.003160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,1536,0.2423324320051405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,1024,0.023991111252042983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,12288,0.05077866713205973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,32,0.010405333505736457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,3584,0.023570666710535686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,10240,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,512,0.0027075554761621687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,4096,0.011424000064531961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,128,0.014513777361975776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,2048,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,1536,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,256,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,3072,0.0063582224150498705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,768,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,128,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,2048,0.007009777757856581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,65536,0.5173759990268284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,7168,0.012829333543777466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,128,0.014890667464998035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,51200,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,8192,0.03931466738382975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,65536,0.5616248978508843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,12288,1.2170915603637695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,1536,0.31503910488552517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,16384,0.23548444112141928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,65536,0.15064356062147352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,5120,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,10240,0.3077270984649658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,10240,0.020632000433074105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,8192,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,1024,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,32,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,51200,0.4400160047743056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,1024,0.29685155550638836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,3584,0.023769777682092454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,51200,0.11442044046190049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,1536,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,512,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,10240,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,16384,0.010168000227875179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,10240,0.30911821789211696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,7168,0.014035556051466199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,5120,0.19787200291951498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,51200,0.0614559981558058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,6144,0.022064000368118286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,3584,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,3584,0.025012445118692186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,1536,0.027907556957668726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,1024,0.013646221823162503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,64,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,10240,0.1709831158320109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,3072,0.019620445039537217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,51200,0.06783555613623725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,2560,0.02456355591615041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,65536,5.115513695610894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,3072,0.18317777580685088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,32,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,5120,0.016751999656359356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,3584,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,512,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,768,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,256,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,1024,0.013754666679435305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,3584,0.007481777833567725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,3584,0.060283554924858936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,1024,0.019888889458444383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,32,0.010504888991514841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,1536,0.015646222564909194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,8192,0.07089600298139784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,1536,0.008229333493444655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,3584,0.09049422211117214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,128,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,16384,0.061101330651177295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,1024,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,32,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,2048,0.016207110550668504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,8192,0.021663111117151048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,3072,0.019725332657496136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,3584,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,64,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,1536,0.009515555368529426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,1024,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,768,0.002797333316670524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,2560,0.10873599847157796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,51200,0.23083466953701445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,12288,0.016558221644825406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,3072,0.03847466574774848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,10240,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,1536,0.0041200001206662925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,5120,0.016136000553766888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,768,0.00309599998096625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,16384,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,7168,0.0795217752456665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,51200,2.9869751400417752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,12288,0.1802133321762085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,2048,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,6144,0.014631999863518609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,1024,0.03151022063361274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,256,0.011824888487656912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,4096,0.008296888735559251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,3072,0.025253333979182776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,8192,0.19771911038292778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,256,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,51200,0.1230800019370185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,12288,0.21510933505164254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,2048,0.022409778502252366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,7168,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,3584,0.10736533006032307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,64,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,5120,0.010648888846238455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,3584,0.10492710934744941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,2048,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,4096,0.021151110529899597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,768,0.005727111051479976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,5120,0.0653395586543613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,64,0.010278222461541494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,512,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,3584,0.047705776161617704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,16384,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,8192,0.24247911241319445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,2048,0.010747555229398938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,768,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,4096,0.05266844564014011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,6144,0.006019555446174409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,51200,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,6144,0.11976977189381917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,12288,0.3927653365665012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,12288,0.017154667112562392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,128,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,2560,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,12288,0.9277067184448242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,1024,0.0042133331298828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,6144,1.4028684828016493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,7168,0.25358488824632436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,10240,0.07046222024493747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,2048,0.006872889068391588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,64,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,7168,0.018008000320858426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,1024,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,768,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,3584,0.01643466618325975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,64,0.007341333561473423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,128,0.01179377817445331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,51200,0.03852622376547919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,12288,0.01908266709910499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,1024,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,7168,0.14303910732269287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,12288,0.08963644504547119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,12288,0.33526844448513454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,256,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,7168,0.8197742038302951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,512,0.01331999980741077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,16384,0.32135023011101616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,7168,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,12288,0.12886399692959255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,6144,0.13297066423628065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,1536,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,5120,0.4129146734873454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,1536,0.015821332732836407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,10240,0.13265244166056314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,768,0.023456000619464453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,64,0.021533333592944678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,65536,0.06883022520277235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,512,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,3584,0.058182219664255776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,512,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,768,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,12288,0.017443555924627516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,5120,0.03865333398183187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,256,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,2560,0.09999733501010472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,5120,0.2614640129937066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,8192,0.1399031082789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,2560,0.0070053330726093715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,16384,2.5094070434570312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,128,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,3584,0.1917466719945272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,512,0.003994666867785984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,3584,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,3584,0.019507555498017203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,51200,0.04829511046409607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,51200,0.445815987057156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,3072,0.05473600162400139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,65536,1.7570675743950739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,10240,0.5031004481845432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,1024,0.029096000724368628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,51200,0.06615733438067965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,256,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,1024,0.012741333080662621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,128,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,3072,0.04231999980078804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,3584,0.11126489109463161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,1536,0.06455555227067736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,65536,0.0962373349401686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,128,0.04544711112976074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,256,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,128,0.013205333716339536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,3072,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,1536,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,65536,0.9587617450290256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,12288,0.8315137757195367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,3072,0.018552000323931377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,32,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,768,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,64,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,6144,0.11635733975304498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,32,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,512,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,768,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,1024,0.01346933344999949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,64,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,5120,0.17886932690938315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,4096,0.029169778029123943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,256,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,2048,0.006694222076071634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,65536,0.07539466354582045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,10240,0.05256444546911451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,128,0.0018622221218215094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,3072,0.048235555489857994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,1536,0.01882488860024346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,10240,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,768,0.13068711757659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,768,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,2560,0.04297599858707852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,16384,0.1381333271662394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,768,0.005127111242877113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,1024,0.02051733268631829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,51200,2.3537769317626953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,51200,0.06554222106933594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,1536,0.015037332971890768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,5120,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,8192,0.06921333074569702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,768,0.023955555425749883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,65536,28.210220336914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,5120,0.04684799909591675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,8192,0.31712267133924693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,10240,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,768,0.018966222802797954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,1024,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,7168,0.1714773310555352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,512,0.008606221940782335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,16384,0.02664888898531596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,128,0.002055111030737559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,128,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,65536,0.17846311463250053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,10240,0.0042035554846127825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,128,0.017543110582563613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,16384,0.041054222318861223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,8192,0.474564446343316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,2560,0.28879555066426593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,8192,0.035360889302359685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,2560,0.020249777370029025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,512,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,256,0.03340177734692892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,128,0.006457777900828256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,1024,0.0028942223224374983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,51200,0.05058133271005419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,51200,0.10373955302768284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,1024,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,128,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,12288,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,6144,0.013182222015327878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,512,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,51200,0.05302222238646614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,4096,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,3072,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,6144,0.017855111095640395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,3584,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,12288,0.11751555071936713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,1024,0.036497778362698026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,3584,0.1294693284564548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,2560,0.02311022248533037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,512,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,512,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,65536,0.4484977722167969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,6144,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,7168,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,1024,0.06408355633417766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,2048,0.006175111151403851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,8192,0.019609777463806998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,6144,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,12288,0.024710221423043147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,4096,0.00665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,3072,0.021965333157115515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,1024,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,768,0.018779555956522625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,10240,0.08616533544328477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,768,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,2048,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,3584,0.012737777498033313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,256,0.011519999967681037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,7168,1.3750283983018663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,16384,0.2548702292972141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,32,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,4096,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,6144,0.016639999217457242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,1024,0.008008889026112026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,65536,4.134229448106554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,10240,0.016466667254765827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,4096,0.009642666412724389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,6144,0.011800000237094032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,2560,0.025591111845440332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,768,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,768,0.015492444237073263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,16384,0.06450488832261828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,128,0.0017911110901170308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,32,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,768,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,8192,0.023951111568344965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,5120,0.012975999878512489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,4096,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,16384,0.017229333519935608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,768,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,12288,0.09439288907580906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,3584,0.03647022114859687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,1024,0.295089774661594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,768,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,2560,0.026144888665941026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,2560,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,16384,0.022130666507614985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,5120,0.2831733226776123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,64,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,3072,0.016190222567982145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,3072,0.009413333402739631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,65536,0.9573688507080078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,512,0.027241778042581346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,512,0.00907555553648207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,8192,0.24335111512078178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,256,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,4096,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,51200,1.6255351172553167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,128,0.02092355489730835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,64,0.0025742221623659134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,256,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,1024,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,10240,0.09164266453848945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,6144,0.09129510985480414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,12288,0.040392001469930015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,12288,0.35975199275546604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,3072,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,51200,0.23220266236199272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,10240,0.25434401300218373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,65536,0.41517334514194065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,512,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,3072,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,1024,0.024124445186720953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,768,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,12288,0.04510133465131124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,7168,0.04237066706021627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,65536,0.26666755146450466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,65536,0.08530755837758382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,6144,0.16188266542222765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,16384,0.012057777908113269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,3072,0.013972444666756524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,256,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,7168,0.004602666530344221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,51200,2.216764450073242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,51200,0.27723556094699436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,5120,0.016900445024172466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,51200,0.451798227098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,65536,0.7267768647935656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,51200,0.06076444519890679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,768,0.06142933501137627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,8192,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,10240,0.04315288861592611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,65536,0.12920355796813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,64,0.00684799998998642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,10240,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,16384,0.3150542312198215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,8192,0.0730453332265218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,2048,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,8192,0.03827733463711209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,1536,0.019337778290112812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,1536,0.025240000751283433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,4096,2.11006842719184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,1024,0.04532266656557719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,256,0.09427111016379462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,51200,0.18139910697937012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,6144,0.04615022076500786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,16384,0.05658577548133003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,512,0.002599999929467837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,32,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,128,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,7168,1.1595884958902996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,2048,0.009975999593734741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,768,0.015584000282817416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,1024,0.017847110827763874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,2560,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,1024,0.020145777199003432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,2048,0.004580444345871608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,768,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,2048,0.30854490068223744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,2048,0.013574221895800697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,128,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,65536,0.744967089758979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,4096,0.23651554849412706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,3584,0.017602667212486267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,7168,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,4096,0.7766648928324381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,51200,1.0799404780069988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,768,0.007696000238259633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,8192,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,3072,0.09388711055119832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,4096,0.047071112526787647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,6144,0.015517334143320719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,51200,0.04699022240108914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,3072,0.009922666682137383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,12288,0.020744888318909537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,51200,0.1025564471880595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,768,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,8192,0.08622666862275864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,6144,0.008368000388145447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,1536,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,51200,0.03661422265900506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,51200,1.7913118998209636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,5120,0.034280889564090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,2560,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,8192,0.011016888750924004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,256,0.0053022222386466126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,3584,0.03527733352449205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,1024,0.06272444460127088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,2560,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,3584,0.0439991123146481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,5120,0.2117013269000583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,8192,0.07871199978722467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,512,0.04982044299443563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,6144,0.023762666516833838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,512,0.00886399961180157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,10240,0.06986933284335665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,512,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,64,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,12288,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,7168,0.13215111361609563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,65536,0.1264293326271905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,51200,1.1499449412027996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,6144,0.03640000025431315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,16384,0.11541599697536892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,5120,0.03148355417781406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,1024,0.1499697764714559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,8192,0.6661253505282932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,5120,0.05588710970348782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,3072,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,4096,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,1536,0.06964177555508085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,4096,0.05099466774198744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,512,0.05682133303748237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,12288,0.05901510847939385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,10240,0.043064888980653554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,5120,0.23735377523634169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,3072,0.4309413433074951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,4096,0.8690693113538953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,1536,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,7168,0.23573067453172472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,16384,0.035593777894973755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,7168,0.030709332889980737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,65536,0.18311556180318198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,2048,0.1022613313462999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,1024,0.045138667027155556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,12288,0.8316568798489041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,32,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,3072,0.010552888943089379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,5120,0.8046053250630697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,3072,0.06405422422620985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,3584,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,128,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,256,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,51200,8.619811164008246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,256,0.016015999846988253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,256,0.011161777708265515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,256,0.011514667007658215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,12288,0.08971466620763142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,128,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,128,0.015450666348139444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,64,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,7168,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,2048,0.017450667089886136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,32,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,2560,0.016647110382715862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,6144,0.011615999870830111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,65536,0.062252442042032875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,16384,0.7053644392225477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,768,0.017834666702482436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,8192,0.0464933349026574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,4096,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,1024,0.08090044392479791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,5120,0.012255111502276527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,3072,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,1536,0.005149333427349727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,6144,0.011688888900809817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,12288,0.5863315794203017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,512,0.003318222239613533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,3072,0.033703111939960055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,64,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,7168,0.04354222284422981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,5120,0.02992266747686598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,51200,0.049060444037119545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,2560,0.013284444808959961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,2048,0.06304622358746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,1024,0.0028880000528362063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,256,0.0032764443506797156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,10240,0.02586666742960612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,3584,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,7168,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,7168,0.010863111250930362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,7168,0.011680000358157687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,256,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,2048,0.07578577597935994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,768,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,7168,0.020652444826232064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,1536,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,512,0.01867644488811493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,768,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,65536,0.17813511689503989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,3072,0.011239111423492432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,8192,0.02628533376587762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,4096,0.009944000177913243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,8192,0.2627813286251492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,2048,0.08274755875269572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,2048,0.07336799965964423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,51200,0.16817866431342232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,128,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,5120,0.008741333252853816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,5120,0.014279110564125909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,2560,0.007000889215204451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,12288,0.013432000246312885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,512,0.023944000403086346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,3584,0.2742951181199816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,8192,0.14412444167666966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,7168,0.05025510986646017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,768,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,7168,0.023888889286253188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,2560,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,6144,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,8192,0.014147554834683737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,128,0.009663110805882348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,1536,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,65536,0.5551342434353298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,51200,0.12541688813103571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,7168,0.0891031093067593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,256,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,16384,0.020679111282030743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,2048,0.024318221542570326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,51200,2.0725875430636935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,12288,0.06184177928500705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,256,0.01163733336660597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,3072,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,1024,0.06828444533877902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,32,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,8192,0.0540906654463874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,7168,0.05072711242569817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,65536,0.17338932885064018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,768,0.02071911096572876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,1024,0.014724445011880664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,768,0.018800000349680584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,10240,0.12109600173102485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,1536,0.04893955588340759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,64,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,64,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,2048,0.07832177480061848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,8192,0.012975999878512489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,2560,0.026879111925760906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,256,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,3072,0.08597066667344834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,5120,0.026149332523345947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,51200,0.045313778850767344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,2048,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,2048,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,64,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,1024,0.16433866818745932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,5120,0.04485422372817993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,1536,0.01535022258758545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,51200,0.15274488925933838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,7168,0.03635377685228983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,12288,0.6437502437167698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,256,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,6144,0.0333262218369378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,1536,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,1536,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,6144,0.10874933004379272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,3584,0.359585788514879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,4096,0.022678222921159532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,768,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,16384,13.085494147406685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,16384,0.024824889169798955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,2560,0.02972888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,3072,0.02750933335887061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,256,0.017998221847746108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,2048,0.027344889110989038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,6144,0.06891910897360908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,16384,0.09643822246127659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,128,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,6144,0.11521244049072266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,12288,0.014690667390823364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,5120,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,2048,0.04523288872506884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,1536,0.03122933374510871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,3072,0.02830666634771559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,1536,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,1536,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,256,0.023757333556811016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,512,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,256,0.0063546668324205614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,6144,0.33449334568447536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,3584,0.027028444740507338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,3072,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,12288,0.5477635595533583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,12288,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,2560,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,5120,0.010205333431561789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,16384,0.017567111386193168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,256,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,2048,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,3584,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,8192,0.19281600581275093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,768,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,12288,0.16086488299899632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,10240,2.5548809899224176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,3072,0.09336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,4096,0.07746399773491754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,2048,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,5120,0.03439199924468994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,4096,0.024989333417680528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,32,0.007105777661005656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,2048,0.007073777417341868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,10240,0.01812533371978336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,3072,0.02482044365670946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,256,0.010803555448849997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,1024,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,512,0.0027706666539112725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,51200,0.17873244815402559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,2560,0.027368000812000696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,128,0.0032559999575217566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,10240,0.35840800073411727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,2560,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,12288,1.790951199001736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,4096,0.012130666938092975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,1024,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,51200,0.11020533243815105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,128,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,65536,0.06412622001436022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,4096,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,3072,0.019544000426928203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,8192,0.03275555703375075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,51200,0.26914223035176593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,32,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,1024,0.03167555398411221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,10240,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,768,0.016488000750541687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,2560,0.00489155575633049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,128,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,128,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,3072,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,1024,0.01260088880856832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,7168,0.05644444624582926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,16384,0.020407110452651978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,64,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,32,0.9441955354478625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,256,0.025040000677108765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,12288,0.052036444346110024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,1536,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,256,0.016193777322769165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,1024,0.0075191110372543335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,512,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,65536,0.13379377788967556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,32,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,5120,0.009759110709031424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,256,0.0587022238307529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,10240,0.035976890060636736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,12288,0.08076266447703044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,12288,0.039911998642815485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,1024,0.003935111065705617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,4096,0.02851466668976678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,1024,0.005980444451173146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,3072,0.046441776884926685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,1024,0.016203555795881484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,12288,0.016911110944218107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,4096,0.03632355398601956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,128,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,4096,0.03204177816708883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,5120,0.013631111217869652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,8192,0.2958933247460259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,3584,0.10480799939897324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,128,0.004127999974621667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,3072,0.01745244363943736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,51200,0.04945155647065905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,12288,0.013195555243227216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,512,0.01587288909488254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,12288,3.4841075473361545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,1024,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,3584,0.008440000315507254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,64,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,12288,0.15242577923668757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,1024,0.018205333087179396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,768,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,12288,0.013422222601042854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,51200,0.04006400042110019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,32,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,4096,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,4096,0.018185777796639335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,32,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,768,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,2048,0.03165600034925673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,6144,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,51200,0.1157288948694865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,768,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,16384,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,65536,0.015113777584499784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,51200,1.6094097561306422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,3072,0.06499822272194757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,768,0.02493688960870107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,32,0.001790222194459703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,1536,0.014724445011880664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,7168,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,64,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,16384,0.060792002413007945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,16384,0.0188417782386144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,5120,0.06924977567460802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,1024,0.008730666504965888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,4096,0.02375288969940609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,8192,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,51200,0.133297774526808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,2048,0.24921600023905435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,512,0.05730399820539686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,2048,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,1024,0.01109333336353302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,128,0.005055111315515307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,12288,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,6144,0.014041778114106921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,6144,0.08838133017222087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,2048,0.10905777745776707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,4096,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,65536,0.31992801030476886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,64,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,2048,0.027752000424597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,6144,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,51200,2.4317972395155167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,2560,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,51200,0.05587644709481133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,768,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,2048,0.10152622063954671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,768,0.09479377667109172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,6144,0.023215111758973863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,65536,0.0645208888583713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,3072,0.08350577619340684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,2560,0.007716444631417592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,32,0.00700711127784517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,5120,0.2929360071818034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,64,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,4096,0.0042151109211974675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,64,0.00792711145348019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,65536,0.494492424858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,1536,0.2599964406755236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,3584,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,2048,0.023834667272037927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,10240,0.04600977897644043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,7168,0.02699111070897844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,5120,0.037436445554097496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,128,0.006686222222116258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,7168,0.01819644371668498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,5120,0.02533600065443251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,6144,0.045231110519833036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,65536,0.600788434346517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,16384,0.9035484525892469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,8192,0.059466666645473905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,1024,0.04483733243412442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,7168,0.009382222261693742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,8192,0.6247759924994575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,64,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,1024,0.008651555412345463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,10240,0.03437155485153198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,16384,0.04518666532304552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,512,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,64,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,256,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,2560,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,3584,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,256,0.025390222668647766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,4096,0.018954666124449838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,32,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,1536,0.01854044364558326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,5120,0.025069332785076563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,256,0.0064488889442549805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,768,0.02438844409253862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,3072,0.02923911147647434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,256,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,2560,0.09310755464765762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,2048,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,256,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,128,0.019719999697473314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,3584,0.008728000024954477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,51200,0.09415288766225179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,16384,0.06778222322463989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,3584,0.19312443998124865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,2560,0.03222666515244378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,6144,0.013376889129479727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,3584,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,6144,0.024170666933059692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,16384,0.2998186747233073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,1024,0.2567697895897759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,3072,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,1024,0.03737777802679274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,2048,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,3584,0.011729777687125735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,4096,0.03752444518937005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,7168,0.0747342242134942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,12288,0.07649599843555026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,64,0.0020417778028382198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,16384,0.14562043878767225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,2560,0.012495111260149213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,768,0.09618044561809963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,2560,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,8192,0.022127999199761286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,51200,0.354494227303399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,10240,0.48697953754001194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,65536,0.06773955292171903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,5120,0.596144888136122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,10240,0.04112977782885233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,256,0.25342223379347056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,7168,0.013350221845838757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,2048,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,3072,0.017335111896197002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,2048,0.013433777623706393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,8192,0.012156444291273752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,1536,0.038438220818837486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,65536,3.284268485175239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,16384,0.07239644394980536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,1536,0.029059555795457628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,2560,0.04959377646446228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,5120,0.014839111102951897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,4096,0.011417778001891242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,65536,0.06995111041598849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,16384,0.3846755557590061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,256,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,8192,0.07409333520465426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,10240,0.0288702216413286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,51200,0.19931199815538195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,128,0.04180711176660326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,2048,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,3072,0.030660443835788306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,7168,0.023416888382699754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,12288,0.03387466735310025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,3072,0.05154755711555481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,768,0.025029333101378545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,51200,0.05187733305825127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,1536,0.025389333566029865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,2560,0.023764444722069636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,12288,0.4576186603969998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,512,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,1024,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,2048,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,2048,0.1704097721311781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,1536,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,1024,0.015797333584891427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,128,0.012060444388124677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,6144,0.026603556341595117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,16384,0.012323555847009024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,10240,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,10240,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,3072,0.5164888699849447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,128,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,12288,0.026714666022194758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,2048,0.0214355554845598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,1024,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,3584,0.05356533328692118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,8192,0.11281243960062663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,6144,0.016439111696349252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,3072,0.0626702242427402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,16384,0.04275733232498169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,8192,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,8192,0.14099289311303034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,7168,0.32797421349419487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,128,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,3584,5.639959123399522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,512,0.047482666042115956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,7168,0.011144888897736868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,1024,0.006416000011894438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,3584,0.24331911404927573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,4096,0.009149333669079674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,128,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,8192,0.010803555448849997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,3584,0.010296000374688042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,128,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,6144,0.10635289218690659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,51200,0.2506728967030843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,4096,0.004563555535342959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,2560,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,3072,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,4096,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,2560,0.0063564446237352155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,4096,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,4096,0.06312800116009183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,1024,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,4096,0.039155556095971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,5120,0.033668445216284856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,10240,0.022364444202846948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,16384,0.1179715527428521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,3072,0.012470222181744046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,2560,0.032413333654403687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,3072,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,51200,0.24869688351949057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,32,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,256,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,128,0.01463644372092353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,12288,0.06600177950329252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,5120,0.042479998535580106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,512,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,512,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,128,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,3584,0.03522488805982802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,1024,0.0028986665937635633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,10240,0.40765955713060165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,64,0.011898666620254517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,1024,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,512,0.05548888776037428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,768,0.0219217770629459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,7168,0.036687112516827054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,512,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,7168,0.03917600048912896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,51200,8.347859700520834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,1024,0.015511110424995422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,2560,0.05219199922349718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,4096,0.023802666200531855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,2048,0.007553777760929531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,32,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,128,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,6144,0.34510310490926105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,768,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,6144,0.02715555495686001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,1536,0.03166933192147149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,512,0.036794665786955096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,51200,0.1238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,12288,0.01982222166326311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,7168,0.5954888661702474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,51200,0.3428160084618463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,7168,0.06363288561503093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,7168,0.16450044843885633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,65536,0.053467555178536304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,3584,0.009087111386987898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,32,0.005296000176005893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,2560,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,1536,0.041343109475241765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,256,0.0022044444663657085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,51200,0.29930488268534344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,1536,0.010187555518415239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,4096,0.10213777754041885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,5120,0.046522667010625206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,1536,0.02720355490843455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,2560,0.07751199934217665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,7168,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,1024,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,8192,0.028370665179358587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,3072,0.07436444362004598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,7168,0.009936889012654623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,10240,0.06486222479078504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,4096,0.022654222117529974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,7168,0.011362666885058085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,2048,0.058944887585110135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,512,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,1536,0.05270311236381531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,1024,0.03552622265285916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,768,0.002625777696569761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,256,0.02162577708562215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,12288,0.01796000036928389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,1024,0.006538666784763336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,8192,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,768,0.00254133323000537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,128,0.005327111317051782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,65536,0.024365334047211543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,32,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,51200,1.1252240075005424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,2560,0.09163822068108453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,64,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,10240,0.0499946673711141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,512,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,6144,2.9582996368408203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,6144,0.1246675517823961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,256,0.0022133334229389825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,6144,0.03286666671435038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,4096,0.008564444051848518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,2560,0.012999110751681857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,2560,0.026050666968027752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,2048,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,768,0.0041351111398802865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,3584,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,65536,0.08447555700937907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,12288,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,5120,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,3072,0.007336888876226213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,2048,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,5120,0.287954674826728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,10240,0.029285334878497656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,7168,0.07595644394556682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,51200,0.045217778947618276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,8192,0.01540177729394701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,6144,0.040194667047924466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,1024,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,256,0.02423022190729777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,5120,0.015474667151769003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,3584,0.033203555477990046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,16384,0.05814488728841146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,2560,0.053898665640089244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,2560,0.08092088831795587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,10240,0.014956444501876831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,2560,0.5016453531053331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,10240,0.11407377984788682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,512,0.03355377912521362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,4096,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,8192,0.09088889095518325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,7168,0.15162221590677896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,10240,0.02425777746571435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,512,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,7168,0.020041777027977836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,256,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,1536,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,16384,0.045455998844570585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,6144,0.05705955293443468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,12288,0.08675733539793228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,32,0.004256000121434529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,1024,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,8192,0.08838578065236409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,256,0.022999111149046157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,5120,0.026297777891159058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,768,0.29139643245273167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,768,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,7168,0.11578844653235541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,4096,0.15388000011444092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,3072,0.06595644685957167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,4096,0.011320888996124268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,16384,0.060784889592064746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,5120,0.012915555801656512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,1024,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,2048,0.046911110480626426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,16384,0.3657653331756592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,51200,0.09017422464158799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,3072,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,1536,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,2560,0.025644444757037695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,7168,0.31441156069437665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,128,0.008371555142932469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,3584,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,512,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,3072,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,3584,0.034536888202031456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,3584,0.03154311246342129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,16384,0.21989066070980498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,51200,0.15090489387512207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,16384,0.09029422203699748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,51200,0.34666045506795246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,768,0.017266665895779926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,512,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,51200,0.050500445895724826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,512,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,7168,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,5120,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,2048,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,6144,0.17215555244021943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,256,0.0229111115137736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,2048,0.029177778297000464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,10240,0.03345422281159295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,4096,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,7168,0.0181377778450648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,2560,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,1024,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,1024,0.016033778587977093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,51200,0.06075466341442532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,1024,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,65536,0.01904088921017117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,2560,0.04535822073618571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,10240,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,12288,0.0733706686231825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,768,0.006326222171386083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,1024,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,768,0.01549600064754486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,51200,0.06682222419314914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,512,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,65536,0.08656799793243408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,768,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,10240,0.01713511182202233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,128,0.017134222719404433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,2048,0.045596443944507174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,2560,0.02968888812594944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,10240,0.010634666515721215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,51200,0.3886684576670329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,7168,0.011657777759763928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,256,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,64,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,128,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,16384,0.16822400357988146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,128,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,4096,0.02092711130777995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,51200,0.061588446299235024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,6144,0.055436445607079395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,8192,0.15860088666280112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,2048,0.01401600076092614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,2560,0.0436062216758728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,3584,0.2976417806413439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,7168,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,64,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,16384,0.07764266596900092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,65536,0.06785777542326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,2560,0.3936462137434218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,5120,0.7863679991828071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,12288,0.04263733492957222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,65536,0.08296977811389498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,16384,0.17727111445532906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,16384,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,128,0.013453333742088742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,5120,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,128,0.0019662222928471034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,7168,0.08602488703197902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,256,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,51200,2.226282755533854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,3072,0.027339556150966223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,3072,0.01237066669596566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,5120,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,1536,0.02479644450876448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,256,0.17236622174580893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,32,0.005335999859703912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,128,0.009479111267460717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,1024,0.01665244499842326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,1024,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,6144,0.00908799966176351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,6144,0.8252488772074381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,768,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,1536,0.009303111169073317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,3584,0.3499857849544949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,7168,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,6144,0.06919288635253906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,32,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,65536,0.12807733482784694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,51200,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,256,0.016542222764756944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,3584,0.03594933284653558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,768,0.003938666648334927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,7168,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,1536,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,6144,0.19711021582285562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,3072,0.06423111094368829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,4096,0.034796443250444196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,16384,0.03439377744992574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,512,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,2560,0.02509511179394192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,6144,0.0358497765329149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,256,0.005296000176005893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,12288,0.5483546786838108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,7168,0.4047368897332086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,10240,0.09607022338443333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,1536,0.02590399980545044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,4096,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,1024,0.02311644454797109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,128,0.013210666676362356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,10240,0.026828444666332666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,6144,0.021851556168662176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,8192,0.5728186501397027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,6144,0.29945511288113064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,12288,0.06016266345977783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,256,0.009558222360081142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,51200,2.446973376803928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,5120,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,6144,0.9451227188110352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,5120,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,3072,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,256,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,5120,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,5120,0.09739110867182414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,8192,0.03722133239110311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,5120,0.02308711078431871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,256,0.024752888414594863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,51200,0.07999733421537611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,256,0.01793599956565433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,1024,0.007166222565703922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,51200,0.6263830926683214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,2048,0.2060204479429457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,2048,0.013824889229403602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,12288,0.08032711346944173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,32,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,768,0.03157777918709649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,4096,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,256,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,1024,0.009115555220180089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,64,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,128,0.005734222216738596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,2560,0.19248178270128039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,4096,0.014312000738249885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,1024,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,256,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,10240,0.02721066607369317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,3072,0.007768889268239339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,7168,0.4978462325202094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,51200,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,5120,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,12288,0.017571555243598092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,51200,0.09380977683597141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,128,0.024998222788174946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,5120,0.04330755604637993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,10240,0.016604445046848722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,2560,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,6144,0.5106177859836155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,2048,0.05720444520314535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,64,0.005085333353943295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,512,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,64,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,512,0.017156443662113614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,128,0.01862844493654039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,1024,0.020224889119466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,16384,0.04470400015513102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,256,0.03019022279315525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,64,0.0020702222569121257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,3072,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,128,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,51200,0.07235111130608453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,128,0.0047386665311124595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,2048,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,1024,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,2560,0.04534222351180183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,128,0.0026231110095977783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,1024,0.14944622251722547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,51200,0.051296889781951904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,1024,0.07109333409203424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,1536,0.02016266683737437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,512,0.014508444401952954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,3072,0.007592000067234039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,5120,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,10240,0.014370666609870063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,4096,0.029093331760830347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,16384,0.055573330985175244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,65536,0.06310311290952894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,64,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,512,0.012915555801656512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,16384,0.023560000790490046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,65536,0.40788178973727757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,5120,0.026935110489527386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,65536,0.07675555679533216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,12288,0.1621626615524292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,3584,0.08139110936058892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,12288,0.03352888756328159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,1536,0.006053333481152852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,64,0.00370133337047365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,3584,0.010983111129866706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,3072,0.01862844493654039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,5120,0.01627200014061398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,65536,0.10358311070336236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,3584,0.23076534271240234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,2048,0.02400622268517812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,1024,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,1024,0.01648177703221639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,65536,0.4703511132134332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,65536,0.16443555884891087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,3584,0.015014222926563688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,3584,0.007904888855086433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,3584,0.009962666365835402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,5120,0.04462933209207323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,64,0.002789333255754577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,2048,0.006551110910044775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,7168,0.04208177659246656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,2560,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,2048,0.004283555679851108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,8192,0.5701431168450249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,8192,0.2856862280103895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,65536,0.2717822127872043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,4096,0.013312888642152151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,8192,0.2285919984181722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,256,0.0181022220187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,16384,0.04105600052409702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,512,0.03905333413018121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,10240,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,16384,2.7093662685818143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,51200,0.8776995340983073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,65536,0.520085334777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,7168,0.6308426856994629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,32,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,256,0.025117332736651104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,512,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,7168,0.029398222764333088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,2560,0.02280355493227641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,768,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,7168,0.22403555446200898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,2560,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,4096,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,1536,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,256,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,5120,0.16152800454033747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,7168,0.11299999554951985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,12288,0.029932445949978296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,3072,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,16384,0.3313475449879964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,7168,0.06528888808356391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,7168,0.020484444167878892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,256,0.0019022222194406723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,2560,0.00905511114332411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,16384,0.02018222212791443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,128,0.0058764442801475525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,51200,0.29021334648132324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,51200,0.3314266734653049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,128,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,256,0.005693333430422678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,768,0.014742222097184924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,10240,0.022356443934970435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,256,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,4096,0.1253982252544827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,12288,0.13302221563127306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,8192,0.03073244293530782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,8192,0.007024889190991719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,3072,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,7168,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,65536,0.16587111685011122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,51200,0.473739571041531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,512,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,7168,0.06938488615883721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,4096,0.04786488744947645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,1536,0.0214373336897956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,16384,0.042837331692377724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,3072,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,1536,0.17382755544450548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,256,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,512,0.003149333306484752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,51200,0.3733022212982178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,2560,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,3584,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,6144,0.03365244468053182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,3072,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,16384,0.04927466644181145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,3584,0.0643484459982978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,65536,0.05995021926032173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,1024,0.03528711199760437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,3584,0.018790221876568265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,512,0.002554666664865282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,64,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,65536,1.8588133917914496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,7168,0.046182221836513944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,1536,0.013336000343163809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,7168,0.00887022250228458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,65536,0.027796443965699937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,1024,0.024653333756658766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,3584,0.01039644413524204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,51200,0.06580355432298449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,12288,0.8337155448065864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,1536,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,768,0.021958221991856892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,8192,0.044569777117835156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,5120,0.024516445067193773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,128,0.03329066766632928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,10240,0.01408266690042284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,12288,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,256,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,3072,0.0628675553533766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,51200,0.8454924689398872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,16384,0.2661093340979682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,512,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,2560,0.015565334094895257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,6144,0.04765333400832283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,512,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,1024,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,2048,0.011370666325092316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,32,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,4096,0.0403128895494673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,768,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,512,0.012749333348539142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,4096,0.06900355550977919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,5120,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,6144,0.02665600015057458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,512,0.026592888765864905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,2560,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,512,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,51200,0.07309955358505249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,51200,0.04432266619470385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,3072,0.06059555874930488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,64,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,512,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,32,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,5120,0.02750933335887061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,10240,0.026145777768558923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,5120,0.06575200292799208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,3584,0.06258933411704169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,512,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,1536,0.022466666168636743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,51200,41.82911851671007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,64,0.019951999187469482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,64,0.003563555578390757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,3072,0.055288000239266284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,2560,0.030905776553683813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,32,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,768,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,8192,0.030934221214718286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,768,0.005231110999981563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,768,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,12288,0.07489955425262451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,512,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,6144,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,768,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,2560,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,3072,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,7168,0.4455564286973741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,64,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,8192,0.08156266477372912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,6144,0.14972000651889378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,128,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,12288,0.5890444649590386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,10240,0.0404097768995497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,1024,0.009704888694816166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,256,0.020591111646758187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,4096,0.008194666769769456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,5120,0.014752000570297241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,6144,0.13673422071668836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,5120,0.011232888533009423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,16384,0.016498666670587327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,1536,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,768,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,65536,0.2839404477013482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,4096,0.01553600033124288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,256,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,7168,0.11279643906487359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,768,0.02015555567211575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,4096,0.008270222279760573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,2048,0.006877333339717653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,65536,1.1554302639431424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,128,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,2560,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,3072,4.130429161919488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,6144,0.1356239981121487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,128,0.025069332785076563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,65536,0.06375466452704535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,51200,0.7209208806355795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,65536,0.20735732714335123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,6144,0.01182577759027481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,5120,0.3838168780008952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,12288,0.041364444626702204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,6144,0.2127093341615465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,3072,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,2560,0.005697777701748743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,4096,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,7168,0.021926222576035395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,64,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,1536,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,2048,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,6144,0.00905777762333552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,3072,0.06176977687411838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,65536,0.46228710810343426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,32,0.003839111162556542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,51200,0.08982933229870266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,768,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,1024,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,3584,1.142823961046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,10240,2.5955626169840493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,768,0.05298755566279093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,64,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,10240,0.04410488737954033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,51200,0.06352533234490289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,7168,0.09478755791982015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,5120,0.08627733257081772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,1536,0.023936000135209825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,768,0.0399582220448388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,1536,0.01108000013563368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,1024,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,8192,0.011120888921949597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,512,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,256,0.023863999380005732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,3584,0.020619556307792664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,51200,0.1081964439815945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,16384,0.0768595536549886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,32,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,10240,0.13404356108771429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,6144,0.01792711185084449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,512,0.007753777835104201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,3584,0.03324888812171088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,2048,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,2560,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,64,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,32,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,51200,0.7705991003248426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,128,0.014734221829308404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,256,0.02362755603260464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,128,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,2048,0.0427724454138014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,51200,0.442704889509413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,16384,4.588137732611762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,8192,0.06324266725116305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,6144,0.05031288994683159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,3072,0.12068977620866563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,256,0.025733333494928148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,10240,0.05176711082458496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,6144,0.023345778385798138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,256,0.017628444565667045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,3584,0.05559644434187147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,6144,0.021652445197105408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,1024,0.07146400213241577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,10240,0.029225776592890423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,7168,0.012710221939616732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,2048,0.03674666749106513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,8192,0.07813066906399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,3584,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,1024,0.011160888605647616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,10240,0.0772248903910319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,8192,0.014351111319330005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,1536,0.10434310966067845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,5120,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,12288,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,3584,0.06866577598783705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,51200,0.0782266656557719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,2048,0.06656177838643391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,2560,0.022856000396940444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,3072,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,2048,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,6144,0.17313510841793486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,6144,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,6144,0.12793955538007948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,1024,0.07004444466696845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,768,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,65536,0.0860862202114529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,10240,0.016805332567956712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,10240,0.02407111061943902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,32,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,768,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,256,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,5120,0.03397688931889004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,10240,0.015201777219772339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,10240,0.21866044733259413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,4096,0.03895911243226793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,5120,0.03279377685652839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,256,0.00981777740849389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,512,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,128,0.0021928888228204516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,768,0.03119644522666931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,2048,0.011679111255539788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,12288,0.2601546711391873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,4096,0.050408889849980675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,51200,0.8296435674031576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,768,0.02073777715365092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,51200,0.07138399945365058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,2048,0.02111022174358368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,8192,0.09185066488054063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,256,0.04485155476464165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,10240,0.32286400265163845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,6144,0.009198221895429824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,6144,0.013761777844693927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,768,0.01824355622132619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,4096,0.08816711107889812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,2048,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,3584,0.08409955766465928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,12288,0.06841778092914157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,8192,0.07236711184183757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,1536,0.015959110524919298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,3584,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,5120,0.4157706631554498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,2560,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,12288,0.16824532879723442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,12288,0.03186488813824124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,2048,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,1024,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,4096,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,64,0.002247111043996281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,65536,2.967373318142361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,7168,0.046282665597067944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,6144,0.018609777092933655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,3584,0.020807110601001315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,32,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,51200,0.011347555451922946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,3072,0.04849955439567566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,2048,0.019516444868511625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,7168,0.016744888491100736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,1536,0.010868444210953183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,768,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,65536,1.3362373775906031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,3584,1.3099768956502278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,10240,0.12573956118689641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,16384,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,6144,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,2048,0.010760889285140567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,128,0.014173333843549093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,1536,0.036525332265430026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,1536,0.0382008883688185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,1024,0.05173333485921224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,1024,0.016551110479566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,2048,0.020696888367335003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,256,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,7168,0.13957156075371638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,7168,0.03195022212134467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,2048,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,2560,0.22218667136298287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,2560,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,8192,0.15513511498769125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,2560,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,3584,0.03562488820817735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,1024,0.41813156339857316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,32,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,51200,0.24812089072333443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,7168,0.013844444519943662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,2048,0.025701334079106648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,6144,0.017909333109855652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,3584,0.028789334826999243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,2048,0.01070755554570092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,51200,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,16384,0.06163822280036079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,256,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,4096,0.031621333625581526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,5120,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,3584,0.010508444574144153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,2048,0.03625955515437656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,256,0.09871111313501994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,2048,0.4300435649024115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,4096,0.05742044581307305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,1536,0.23684888415866426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,1536,0.03512888815667894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,7168,0.02149244480662876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,51200,0.38797333505418563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,65536,0.5253582000732422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,3584,0.06287022431691487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,32,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,4096,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,7168,0.14680266380310059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,12288,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,6144,0.1389626661936442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,2560,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,1536,0.0678079989221361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,1024,0.2768799993726942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,65536,0.3938311205969916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,16384,0.045371555619769625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,2560,0.020555555820465088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,768,0.010016889207892949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,3072,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,12288,0.012616888516479068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,1024,0.004935111022657818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,10240,1.1886328591240776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,5120,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,3584,0.011437333292431302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,256,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,2048,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,2048,0.005493333356248007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,5120,0.036095112562179565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,5120,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,3072,0.013352889153692456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,5120,0.042544888125525586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,2048,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,8192,0.021535999245113794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,51200,0.050923556089401245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,1024,0.03603555427657233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,768,0.016926222377353244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,768,0.00369777778784434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,16384,0.02571822206179301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,3584,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,256,0.004217777815130022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,2560,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,12288,0.04179111123085022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,4096,0.020466667082574632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,2048,0.02388711108101739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,768,0.034985777404573225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,12288,0.08232977655198839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,3584,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,7168,0.02196088929971059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,12288,0.03401866555213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,10240,0.02919911013709174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,3584,0.032446222172843084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,6144,0.030236442883809406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,5120,0.014673777752452426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,5120,0.21125511328379312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,256,0.021564443906148274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,65536,0.05990488661660088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,2560,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,768,0.030704001585642498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,2048,0.1664711104498969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,1536,0.022780444886949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,128,0.0063528890411059065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,256,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,64,0.04383200075891283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,4096,0.04306844539112515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,16384,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,65536,0.10855911175409953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,8192,0.07811378108130561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,32,0.0025555555605226094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,2048,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,51200,0.03974399964014689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,4096,0.06866488854090373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,4096,0.08060444725884332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,51200,11.134292602539062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,2048,0.022668444448047217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,10240,0.05542666713396708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,4096,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,4096,0.03958400090535482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,4096,0.015492444237073263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,3584,0.01863555610179901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,12288,0.4425671100616455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,16384,3.060502158270942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,128,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,65536,0.06814933485454984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,2560,0.02164622147878011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,10240,0.07305066453086005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,768,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,512,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,2048,0.03542133172353109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,256,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,512,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,6144,0.11226488484276666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,2048,0.09069155322180854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,12288,0.013561777770519257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,8192,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,768,0.02402044501569536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,3584,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,10240,0.018599111172888014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,128,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,6144,0.017916444275114272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,5120,0.01644799941115909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,8192,0.016527111331621807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,2048,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,3072,0.007182222273614671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,2560,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,12288,0.0836835569805569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,768,0.025281776984532673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,16384,0.6025537914699978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,3584,0.03624177641338772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,6144,0.023952888117896184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,6144,0.4892622099982367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,8192,0.023367111881573994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,6144,0.17034222020043266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,4096,0.03815822137726678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,256,0.02719466719362471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,3072,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,12288,0.2808106740315755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,10240,0.03642755415704515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,12288,0.027094221777386133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,512,0.002982222164670626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,10240,0.012260444462299347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,3072,0.00674044465025266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,6144,0.03100799851947361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,10240,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,256,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,7168,0.030414223670959473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,1024,0.018128888474570382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,6144,0.032249778509140015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,1536,0.03153689040078057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,3072,0.02947910957866245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,1536,0.012057777908113269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,4096,0.10462755627102321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,1536,0.0816275543636746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,7168,0.02418222195572323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,768,0.01779022150569492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,768,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,2560,0.0644764436615838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,7168,0.028904888365003798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,4096,0.1704853375752767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,1536,0.014096000128322177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,512,0.02296888828277588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,128,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,32,0.034083555142084755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,768,0.0031315556002987754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,32,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,128,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,4096,0.007877333296669854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,768,0.027521777484152053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,12288,0.09989155663384332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,4096,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,16384,0.011699555648697747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,2560,0.005865777946180767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,65536,1.2051270802815754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,10240,0.0229777776532703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,3584,0.0407680008146498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,256,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,512,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,2560,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,64,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,10240,0.09532533089319865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,768,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,8192,0.014182221558358936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,768,0.028910222980711196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,3072,0.041323555840386286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,3072,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,256,0.017483555608325534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,2048,0.007037333316273159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,5120,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,1024,0.03052355514632331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,6144,0.06850755876964994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,32,0.0240346673462126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,256,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,512,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,2048,0.061034666167365186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,6144,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,7168,1.6837573581271703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,2048,0.061848892105950244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,512,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,12288,0.01439377831088172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,65536,0.09477155738406712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,16384,0.032079110542933144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,32,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,65536,0.046907554070154824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,1536,0.4460826449924045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,3072,0.0443466669983334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,32,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,6144,0.021895110607147217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,2048,0.0074924445814556545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,64,0.0019102222803566193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,1024,0.018970666660202872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,8192,0.011592888997660743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,2048,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,10240,0.4042088985443115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,256,0.04220622115665012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,12288,0.009768000079525841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,32,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,65536,0.06353600157631768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,12288,0.019718221492237516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,51200,0.05136177937189738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,1024,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,128,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,1536,0.01146311147345437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,1536,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,1024,0.009877333210574256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,8192,0.11830666330125596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,3072,0.18635022640228271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,8192,0.019835554891162448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,8192,0.01808088852299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,16384,0.02525866693920559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,8192,0.030051555898454454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,2048,0.05328177743487888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,4096,0.01648177703221639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,8192,0.04439111219512092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,256,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,3072,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,51200,0.07088533374998304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,64,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,256,0.02713511056370205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,2048,0.014514666464593677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,65536,0.06308444341023763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,6144,0.03238222334120009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,1536,0.0035786665976047516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,1536,0.16492356194390193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,2048,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,65536,0.6096844673156738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,8192,0.01295199990272522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,256,0.0023022221608294383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,10240,0.06009066767162747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,12288,0.0829111138979594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,5120,0.04274222254753113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,128,0.17158577177259657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,4096,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,3072,0.022337777747048274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,3584,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,1536,0.1146426730685764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,2048,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,1024,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,6144,0.027285334136750963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,5120,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,256,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,3072,0.008946666287051307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,7168,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,128,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,1024,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,7168,0.009462222456932068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,256,0.016143111719025504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,1536,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,12288,0.20767554971906874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,10240,0.181604438357883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,3584,0.02688177757793003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,768,0.01812177730931176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,32,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,3584,0.012353777885437012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,12288,0.06026577949523926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,51200,0.12644444571601018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,65536,0.06586488750245836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,3584,0.0904044442706638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,4096,0.010999999940395355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,16384,0.08574399683210586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,7168,0.10058310959074233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,65536,3.106168958875868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,65536,0.12513421641455755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,4096,0.07675111293792725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,64,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,10240,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,65536,0.4280382262335883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,12288,0.011199110911952125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,6144,0.03018666638268365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,4096,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,768,0.018985778093338013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,3072,0.05286222365167406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,4096,0.06055821975072225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,4096,0.010505778094132742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,7168,0.010658666491508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,1536,0.018086221483018663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,2560,0.16728356149461535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,12288,0.10223378075493707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,16384,0.020449777444203694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,7168,0.12797066900465223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,512,0.05923644701639811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,768,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,512,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,6144,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,2048,0.006839111033413146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,8192,0.014886221951908536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,5120,0.3093697759840223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,10240,0.10930399762259589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,768,0.011711111499203576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,256,0.042131556404961475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,16384,0.0705359975496928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,128,0.010209778116808997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,128,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,1024,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,4096,0.03176800078815884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,2560,0.006112888869312074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,256,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,1536,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,65536,0.07951911290486653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,1536,0.015400888191329109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,512,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,128,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,3584,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,51200,112.95414903428819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,1536,0.03633155425389608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,7168,0.35307910707261825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,768,0.009720889230569204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,768,0.022804444034894306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,1536,0.005173333403136996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,12288,0.0196115556690428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,768,0.023201778531074524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,256,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,51200,0.07701777749591403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,2048,0.0727564427587721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,3072,0.004238222208287981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,5120,0.025032889511850145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,256,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,16384,0.1339893341064453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,2048,0.00509688879052798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,6144,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,1024,0.05341955688264635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,65536,1.2037466896904838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,16384,0.023008000519540574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,65536,0.06500355402628581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,1536,0.0047048889100551605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,128,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,12288,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,4096,0.029623998536003962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,51200,0.5424906412760416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,51200,0.121960891617669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,512,0.1448391146130032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,65536,0.1827119986216227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,10240,0.7693751123216418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,768,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,6144,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,10240,0.06603111161126031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,256,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,128,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,128,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,4096,0.012825777961148156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,1536,0.016209777858522203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,10240,0.01587022178702884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,3072,0.016924444172117446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,512,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,8192,0.2045688894059923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,768,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,16384,0.20426577991909453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,1024,0.01998933321899838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,128,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,512,0.016153777639071148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,2048,0.046514666742748685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,51200,0.4215706719292535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,51200,0.6046000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,1536,0.008103999826643202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,512,0.022406222091780767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,2048,0.0852960016992357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,8192,1.9020587073432074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,3584,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,10240,0.07038399908277723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,8192,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,51200,0.36541154649522567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,128,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,4096,0.015804444750150044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,12288,0.02680711117055681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,2048,0.007043555378913879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,256,0.003137777869900068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,2048,0.1211928923924764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,6144,0.034075554874208235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,128,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,51200,0.1263306670718723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,6144,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,64,0.001771555592616399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,2560,0.017719111508793302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,1024,0.06171288755204943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,2560,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,65536,0.5752799775865343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,128,0.010195555786291758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,2560,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,16384,0.03823111123508877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,51200,0.15015999476114908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,10240,0.009938666390048133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,512,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,128,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,1536,0.0047591109242704184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,64,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,6144,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,2048,0.07158755593829684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,6144,0.05160266823238797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,128,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,256,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,6144,0.016945777667893302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,8192,0.026147555973794725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,8192,0.09305155277252197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,2048,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,16384,0.017268444101015728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,2048,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,2560,0.020598222812016804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,65536,0.30295199818081325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,256,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,512,0.040638221634758845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,16384,0.1312533352110121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,8192,0.6876871320936414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,1536,0.01755199995305803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,6144,0.016650666793187458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,768,0.02083200050724877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,1536,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,12288,0.02815822097990248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,3072,0.11490221818288167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,51200,15.50828382703993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,6144,0.043206221527523465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,512,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,16384,0.05630133549372355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,3072,0.02161600026819441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,768,0.02533333334657881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,10240,0.09500000211927627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,512,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,16384,0.02212711175282796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,1536,0.02365066607793172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,256,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,6144,0.1051902241177029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,512,0.07744710975223117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,128,0.016237333416938782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,3072,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,2048,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,8192,0.43037599987453884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,768,0.021578666236665513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,10240,0.04283644424544441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,16384,0.14066399468315974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,8192,0.647783120473226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,256,0.010948444406191507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,1536,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,16384,0.10504355695512559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,10240,0.0210826661851671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,512,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,256,0.01018488903840383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,5120,0.027816000911924574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,512,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,5120,0.013657777673668332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,128,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,512,0.019362666540675692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,7168,0.011971555650234222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,1024,0.029748443100187514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,16384,0.19579466183980307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,1536,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,7168,0.022651554809676275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,2560,0.026064000195927087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,768,0.01856355534659492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,65536,0.16806222332848442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,5120,0.027456000447273254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,2560,0.01309333327743742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,65536,0.07842666572994657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,5120,0.009121778110663096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,256,0.018353778454992507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,6144,0.007110222346252865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,8192,0.025309332542949255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,64,0.007516444557242923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,128,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,512,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,8192,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,512,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,12288,0.017461333009931777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,2560,0.004951999833186467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,768,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,51200,0.06638755400975545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,2560,0.013356444736321768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,768,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,10240,0.023366222778956097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,1024,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,1024,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,32,0.035622222556008235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,2560,0.02125599980354309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,2560,0.024880000286632117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,10240,0.13505155510372585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,5120,0.16477155685424805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,128,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,2560,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,12288,1.5079964531792536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,5120,0.026131555438041687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,65536,0.078639997376336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,256,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,8192,0.04333422250217862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,65536,1.0612701839870877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,51200,0.06528000036875407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,8192,0.06891822152667575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,6144,0.47931909561157227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,3584,0.04693511128425598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,65536,0.2536097897423638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,64,0.0021991110924217436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,12288,0.04700266652637058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,8192,0.07283288902706571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,32,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,3072,0.015881778465376962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,1024,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,65536,0.12222933769226074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,4096,0.12753421730465359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,10240,0.0436462230152554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,5120,0.028824888997607764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,5120,0.014985778265529208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,2560,0.02869066596031189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,3584,0.08380533589257134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,3072,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,2048,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,1536,0.022912888063324824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,51200,0.07087822092903985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,768,0.003181333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,3584,0.007359111474619971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,16384,0.024147555232048035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,768,0.03419288992881775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,32,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,7168,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,1536,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,5120,0.01667644414636824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,1024,0.06241244739956326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,8192,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,256,0.022805333137512207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,65536,0.5953111118740505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,5120,0.019253333409627277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,2560,0.03466488917668661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,32,0.002415999977125062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,10240,0.0864844454659356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,1024,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,32,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,1536,0.014874666929244995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,128,0.009550222092204625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,10240,0.02128533356719547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,12288,0.041189332803090416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,6144,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,7168,0.15916621685028076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,128,0.015535111228624979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,8192,0.11573778258429633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,10240,0.016896888613700867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,3072,0.03018310997221205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,7168,0.04507199923197428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,65536,0.4188506603240967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,8192,0.5986702177259657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,128,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,2048,0.016514667206340365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,128,0.008012444608741337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,12288,0.018008000320858426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,768,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,6144,0.007901333272457123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,256,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,768,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,12288,0.08457777897516887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,768,0.03083289000723097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,1024,0.012736000120639801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,2560,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,256,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,3584,0.051348444488313466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,51200,0.17121422290802002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,1536,0.028078221612506445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,8192,0.02608088817861345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,12288,0.03151822090148926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,3584,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,6144,0.07547999752892388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,3072,0.007068444457319047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,8192,0.03269688950644599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,7168,0.02846222122510274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,512,0.02469066613250309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,512,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,512,0.005524444497293896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,1536,0.016757332616382174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,512,0.011876444849703046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,6144,0.010827555424637265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,8192,0.14243022600809732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,16384,0.04390666551060147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,16384,0.018406222263971966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,8192,0.19050844510396323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,10240,1.6047671635945637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,64,0.010060444474220276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,2048,0.16045155790117052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,2048,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,4096,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,3584,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,65536,1.1155030992296007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,128,0.002557333351837264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,2048,0.025080889463424683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,8192,0.037455999188952975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,1536,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,7168,0.013743110828929476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,1024,0.03384977910253737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,3584,0.03160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,768,0.008912000391218398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,16384,0.0936773353152805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,3072,0.023372444841596816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,6144,0.00907555553648207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,256,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,8192,0.16507555378807917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,768,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,8192,0.039172444078657366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,3072,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,6144,0.038479109605153404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,128,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,128,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,8192,0.014828445182906257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,3584,0.1749040020836724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,512,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,768,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,5120,0.027447111076778833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,256,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,10240,0.1355066696802775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,3072,0.028551999065611098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,7168,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,10240,0.05830133623547024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,1024,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,12288,0.24519999821980795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,16384,0.044059554735819496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,4096,0.100673774878184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,12288,0.02827199962404039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,51200,0.348582214779324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,65536,0.06123822265201145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,1536,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,8192,0.02195466723706987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,5120,0.08726933267381455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,5120,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,512,0.01365600029627482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,768,0.02087733315096961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,6144,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,10240,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,1024,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,7168,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,3584,0.02772888872358534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,128,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,4096,0.016697777642144095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,768,0.00869066682126787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,8192,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,65536,0.24307645691765678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,4096,0.015295111470752291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,512,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,4096,0.064337776766883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,1536,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,2560,0.009836444424258338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,10240,0.010807111031479306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,1024,0.0122079998254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,3584,0.028925332758161757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,128,0.006199999815887875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,7168,0.009731555150614845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,3072,0.18755911456214058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,3584,0.007724444071451823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,1536,0.016215999921162922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,3584,0.019254222512245178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,512,0.022798221972253587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,768,0.01329066687160068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,512,0.2154524458779229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,65536,0.09786488612492879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,512,0.015622221761279635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,512,0.010800888968838586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,10240,0.04248799880345663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,51200,0.0640266670121087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,128,0.09011644124984741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,5120,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,1536,0.019810666640599568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,51200,0.4889341990152995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,3584,0.40490044487847227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,51200,0.08721599976221721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,10240,0.18149333530002168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,3584,0.05583466423882378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,3584,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,3072,0.017135999268955655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,5120,0.06858044200473361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,4096,0.030007110701666936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,4096,0.008020444048775567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,3584,0.02937600016593933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,512,0.03827555643187629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,6144,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,65536,1.0833066304524739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,512,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,3072,0.033144887950685285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,128,0.010107555323176915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,3072,0.005326222214433882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,8192,0.042183111111323036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,4096,0.015265777707099915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,1536,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,128,0.007118222614129384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,2048,0.013847110999955071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,1024,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,65536,0.4794666502210829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,5120,0.008815111385451423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,6144,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,1024,0.020615999897321064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,65536,0.06832800308863322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,256,0.016160888804329764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,128,0.022289777795473736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,1024,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,4096,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,3584,0.012119111087587146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,51200,0.06018222040600247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,64,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,128,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,128,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,2048,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,32,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,2048,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,2560,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,768,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,1536,0.03891288903024461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,256,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,7168,0.3372231059604221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,3584,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,10240,0.14486132727728948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,8192,0.03492355677816603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,512,0.009621333744790819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,2048,0.062189333968692355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,65536,0.10434666607115005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,64,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,64,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,16384,0.03571110963821411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,5120,0.006334222025341458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,4096,0.024358222881952923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,128,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,1024,0.05114755696720547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,12288,0.03488711184925503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,2560,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,10240,0.017453332742055256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,4096,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,2048,0.18541510899861655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,512,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,8192,0.1519040001763238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,12288,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,128,0.04253422220547994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,16384,0.06915911038716634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,4096,0.010632000035709806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,65536,0.07003200054168701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,1024,0.01648977730009291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,7168,0.04450311263402303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,1024,0.013083555632167392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,768,0.03901155458556281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,6144,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,51200,0.12967111004723444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,12288,3.080720901489258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,12288,0.030928889910380047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,3584,0.028607110182444256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,2048,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,128,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,10240,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,5120,0.039312889178593956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,512,0.16224533981747097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,512,0.08742488755120172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,16384,0.7309182484944662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,3584,0.31160977151658803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,2560,0.2389289008246528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,3072,0.032976001501083374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,128,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,7168,0.06359199682871501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,51200,0.053222222460640796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,8192,0.06933066580030653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,2560,0.024480000138282776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,10240,0.043140444490644664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,3072,0.08665955728954738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,128,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,16384,0.019699555304315355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,2048,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,1536,0.027270222703615826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,1024,0.010057777994208867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,64,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,65536,0.09734488858116998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,256,0.007032000356250339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,2560,0.49091646406385636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,3072,0.009750222166379293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,2560,0.01700266698996226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,256,0.007343110938866933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,51200,0.9385378095838758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,128,0.018544889158672757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,128,0.0199955552816391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,12288,0.021359999974568684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,512,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,65536,0.0162764439980189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,65536,0.12879378265804714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,5120,0.023931556277804907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,64,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,3072,0.016887111796273124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,3072,0.17183378007676867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,8192,0.014880888991885714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,32,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,2560,0.004552000098758274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,7168,0.07913688818613689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,64,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,5120,0.012061333490742577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,32,0.0019333333604865605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,512,0.10574755403730605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,3072,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,512,0.030928889910380047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,12288,0.01407999959256914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,5120,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,768,0.016170667277442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,1024,0.03832622369130453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,51200,0.045828445090187915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,768,0.040186666780047946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,2560,0.033366223176320396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,12288,0.10396533542209202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,65536,0.11982488632202148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,3584,0.023670223024156358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,16384,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,3584,0.02366755571630266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,8192,0.02270577847957611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,4096,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,3584,0.013507555756303998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,1024,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,12288,0.26808354589674205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,51200,0.40938933690388996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,8192,0.020121778051058452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,7168,0.033533334732055664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,4096,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,4096,0.012293332980738746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,4096,0.06553066439098783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,64,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,65536,4.9253794352213545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,12288,0.04422044422891405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,1024,0.00962755580743154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,3584,0.013537777794731988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,1536,0.009043555292818282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,1536,0.037864890363481306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,10240,0.2320542335510254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,3584,0.017253332667880587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,3072,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,12288,0.02120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,12288,0.01369955556260215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,128,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,3072,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,8192,0.013959111438857185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,16384,0.04386933313475715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,2560,0.00452888881166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,1024,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,768,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,5120,1.021691534254286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,8192,0.07133422295252483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,6144,0.009894222021102905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,51200,0.05958400170008341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,256,0.020617778102556866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,2048,0.01815555493036906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,5120,0.35117244720458984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,16384,0.07231466637717353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,128,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,3072,0.016573333077960543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,65536,0.021385777327749465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,12288,0.038027554750442505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,512,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,3584,0.008422222402360704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,1024,0.020690666304694284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,65536,21.15234375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,64,0.02612888813018799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,3072,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,8192,0.15764089425404867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,7168,0.2561546696556939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,2560,0.022859555151727464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,10240,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,65536,0.13434133264753553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,256,0.08535289102130467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,64,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,64,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,768,0.020296888218985665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,256,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,1536,0.06635822190178765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,16384,1.6429110633002386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,2048,0.38170933723449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,8192,0.019959110352728102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,10240,0.20683111084832084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,512,0.031191110610961914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,12288,0.042279998461405434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,768,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,10240,0.18091200457678902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,3584,0.06263733572430082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,12288,0.12930933634440103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,2048,0.0544177790482839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,12288,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,4096,0.029483556747436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,768,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,256,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,1536,0.020780444145202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,2560,0.025817778375413682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,8192,0.07466933462354872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,5120,0.5572862095303005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,4096,0.13547644350263807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,1536,0.03489244315359328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,4096,0.06385333008236356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,2048,0.012387555506494312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,4096,8.006689283582899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,4096,0.03007555670208401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,4096,0.029267556137508813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,8192,0.14298310544755724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,1024,0.052579555246565074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,3072,0.02249777813752492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,10240,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,51200,0.11358933978610569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,2048,0.5435377756754557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,7168,0.032497776879204646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,3072,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,51200,0.8404880099826388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,256,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,16384,0.19807022147708467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,2560,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,5120,0.06232711341645983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,1024,0.011760888828171624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,12288,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,256,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,1024,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,768,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,768,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,7168,0.04110044572088454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,128,0.03555999861823188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,12288,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,128,0.026351999905374315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,2048,0.028258668051825628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,512,0.01888088881969452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,1024,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,512,0.006748444504208035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,10240,0.06316888994640775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,1024,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,10240,0.05438311232460869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,16384,2.0205902523464627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,8192,0.46316178639729816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,8192,0.0398960014184316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,2560,0.06958666774961683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,3584,2.3377929263644748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,256,0.03013422091801961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,768,0.023401778605249193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,2048,0.014661333627170987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,2560,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,51200,0.32515109909905326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,1536,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,1536,0.030663109487957422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,7168,0.12924355930752224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,64,0.0017946666727463405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,3584,0.04326311084959242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,51200,0.056461334228515625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,16384,0.8794977929857043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,2560,0.022691556149058874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,65536,0.08077866501278348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,2048,0.0382008883688185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,65536,0.06333066357506646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,32,0.005509333478079901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,2560,0.034063110748926796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,128,0.004967999955018361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,5120,0.02753688891728719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,1024,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,2048,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,1024,0.17930044068230522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,1024,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,1536,0.022967111733224656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,8192,0.04962311188379923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,4096,0.00981955561372969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,6144,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,3584,0.03308355477121141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,256,0.0021093332519133887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,512,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,65536,0.10612089104122585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,768,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,3072,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,2560,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,768,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,7168,0.01752622259987725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,3072,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,128,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,8192,0.03576088945070902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,512,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,768,0.04319466816054451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,10240,0.029651555750105116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,256,0.012333333492279053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,8192,0.028755555550257366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,5120,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,7168,0.2768488989935981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,1024,0.003301333429084884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,16384,0.06359022193484835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,4096,0.09328622288174099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,256,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,2560,0.020576000213623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,2560,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,512,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,5120,0.030943112240897283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,512,0.007756444315115611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,65536,0.06808977656894259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,6144,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,12288,0.01809955636660258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,1024,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,2560,0.027318222655190363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,256,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,5120,0.020036444067955017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,2560,0.056477334764268666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,65536,0.9956080118815104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,256,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,12288,0.014468444718254937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,65536,0.046691556771596275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,7168,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,5120,0.024694222542974684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,256,0.015476443701320224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,7168,0.019137778215938143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,16384,0.04167644513977898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,3072,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,8192,0.02274488906065623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,6144,0.019544000426928203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,128,0.007890666524569193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,1536,0.07918044593599108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,65536,0.08598222335179646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,128,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,2048,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,3072,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,8192,0.03980177640914917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,1024,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,12288,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,3072,0.022249778111775715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,4096,0.034959110948774554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,3072,0.01642577846844991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,7168,0.08215200238757663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,32,0.009033777647548253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,128,0.021597334080272253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,51200,0.838081783718533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,3584,0.021613332960340712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,512,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,3584,0.02961155441072252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,10240,0.43388978640238446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,1024,0.017456000049908955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,64,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,4096,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,4096,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,7168,0.12514844205644396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,7168,0.012338666452301873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,1024,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,128,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,2560,0.010223111344708337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,256,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,51200,0.29576889673868817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,65536,0.06254844533072577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,51200,0.04650400082270304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,65536,0.01333866682317522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,5120,0.04068266683154636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,3584,0.009897777603732215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,768,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,6144,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,3584,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,51200,1.1145626703898113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,65536,0.7460080252753364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,7168,0.017823111679818895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,3584,0.031656887796190046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,8192,0.035338666703965925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,16384,0.04723199870851305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,2560,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,5120,0.014276444911956787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,128,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,512,0.030113776524861652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,4096,0.027911110056771174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,8192,0.020204444726308186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,128,0.025871111287011042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,51200,0.10515555408265854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,2560,0.045906666252348155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,7168,0.01942044496536255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,32,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,3072,0.07317155599594116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,1024,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,2048,0.007862221863534715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,3584,0.04003289010789659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,2560,0.009734222458468543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,8192,0.015042665931913586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,256,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,6144,0.06872800323698255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,256,0.018430221411916945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,3584,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,10240,0.04110666778352525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,768,0.02840533190303379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,7168,0.05488444368044535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,6144,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,3584,0.007690666450394525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,65536,0.06713244650099012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,12288,0.09868799977832371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,4096,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,64,0.01385600037044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,4096,0.019779556327395968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,4096,0.09815822045008342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,8192,0.19008444415198433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,65536,0.14820088280571833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,51200,0.7338924407958984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,12288,0.061456885602739125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,10240,0.014808000789748298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,16384,0.46160798602634007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,256,0.03753688931465149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,1024,0.3148053222232395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,10240,0.04519911275969612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,1536,0.013019555144839816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,3584,0.013756444056828817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,2048,0.048446221484078296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,8192,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,10240,0.0165928883685006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,1536,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,256,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,2560,0.01687466601530711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,1024,0.0028986665937635633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,768,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,51200,0.051889777183532715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,2560,0.015271110667122735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,512,0.03180799881617228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,10240,0.04112977782885233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,3584,0.19858488771650526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,65536,0.19720178180270723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,8192,0.010395555860466426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,3072,0.015255999233987598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,51200,0.25382399559020996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,3584,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,1536,0.003559111307064692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,512,0.01902933253182305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,64,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,256,0.01108888867828581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,64,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,5120,0.07254844241672091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,3072,0.01787911189926995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,32,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,1536,0.04266044497489929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,3072,0.013081777426931592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,3584,0.10418133603201972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,3072,0.026299556096394856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,10240,0.17260977956983778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,3584,0.010307555397351583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,12288,0.06175733274883694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,6144,0.08265866835912068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,8192,0.03803199860784743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,8192,0.0491466654671563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,128,0.006656000183688269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,51200,12.558124118381075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,51200,4.690377977159288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,65536,0.7210248841179742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,12288,0.023324444890022278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,256,0.003842666745185852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,10240,0.04051822092798021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,1024,0.014301333162519666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,2560,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,512,0.0511813329325782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,256,0.015170666906568738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,256,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,3584,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,16384,0.15469511349995932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,2048,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,3072,0.07403022050857544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,3072,0.03589955634540982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,3584,0.11301955911848281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,128,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,128,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,5120,0.2858355575137668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,51200,0.05031377739376492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,3072,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,3072,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,1024,0.033732444047927856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,2560,0.03065599997838338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,3584,0.02757866680622101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,8192,0.17556977272033691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,5120,0.22838489214579263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,2560,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,2048,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,32,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,6144,0.011971555650234222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,10240,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,3584,0.20922133657667372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,1536,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,512,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,768,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,3072,0.014646222194035849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,51200,0.3569457795884874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,10240,0.018029333816634286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,8192,0.0318177772892846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,256,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,3072,0.019157333506478202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,3072,0.021741333934995864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,2560,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,51200,0.22067377302381727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,3584,0.007894222107198503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,65536,0.12399821811252171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,1536,0.010885333021481832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,128,0.014841778410805596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,2048,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,64,0.002605333303411802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,6144,0.23601065741644967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,12288,7.725016276041667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,8192,0.03240711159176297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,3072,0.0042151109211974675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,2560,0.011897777517636618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,1536,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,512,0.02755555510520935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,65536,0.368106656604343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,4096,0.027127999398443434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,256,0.007123555574152205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,4096,0.24986844592624238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,8192,0.07679111427730985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,256,0.03923111160596212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,51200,1.151004473368327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,3584,0.024382222029897902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,5120,0.012355555262830524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,8192,0.1202239990234375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,64,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,768,0.030914667579862807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,4096,0.6922933260599772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,768,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,1536,0.03889866669972738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,12288,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,32,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,6144,0.010681777364677854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,51200,0.02146044373512268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,256,0.006677333265542984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,51200,0.23443288273281523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,2560,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,1536,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,4096,0.020149333609475028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,2048,0.00665155549844106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,32,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,65536,0.027823110421498615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,5120,0.021947556071811255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,16384,0.26293688350253636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,2560,0.5764026641845703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,4096,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,256,0.02664888898531596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,12288,0.05359199974271986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,10240,0.12159732977549235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,12288,0.315342214372423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,128,0.0229768885506524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,3584,0.013471110827393003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,768,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,32,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,51200,0.06322044134140015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,512,0.022611555125978258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,512,0.07711111174689399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,10240,0.018179555733998615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,256,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,12288,0.01626311077011956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,2560,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,4096,0.11865333716074626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,12288,0.06513333320617676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,7168,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,16384,0.149052447742886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,8192,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,128,0.005381333331267039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,6144,0.029118223322762385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,2048,0.07460533248053657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,4096,0.021379555265108745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,4096,0.16849422454833984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,768,0.05188622077306112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,10240,0.025419554776615564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,512,0.0053839998112784485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,16384,0.2070666684044732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,512,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,6144,0.021300445000330608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,8192,0.07326755258772108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,2048,0.06176533301671346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,2560,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,12288,0.029623111089070637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,6144,0.02680711117055681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,1536,0.14108178350660536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,1024,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,6144,0.05822222100363838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,51200,1.5718328687879775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,16384,0.020449777444203694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,5120,0.044919110006756253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,7168,0.5320089128282335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,65536,0.3855626583099365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,51200,0.06203999784257677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,8192,0.013919111755159167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,3584,0.01484444406297472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,3072,0.03347911106215583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,4096,0.024680889315075342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,65536,0.5694488949245876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,768,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,51200,0.7679662174648709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,512,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,128,0.019886222150590684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,6144,0.024616888827747766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,7168,0.06702488660812378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,8192,0.0136808885468377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,512,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,768,0.10440178049935235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,1024,0.019347555107540555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,5120,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,12288,0.23369778527153862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,64,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,512,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,5120,0.020198222663667467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,12288,0.04240355557865567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,65536,0.2708231078253852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,16384,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,3584,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,16384,0.09213244252734715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,1536,0.023383110761642456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,512,0.06260622209972806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,64,0.0329368876086341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,512,0.2131955491171943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,65536,0.07476177480485705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,51200,0.01315199997689989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,7168,0.004280888785918553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,6144,0.016940444707870483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,7168,0.06380622254477607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,10240,0.01707466608948178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,2048,0.04944088723924425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,3584,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,2560,0.05837421947055393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,3584,0.06320444742838542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,8192,0.022679110368092854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,12288,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,16384,0.1915271149741279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,12288,0.04988622334268358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,8192,0.0413973331451416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,1536,0.01679733395576477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,256,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,12288,0.3351999918619792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,5120,0.013728888498412238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,51200,0.22396000226338705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,1024,0.14182755682203504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,16384,0.03129511078198751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,65536,3.0056817796495228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,2560,0.2318204508887397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,7168,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,128,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,256,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,32,0.05937244494756063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,5120,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,2048,0.007197333706749811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,7168,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,6144,0.01034755590889189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,51200,0.5165173212687174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,512,0.0029057777590221832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,1536,0.06664178106519911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,32,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,51200,0.03395911057790121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,128,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,4096,0.013632888595263163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,6144,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,128,0.02234488891230689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,768,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,4096,0.09387377897898357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,65536,1.086419529385037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,2048,0.036532445086373225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,65536,0.0644328859117296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,256,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,3584,0.060773332913716636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,32,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,8192,0.09061421950658162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,10240,0.04390577806366814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,16384,0.04001244571473863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,12288,0.2137297789255778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,16384,0.03886311252911886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,4096,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,5120,0.6638302273220485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,128,0.04822133315934075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,51200,1.413711971706814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,512,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,1024,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,6144,0.028552889823913574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,1024,0.07228710916307238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,12288,0.017101332545280457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,3072,0.041963554090923734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,3584,0.040435555908415056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,51200,0.16820977793799508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,1536,0.024861334098709956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,32,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,1024,0.023954666323131982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,1536,0.029915554655922785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,5120,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,1024,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,12288,0.04168355464935303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,10240,0.9176355997721354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,3072,0.04338577720854017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,1024,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,8192,0.013678222066826291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,256,0.010375111467308467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,16384,0.034244444635179304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,7168,0.09034755494859482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,768,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,16384,0.2422942320505778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,3072,0.020854221449957952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,1536,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,768,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,2560,0.02497600018978119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,2048,0.0365973346763187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,4096,0.06621599859661527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,3584,0.01254666679435306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,1024,0.00599644457300504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,5120,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,6144,0.018945778409639995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,8192,0.17874044842190215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,16384,0.1042871144082811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,2560,0.015896888242827523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,64,0.01312355531586541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,3072,0.07537333170572917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,51200,0.4487937821282281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,2048,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,10240,0.6131511264377171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,65536,0.13114933172861734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,3584,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,3072,0.015139554937680563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,3584,0.023208000593715247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,16384,0.07867022355397542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,51200,0.21738489468892416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,768,0.018224888377719455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,3584,0.015991111596425373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,768,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,16384,0.02397955622937944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,65536,0.39791732364230686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,512,0.002584888910253843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,512,0.225966215133667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,512,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,1536,0.050797333319981895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,1536,0.04680622286266751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,4096,0.12351644039154053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,8192,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,5120,0.031698667340808444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,768,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,7168,0.0351475543446011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,12288,0.02614044480853611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,128,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,3584,0.6719475322299533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,4096,0.02442666722668542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,12288,0.0163093325164583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,3072,0.046395556794272534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,5120,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,256,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,51200,11.567552354600695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,6144,0.038183109627829656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,8192,0.10130133231480916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,7168,0.03622488843070136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,2560,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,128,0.017343999611006845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,256,0.013543999857372709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,1536,0.005993777679072485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,4096,0.09406844774881999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,12288,0.06238933404286703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,512,0.005678222411208683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,3584,0.020008888509538438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,6144,0.08771110905541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,16384,0.9930737813313802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,1536,0.02466222147146861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,2560,0.0120319997270902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,6144,0.04535555508401659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,256,0.01650311052799225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,4096,0.01590399940808614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,1536,0.045151111152437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,7168,0.04442310995525784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,3584,0.08296800322002836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,2048,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,2560,0.08761777480443318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,7168,0.028357333607143823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,4096,0.19449244605170357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,10240,0.5815591282314724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,64,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,32,0.008648888932334052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,7168,0.017875555488798354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,3584,0.04385066694683499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,3072,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,3072,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,128,0.009871111147933537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,6144,0.35601245032416445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,12288,1.0495920181274414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,4096,0.01405244403415256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,768,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,4096,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,12288,0.131294223997328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,8192,0.27668534384833443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,7168,0.22548532485961914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,1536,0.03140799866782294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,51200,0.500382211473253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,4096,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,32,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,128,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,768,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,1024,0.30614932378133136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,512,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,2048,0.04141244292259216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,8192,0.025036444266637165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,512,0.023230221536424425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,768,0.0041742221348815495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,8192,0.08027199904123943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,1024,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,10240,0.036632001399993896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,512,0.013389333254761167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,51200,0.19792889224158394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,65536,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,768,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,6144,0.29786756303575307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,1024,0.06158133347829183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,6144,0.009928888744778102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,65536,0.35441332393222386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,6144,0.08784977595011394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,5120,0.00981244444847107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,65536,0.06391555733150907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,16384,0.3618977864583333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,3072,0.08787111441294353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,7168,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,768,0.0273626661962933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,10240,0.04225688841607836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,5120,6.403536054823134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,2048,0.019191111127535503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,1024,0.21144178178575304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,1536,0.03964266512129042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,8192,0.262103107240465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,5120,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,10240,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,51200,0.5558817651536729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,8192,0.023678221636348303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,7168,0.027667555544111464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,16384,0.02481066683928172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,3072,0.007522666619883643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,128,0.0022400000856982335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,2048,0.04128444525930617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,1024,0.01016711112525728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,10240,6.262310028076172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,5120,0.04470577836036682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,768,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,512,0.026280888252788123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,51200,1.304642677307129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,10240,0.5687342219882542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,4096,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,8192,0.042916443612840444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,5120,0.02481599979930454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,7168,0.03957510987917582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,1536,0.005967999911970562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,4096,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,1536,0.03830577929814657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,3584,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,65536,0.29348267449273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,5120,0.5981973542107476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,5120,0.26851820945739746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,512,0.03595111105177138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,6144,1.6370355818006728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,512,0.02254666719171736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,256,0.009373333719041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,7168,0.03056266572740343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,12288,0.01419911119672987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,12288,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,6144,0.009763555394278632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,5120,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,10240,0.1771555609173245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,7168,0.16100266244676378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,64,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,16384,0.07072799735599093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,2048,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,7168,0.40400266647338867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,5120,0.027091556125217017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,4096,0.021190222766664293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,51200,2.211481730143229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,128,0.01742488808102078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,1024,0.06621599859661527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,32,0.06697333521313138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,512,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,1536,0.019902222686343722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,4096,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,8192,0.9033084445529513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,5120,0.01258488910065757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,51200,0.5025484297010634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,65536,0.18136266867319742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,16384,0.0583475563261244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,4096,0.010258666343159145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,5120,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,3072,0.05405511127577888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,5120,0.0992444422509935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,7168,1.5043057335747614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,65536,0.4862382147047255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,512,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,5120,0.007837333612971837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,16384,0.04445688923199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,768,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,12288,0.016013332539134555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,51200,0.9402462641398112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,5120,0.49762757619222003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,1024,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,8192,0.0662355555428399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,32,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,1536,0.015676443775494892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,128,0.02130844526820713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,10240,0.10248800118764241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,16384,0.19057867262098524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,12288,0.009758222434255812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,1024,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,12288,0.09369511074490017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,1536,0.0063386667105886675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,12288,0.058437334166632764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,6144,0.1000231107076009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,2560,0.014498665928840637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,4096,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,10240,0.07473244269688924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,5120,0.0329440004295773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,16384,0.4228257868025038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,1536,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,3584,0.040639999839994646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,65536,4.649677276611328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,2048,0.009905777871608734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,128,0.0181697772608863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,1536,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,16384,0.10885244607925415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,3072,0.2390657795800103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,6144,0.04916710986031426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,16384,0.04643555482228597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,768,0.022472888231277466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,12288,0.01903022163444095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,10240,0.13683556185828313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,1536,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,1024,0.03755555550257365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,6144,0.008779555559158325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,3072,0.02241955531968011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,128,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,2560,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,16384,2.5055698818630643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,2560,0.019665777683258057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,1536,0.018142221702469718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,65536,0.13536889023251003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,128,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,7168,0.009863999982674917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,1024,0.00327822214199437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,3072,0.09864711099200779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,12288,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,5120,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,64,0.018583110637134977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,1024,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,6144,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,2048,0.04122844338417053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,3072,0.007213333414660559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,128,0.025215999947653875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,4096,0.10007733106613159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,12288,0.027237332529491846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,7168,0.4367786778344049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,12288,0.012192000117566852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,3072,0.0074311114019817775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,1024,0.061675555176205106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,5120,0.026181333594852026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,32,0.005642666584915585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,10240,0.025955556167496577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,4096,0.014647111296653748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,65536,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,65536,0.21731911765204537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,1536,0.006666666517655055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,4096,0.41824889183044434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,51200,0.38753332032097715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,16384,0.0210844443904029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,512,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,2048,0.079839997821384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,65536,0.053393777873780995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,2560,0.012484444512261285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,51200,0.32265154520670575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,1536,0.021087111698256597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,2048,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,7168,0.013704000247849358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,4096,0.009062222308582729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,128,0.005991999887757831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,1024,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,3584,0.015284443895022074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,8192,0.01864355636967553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,32,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,3584,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,5120,0.01130488928821352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,8192,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,2560,0.1164204412036472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,6144,0.04828000068664551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,512,0.12276977962917751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,7168,0.004944889081848992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,16384,0.02124711043304867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,1024,0.022653333014912073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,65536,0.41579733954535586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,64,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,7168,0.028196444114049275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,3072,0.03012711140844557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,16384,0.016187555260128446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,256,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,8192,0.03084711233774821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,51200,0.30568001005384654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,5120,0.15686578220791286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,2048,0.025630222426520452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,65536,1.048192024230957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,12288,0.014021333720948962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,65536,0.9190248913235135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,256,0.017796445224020217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,4096,0.010072888599501716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,3072,0.009081777599122789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,1536,0.027950220637851294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,1024,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,51200,0.03498311175240411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,10240,0.2500222259097629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,3584,0.010079999764760336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,1536,0.038154668278164335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,12288,0.03190844588809543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,51200,2.6052534315321183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,7168,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,51200,0.05974666939841377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,1024,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,1024,0.019904888338512845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,6144,0.011784888803958893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,1024,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,10240,0.015537778536478678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,16384,0.18969778219858804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,7168,0.04720088839530945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,65536,0.14248889022403294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,512,0.01862133377128177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,65536,0.09791733158959283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,51200,1.2860221862792969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,1536,0.02370311154259576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,2560,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,16384,0.2715333302815755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,51200,0.9913138283623589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,256,0.055607113573286265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,32,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,768,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,1536,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,128,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,7168,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,128,0.019215111931165058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,7168,0.14357511202494302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,7168,0.05261422197024027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,8192,0.8918755849202474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,12288,0.37488712204827207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,1024,0.025399110383457605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,65536,0.27657688988579643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,8192,0.011125333607196808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,65536,0.2116702265209622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,10240,0.38798220952351886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,51200,2.0655199686686196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,4096,0.02695200012789832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,3072,0.031685332457224526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,768,0.023207111491097346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,65536,0.28163467513190377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,51200,0.07585510942671034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,2560,0.08631377749972874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,10240,0.039339555634392634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,5120,0.036814222733179726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,128,0.01552088889810774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,3072,0.008099555969238281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,2048,0.011135111252466837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,5120,0.028192000256644353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,4096,0.007883555359310573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,6144,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,10240,0.1941173341539171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,6144,0.057116442256503634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,64,0.0018088887963030075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,8192,0.03923377725813124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,5120,0.1328222221798367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,4096,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,16384,0.025809778107537165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,10240,0.02127644419670105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,10240,0.1097795565923055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,2048,0.011612444288200803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,64,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,1536,0.061167114310794406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,256,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,5120,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,128,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,3072,0.03750133183267381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,16384,0.03983111182848612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,10240,0.011778666741318174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,8192,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,5120,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,12288,0.11186577214135064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,2560,0.029197331931855943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,16384,0.04303644431961907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,3584,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,65536,7.319736904568142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,512,0.03666399916013082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,4096,0.0083155557513237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,12288,0.03150666753451029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,2048,0.027621333797772724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,16384,0.04500622219509549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,5120,0.20095466242896187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,256,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,51200,2.3694631788465714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,10240,0.050512890021006264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,32,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,12288,0.04762399858898587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,12288,0.3066844410366482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,3584,0.3077751000722249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,3584,0.03695644603835212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,8192,0.02370133333735996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,12288,4.4917856852213545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,128,0.03306577934159173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,128,0.02165599995189243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,12288,0.09841244750552708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,4096,0.0346959994898902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,2048,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,10240,0.17491200235154894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,2560,0.05576622486114502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,4096,0.04137244489457872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,128,0.015798222687509324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,256,0.019464888506465487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,2560,0.017649778061442904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,64,0.10173333353466457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,12288,0.3362026744418674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,8192,0.10942577653461033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,8192,0.014850666125615438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,65536,0.1621519989437527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,768,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,1024,0.017870222528775532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,1536,0.03224800030390421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,10240,0.03154133425818549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,1536,0.03714222378200955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,32,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,10240,0.03705599904060364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,16384,0.03499822152985467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,768,0.02049333353837331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,768,0.008028444316652086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,16384,0.05630489190419515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,64,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,2048,0.05153600043720669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,8192,0.010618666807810465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,32,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,10240,0.16376266214582655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,768,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,12288,0.03641600079006619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,512,0.0026337777574857077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,5120,0.01661155621210734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,1536,0.008047999607192146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,2048,0.18646577994028726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,12288,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,16384,0.018914666440751817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,4096,0.12131467130449082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,3584,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,3584,0.07568622297710843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,2560,0.04976088802019755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,768,0.006762666834725275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,7168,0.013843555417325763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,512,0.005751111027267244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,64,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,6144,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,2048,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,5120,0.02104088829623328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,6144,0.05910399887296888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,1536,0.009504000345865885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,10240,0.22764534420437285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,10240,0.058913780583275684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,2048,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,256,0.0022248888595236675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,5120,0.010102222363154093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,512,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,3072,0.04416799876425001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,64,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,51200,0.4076275560590956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,1536,0.027056000298923914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,512,0.15882578161027697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,512,0.013824889229403602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,7168,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,32,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,4096,0.008591999610265097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,6144,0.009280888570679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,65536,0.6788969039916992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,6144,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,12288,0.042140444119771324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,3072,0.027629334065649245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,8192,0.07579822010464139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,256,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,768,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,3072,0.38134490119086367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,6144,0.14645244015587702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,16384,0.022648000054889258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,2048,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,32,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,512,0.029458665185504492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,1536,0.05717155668470594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,128,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,2048,0.004257777912749184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,7168,0.09580355882644653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,5120,0.0655919975704617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,1024,0.03596977723969354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,64,0.013648889131016202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,5120,0.032419555717044406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,128,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,5120,0.015636444091796875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,51200,0.12734311156802708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,16384,0.11807289388444688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,7168,0.040729777680502996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,10240,0.022987556126382615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,1024,0.010925333533022137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,256,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,7168,0.07512266768349542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,7168,0.1354693306816949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,2560,0.3857022126515706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,64,0.007725333174069722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,768,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,5120,0.03129511078198751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,5120,0.040729777680502996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,128,0.012262221839692859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,5120,0.01684799955950843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,256,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,2048,0.016953777935769822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,8192,0.018787556224399142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,5120,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,16384,3.588708241780599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,1024,0.02660622199376424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,768,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,3072,0.025750221477614507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,2048,0.015297777122921415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,6144,0.06907200151019625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,65536,0.725136015150282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,2048,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,4096,0.016944888565275405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,512,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,3584,0.06626755661434598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,5120,0.07668622334798177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,16384,0.08192711406283908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,6144,0.018912000788582694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,2560,0.02274844381544325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,4096,0.007870222131411234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,51200,0.02755555510520935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,65536,0.35943465762668186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,16384,0.07703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,10240,0.02015555567211575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,6144,0.03681244452794393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,8192,0.024844444460339014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,2560,0.0961644450823466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,4096,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,64,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,2048,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,7168,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,12288,0.017156443662113614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,128,0.15573599603441027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,51200,0.020727111233605277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,2048,0.08232977655198839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,2048,0.02563200063175625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,16384,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,7168,0.06379199690288968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,2048,0.021163556310865615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,65536,0.1793768935733371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,4096,0.024413333998786077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,256,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,12288,0.03558311197492812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,16384,0.022622222701708477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,12288,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,65536,0.10738578107621934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,3584,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,7168,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,64,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,256,0.01200888885392083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,2560,0.0075253330998950535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,12288,0.025973333252800837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,4096,0.022710222336981032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,65536,0.9150942696465386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,16384,0.04721066686842176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,51200,0.35772355397542316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,65536,0.03769422239727444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,768,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,65536,0.25264178382025826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,512,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,3584,0.16678310765160453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,2048,0.006772444479995304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,51200,0.052991112073262535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,7168,0.02569599946339925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,7168,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,6144,0.04067555401060317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,12288,0.13753599590725368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,128,0.008377778033415476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,4096,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,7168,0.025382222400771245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,128,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,12288,0.10899111297395493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,16384,0.041492445601357356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,51200,0.04840622345606486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,256,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,8192,0.18448889255523682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,6144,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,3072,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,1536,0.06904711325963338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,2560,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,65536,0.06269511249330309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,65536,1.2647875679863823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,768,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,1024,0.04180444280306498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,12288,0.018274666534529794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,7168,0.028027554353078205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,51200,0.514827569325765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,12288,0.029186666011810303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,51200,0.16119999355740017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,1024,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,10240,0.035827553934521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,3072,0.02758666707409753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,768,0.02295555505487654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,256,0.012254222399658628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,256,0.009190222455395592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,7168,0.029731555117501154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,6144,0.08066044251124065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,32,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,12288,0.06014221906661987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,2560,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,4096,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,4096,0.008199999729792276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,3072,0.1847511132558187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,51200,0.25806132952372235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,512,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,1024,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,2560,0.007189333438873291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,65536,1.5688435236612956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,8192,0.01794933279355367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,128,0.02264533274703556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,6144,0.016961778203646343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,3584,0.08794577916463216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,10240,0.022628444764349196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,10240,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,12288,0.07689688603083293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,512,0.03587111168437534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,32,0.01036444471942054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,7168,0.1723804473876953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,4096,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,6144,0.07751644319958158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,51200,0.01201244443655014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,8192,0.17088977495829263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,12288,0.029191109869215224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,256,0.3587635623084174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,65536,0.06916177935070462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,2560,0.027375110321574744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,1024,0.03876266545719571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,64,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,51200,0.05077244506941902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,1536,0.02535022298494975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,7168,0.04264533188607958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,6144,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,5120,0.033598221010631986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,64,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,768,0.01716977854569753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,8192,0.052912000152799815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,1024,0.029478222131729126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,8192,0.029397332006030615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,768,0.038147555457221136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,16384,0.051642666260401406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,3584,0.037477334340413414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,128,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,65536,0.061101330651177295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,128,0.01997866729895274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,5120,0.678443537818061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,1024,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,1536,0.02365155518054962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,8192,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,10240,0.10107466909620498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,16384,0.2841840055253771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,12288,0.07502310805850558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,12288,0.03450489044189453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,128,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,1024,0.009053332938088311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,2560,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,6144,0.01822933389080895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,65536,0.10124711195627849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,3584,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,16384,0.09179288811153835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,32,0.013668444421556262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,10240,0.020442666278945077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,256,0.03700444433424208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,1536,0.020488889680968392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,8192,0.009541333549552495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,51200,0.10427111387252808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,768,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,768,0.013607111242082385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,1024,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,512,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,5120,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,65536,0.3060133457183838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,5120,0.025176000263955858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,8192,0.01349422252840466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,51200,0.578205320570204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,16384,0.03804888990190294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,1536,0.003559999995761447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,51200,0.5149280230204264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,32,0.09345244699054295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,4096,0.21282845073276094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,2048,0.007104000283612146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,512,0.009752000371615091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,8192,0.016208888755904306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,8192,0.00701333334048589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,7168,0.0645973351266649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,3072,0.03386755453215705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,768,0.006719111154476802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,5120,0.04496622085571289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,64,0.0074035558435651995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,4096,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,512,0.09841510984632705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,2560,0.020375111036830477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,4096,0.028667555914984807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,128,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,64,0.017676444517241586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,1536,0.03173777792188857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,2048,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,6144,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,12288,0.09785244199964736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,65536,0.08663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,16384,0.07908888657887776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,4096,5.011341518825955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,512,0.21355022324456108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,51200,0.08107999960581462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,5120,0.025582222474945918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,3072,0.03494400117132399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,768,0.0036195553839206696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,51200,0.11516888936360677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,256,0.11180799537234837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,256,0.034816887643602155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,8192,0.023815110325813293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,32,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,7168,0.028703110085593328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,3584,0.03289777702755398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,6144,0.02211644417709774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,2048,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,6144,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,12288,0.04298311140802172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,256,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,16384,0.3227333227793376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,768,0.010643555886215635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,256,0.07505955298741658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,768,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,16384,0.03139999839994643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,12288,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,2048,0.06459999746746488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,7168,0.01147911118136512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,65536,0.12075822883182102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,7168,0.12482399410671657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,128,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,2560,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,1536,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,10240,0.07889155546824138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,12288,0.027091556125217017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,65536,0.5580568843417698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,16384,0.06718577941258748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,1024,0.0038728887836138406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,3584,0.013627555635240344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,7168,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,128,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,512,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,12288,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,6144,0.3614542219373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,128,0.0041964443193541635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,1536,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,256,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,10240,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,64,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,2048,0.021194666624069214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,5120,0.15002666579352483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,8192,0.040475553936428495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,3072,0.04598933458328247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,768,0.0032213332338465583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,768,0.003539555602603488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,65536,0.04688355657789442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,512,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,5120,0.03208355440033807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,8192,0.02331377731429206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,65536,0.13055022557576498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,4096,0.6715413199530708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,1024,0.008023111356629265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,8192,0.040431999497943454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,5120,0.009071110851234859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,1536,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,2048,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,256,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,512,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,1024,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,64,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,8192,0.03489333391189575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,2048,0.020023999942673575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,16384,0.07559644513660006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,768,0.059750219186147056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,2560,0.007575111256705389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,2048,0.01311644415060679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,2560,0.1615333292219374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,2560,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,1536,0.0741786691877577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,512,0.020586666133668687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,256,0.003338666632771492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,3072,0.09363200267155965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,8192,0.019260444574885897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,64,0.006775110960006714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,12288,0.14291111628214517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,1536,0.8056355582343208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,16384,0.685931576622857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,1024,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,51200,0.05706933471891615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,8192,0.14897067017025417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,2048,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,32,0.043959998422198825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,1536,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,51200,0.06582044230567084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,16384,0.020015110572179157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,256,0.008379555410808986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,128,0.006539555473460092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,4096,0.4955724610222711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,256,0.1232266690995958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,256,0.012833777401182385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,51200,0.07054311037063599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,16384,0.23024090131123862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,8192,0.035063998566733465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,1536,0.006689777804745569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,4096,0.04943022131919861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,5120,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,12288,0.07641600237952338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,2048,0.031173331869973078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,8192,0.34923821025424534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,8192,0.07009689013163249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,12288,0.05686933464474148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,6144,0.017647999856207106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,512,0.010643555886215635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,3072,0.04633333285649618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,3072,0.30808620982699925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,6144,0.019863999552196927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,3072,0.021640888518757288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,256,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,51200,1.4648400412665472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,2560,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,7168,0.02850133180618286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,16384,0.030748443471060857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,512,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,5120,0.05124800072775947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,6144,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,1536,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,65536,0.8184764650132922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,51200,0.1942951149410672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,3072,0.016532444291644625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,3072,0.025136889682875738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,2560,0.0064195555945237475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,6144,0.2190399964650472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,1536,0.006309333360857434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,2048,0.05171111226081848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,1536,0.008753777378135258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,3072,0.14441955089569092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,6144,0.1001982225312127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,2560,0.0808302230305142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,128,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,65536,0.07638488875495063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,3584,0.06231111288070679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,4096,0.004548444516128964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,32,0.0029111111329661477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,4096,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,256,0.016919111212094624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,16384,0.15690667099422878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,6144,0.15428800053066677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,3584,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,7168,0.018823999497625563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,51200,0.10794311099582249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,16384,0.24586576885647246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,8192,0.06695289081997342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,4096,0.04960088928540548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,1536,0.011214222345087262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,1536,0.008513777620262569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,2048,0.010393777655230628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,1024,0.04047288828425937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,65536,0.17341332965426973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,51200,0.8683030870225695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,2560,0.02188355558448368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,3072,0.012489777472284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,768,0.016896888613700867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,16384,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,768,0.025778666138648987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,8192,0.03832711113823785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,2048,0.03891999853981866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,1024,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,1536,0.0282008879714542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,5120,0.47019290924072266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,10240,0.030524445904625788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,1536,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,1536,0.08210666974385579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,8192,0.1541235579384698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,10240,0.04526844289567736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,2048,0.018417777286635507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,32,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,5120,0.03453777896033393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,16384,0.042678223715888135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,32,0.015388444066047668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,8192,0.02275288932853275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,6144,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,3584,0.03558489018016391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,16384,0.14484000205993652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,768,0.024832000335057575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,10240,0.02235466738541921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,5120,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,1024,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,65536,0.07674400011698405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,1024,0.011949333051840464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,256,0.07664977841907077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,3584,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,1024,0.040468444426854454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,51200,0.0533537765343984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,768,0.017430222696728177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,2048,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,32,0.017455110947291057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,3584,0.029002666473388672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,256,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,512,0.017278222574128043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,128,0.018872888551818002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,65536,0.11052711142434014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,10240,0.5141448974609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,4096,0.08382044235865276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,6144,0.0315048893292745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,128,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,2560,0.01907111042075687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,7168,0.3292942312028673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,12288,0.042415112257003784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,2048,0.011492444409264458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,10240,0.1030622190899319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,256,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,2560,0.020987555384635925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,7168,0.024041778511471216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,64,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,512,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,2560,0.00739555557568868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,2560,0.01569422251648373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,4096,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,64,0.006005333529578314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,6144,0.05734044313430786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,256,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,7168,0.35666489601135254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,1024,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,64,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,128,0.0064533332155810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,8192,0.026893332600593567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,512,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,7168,0.03576711151334975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,3584,0.22266756163703072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,1024,0.0031475555151700974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,16384,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,6144,0.021935110290845234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,3072,0.04862577716509501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,10240,0.016206221448050607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,32,0.013565333353148567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,7168,0.0288355549176534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,8192,0.06760978036456637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,32,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,768,0.005707555760939916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,64,0.010431999961535135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,12288,0.06542933649486966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,2560,0.01571466690964169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,10240,0.060124443637000195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,5120,0.04375377628538343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,12288,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,512,0.006079111248254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,256,0.04021688964631822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,3584,0.007428444094128079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,8192,0.026405332816971674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,12288,0.1060853335592482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,5120,0.029779556724760268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,5120,0.050048000282711454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,8192,0.00962222201956643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,5120,0.04792266752984789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,4096,0.017272889614105225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,5120,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,12288,0.2567608886294895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,12288,0.052427556779649526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,1024,0.022459555003378127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,3584,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,65536,0.07179822524388631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,3584,0.05341955688264635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,5120,0.05761866437064277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,6144,0.0304097765021854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,10240,0.07082666953404744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,8192,0.02595022155178918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,3584,0.3323964542812771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,3072,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,51200,0.22962045669555664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,64,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,5120,0.009535999761687385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,8192,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,10240,0.07073777914047241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,128,0.021715554926130507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,6144,0.10537244213951959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,2560,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,128,0.00360622215602133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,12288,0.07350399759080675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,1024,0.005713777823580637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,16384,0.04866400029924181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,16384,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,8192,0.1697324381934272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,128,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,2560,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,5120,0.020075556304719713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,7168,0.021260443660948012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,65536,0.2227395640479194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,1536,0.006315555423498154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,2560,0.013972444666756524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,4096,0.014727110664049784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,2560,0.05099022057321337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,2048,0.007327111230956183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,1536,0.01982222166326311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,4096,0.012776888906955719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,7168,0.012723555167516073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,2560,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,1536,0.03203999996185303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,64,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,12288,0.025403555896547105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,4096,0.7064986758761935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,768,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,7168,0.02091466718249851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,4096,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,2560,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,3072,0.17523111237419975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,7168,0.03710844450526767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,12288,0.5731084611680772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,64,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,16384,0.0454328887992435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,16384,0.06455288992987739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,64,0.008608889248636033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,10240,0.21520445081922743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,6144,0.038210666841930814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,8192,0.01748533381356133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,3072,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,5120,0.03120622369978163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,2048,0.038380444049835205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,8192,0.45117775599161786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,65536,1.2056248982747395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,4096,0.01276533305644989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,10240,0.053040888574388295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,128,0.014757333530320061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,7168,1.0042212804158528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,8192,0.050682667228910655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,512,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,16384,0.294959995481703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,4096,0.01127911110719045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,5120,0.09165422121683757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,1536,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,51200,0.22791732682122123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,128,0.025052444802390203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,4096,0.013812444276279874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,1536,0.013560000393125745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,8192,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,3072,0.005706666658322017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,1536,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,256,0.017490666773584154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,64,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,1536,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,256,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,1536,0.003555555724435382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,65536,0.05843200286229452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,12288,0.08510844575034247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,10240,0.03148711058828566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,16384,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,3584,0.010242666635248397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,32,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,5120,0.009062222308582729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,7168,0.027136888768937852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,8192,0.010748444332016839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,32,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,1536,0.014178666803571912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,64,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,51200,0.02994133366478814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,5120,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,7168,0.013188444077968597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,5120,0.07738133271535237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,3072,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,51200,0.04003644320699904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,8192,0.15829688972897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,2048,0.17297511630588105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,32,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,2048,0.03911022345225016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,128,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,16384,0.043644444810019604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,512,0.0041271112859249115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,32,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,51200,0.48133330874972874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,6144,0.20207022296057808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,3072,0.015357333752844067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,2048,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,6144,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,1536,0.06506133079528809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,4096,0.05436266793145073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,51200,0.07260800070232816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,512,0.02978666623433431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,1024,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,4096,0.030264000097910564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,3584,0.05968533621893989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,8192,0.035014222065607704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,8192,0.011138666835096149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,32,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,4096,0.009938666390048133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,16384,0.01687733332316081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,1536,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,5120,0.0988044473859999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,3584,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,2560,0.021623111433453027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,7168,0.7297680113050672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,6144,7.5883695814344625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,10240,0.17875199847751191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,32,0.0649706655078464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,65536,0.454445309109158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,7168,0.08393066459231907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,5120,0.018960888187090557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,1536,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,256,0.021915555000305176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,3072,0.055367112159729004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,65536,1.9175306955973308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,3584,0.047761778036753334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,3584,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,65536,1.1861786312527127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,4096,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,16384,0.04052711195415921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,4096,0.047116445170508496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,5120,0.027311111489931744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,1536,0.02156266735659705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,51200,8.214770846896702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,16384,0.023718222975730896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,4096,0.11353422535790338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,1024,0.10901066992017959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,16384,0.7323039902581109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,64,0.08459555440478855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,7168,0.012659555508030785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,3584,0.022504889302783545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,32,0.08644888798395793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,3072,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,3072,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,65536,0.0667111078898112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,3584,0.016599110431141324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,1536,0.019745777050654095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,51200,0.6322275797526041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,4096,0.9870835410224067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,3072,0.04165066613091363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,3072,0.04887466629346212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,2048,0.06463644239637586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,2048,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,16384,0.019181332654423185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,51200,0.04828088813357883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,5120,0.07437511285146077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,16384,0.04898755417929756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,256,0.0032417778339650896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,128,0.014848889576064216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,4096,0.0928586655192905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,10240,1.997820536295573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,16384,0.027091556125217017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,64,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,8192,0.9270346959431967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,128,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,512,0.005500444521506627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,51200,1.7045422659979925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,3584,0.011695110963450538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,10240,0.07204266389211018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,10240,0.016467556357383728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,51200,0.1283982197443644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,3584,0.013832000394662222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,12288,0.030064887470669214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,3584,0.03607733382119073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,10240,0.03341333402527703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,1536,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,1024,0.11991822719573975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,3584,0.28812265396118164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,12288,0.006329777754015393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,12288,0.05046311020851135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,512,0.014926221635606555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,8192,0.11662222279442681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,16384,0.33654401037428117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,3584,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,64,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,3584,0.020967110991477966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,12288,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,32,0.08993155426449245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,3584,0.012664000193277994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,4096,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,3584,0.8202382193671333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,2048,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,3072,0.00942755573325687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,5120,0.22921511862013075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,8192,0.03633866707483927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,16384,1.2380506727430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,128,0.04296444522009956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,5120,0.29817067252265084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,3584,0.09478400150934856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,64,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,512,0.0361768901348114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,10240,0.018926221463415358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,6144,0.12463643815782334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,1024,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,1536,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,65536,0.9785262213812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,7168,0.07202044460508558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,65536,6.991554684109158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,1024,0.022837332553333704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,32,0.016688000824716356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,7168,0.03539911243650649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,32,0.24614577823215059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,16384,0.017849778135617573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,1536,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,768,0.0881262222925822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,4096,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,6144,0.11680266592237686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,5120,0.01756088932355245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,2560,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,3072,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,10240,0.3098204400804308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,2048,0.009528888596428765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,10240,0.02678666677739885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,1536,0.007375111182530721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,2560,0.12474311722649468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,51200,0.1174515618218316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,2048,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,2560,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,65536,0.9343421724107531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,512,0.027183110515276592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,10240,0.05320177806748284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,1024,0.004534222185611725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,51200,0.05128088924619886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,7168,0.14875911341773138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,3072,0.013280889226330651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,7168,0.01969688965214623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,3072,0.2769271002875434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,2048,0.017200888858901132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,1024,0.02627022233274248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,1536,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,65536,0.7211173375447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,10240,0.6523306634691026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,32,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,7168,5.6172985500759545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,6144,0.49970044030083555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,7168,0.01999288962946998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,2048,0.008560000194443597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,16384,0.016211556063758004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,1024,0.18671288755204943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,5120,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,2048,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,32,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,65536,0.31581065389845103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,51200,0.11419822110070123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,7168,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,512,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,2048,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,16384,0.06883200009663899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,1024,0.023177777727444965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,5120,0.06327466832266913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,256,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,256,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,7168,0.03461066550678677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,2048,0.013630222115251752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,7168,1.3742969301011827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,256,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,4096,0.08999733130137126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,7168,0.20888888835906982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,51200,2.141510221693251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,768,0.11694222026401097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,65536,0.5774275461832682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,3584,0.0964755548371209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,10240,1.1021893819173176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,128,0.016241777274343703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,8192,0.46512619654337567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,512,0.14056177934010824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,32,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,6144,0.011355555719799466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,1536,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,64,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,1536,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,64,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,2048,0.09819377793206109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,1536,0.015279110934999255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,10240,0.02331999937693278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,128,0.008394666843944127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,10240,0.018575110369258456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,8192,0.14398045010036892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,3584,0.054233776198493115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,128,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,512,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,12288,0.10709422164493138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,8192,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,1024,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,10240,0.1286275519265069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,4096,0.05733422438303629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,16384,0.02015822297996945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,4096,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,1536,0.011247110863526663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,8192,0.458238230811225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,4096,0.03408711155255636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,3584,0.018937778141763475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,12288,0.02049511174360911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,2560,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,3072,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,768,0.07411466704474555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,3072,0.013087111214796701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,256,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,6144,0.021494223011864558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,32,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,7168,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,6144,0.03413777881198459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,65536,0.01924533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,51200,5.6168089972601996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,3584,0.01071466671095954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,16384,0.16384711530473497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,1024,0.19452977180480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,128,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,12288,0.08842577536900838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,65536,0.06249155600865682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,1024,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,7168,0.04776800009939405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,512,0.03122044603029887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,16384,0.06367111206054688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,65536,0.09485689136717056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,3584,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,128,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,16384,0.14783467186821833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,512,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,768,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,65536,0.09824355443318684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,51200,0.19845333364274767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,2560,0.03296800123320685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,5120,0.0719839996761746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,1536,0.012808000048001608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,16384,0.07140444384680854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,4096,0.15572888321346706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,16384,0.02197866638501485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,128,0.1919004387325711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,4096,0.025751110580232408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,65536,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,51200,0.10666400194168091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,2560,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,12288,0.00851644410027398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,12288,0.020277332928445604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,2560,0.01961688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,51200,11.658433702256945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,768,0.03651555710368686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,512,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,3072,0.021367111139827307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,64,0.05664800273047554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,32,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,3072,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,10240,0.43952533933851456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,8192,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,256,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,2560,0.03085600005255805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,128,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,2560,0.016965332958433364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,2048,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,128,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,4096,0.016019556257459853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,5120,0.04549422197871738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,10240,0.017530666457282174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,10240,0.012351999680201212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,256,0.013384888569513956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,32,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,5120,0.058444446987575956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,2560,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,4096,0.0373253325621287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,2560,0.8545093536376953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,256,0.02254044512907664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,6144,0.1942088868882921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,256,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,1536,0.016930666234758165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,10240,0.027428444888856676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,5120,0.07283466392093234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,4096,0.025368889172871906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,512,0.021942221456103857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,32,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,3072,0.03756355577045017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,51200,0.0929715567164951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,65536,0.09067822164959377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,65536,1.0446622636583116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,1024,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,65536,0.11292978127797444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,51200,0.013864888913101621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,10240,0.012806221842765808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,2560,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,3584,0.014567999376191033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,8192,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,6144,0.041760888364579946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,16384,0.21176444159613717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,3072,0.015990222493807476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,256,0.03103289008140564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,10240,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,3584,0.00835111074977451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,3584,0.041144887606302895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,16384,0.12456177340613471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,2560,0.006865777903132969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,8192,0.3780968983968099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,12288,0.02023555503951179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,64,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,64,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,4096,0.011414222419261932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,128,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,2048,0.008015111088752747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,2560,0.007479111353556315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,16384,0.18156088723076713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,16384,0.1286097764968872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,8192,0.03450399968359206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,3584,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,7168,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,512,0.017863111363516915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,5120,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,6144,0.0358586675590939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,512,0.023989333046807185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,3584,0.014357333381970724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,10240,0.016470222009552848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,64,0.09247733487023248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,51200,0.12385422653622097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,1536,0.03829777903027005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,65536,0.338227563434177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,1536,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,3584,0.01680888897842831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,2048,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,51200,0.10950311024983723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,512,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,10240,0.10325688785976833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,2048,0.009798222117953831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,256,0.013155555559529198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,4096,0.049754665957556836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,1024,0.005425777700212266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,8192,0.02160355614291297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,10240,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,3584,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,32,0.0018622221218215094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,16384,0.11051911115646362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,8192,0.007818666597207388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,12288,0.09536088837517633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,256,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,512,0.02723377777470483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,1024,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,2048,0.008473777936564552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,1024,0.014113777213626437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,8192,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,256,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,10240,0.03987200061480204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,5120,0.3840160104963515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,2048,0.06673333379957411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,5120,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,2048,0.04323200053638882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,2560,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,16384,0.09037866857316758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,7168,0.08010222514470418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,12288,0.06268533070882161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,3584,0.010103111465771994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,16384,0.02812533246146308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,256,0.017988445030318368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,3072,0.03677155574162801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,8192,0.02460977766248915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,32,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,6144,0.012682666381200155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,10240,0.03738311264250014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,768,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,10240,0.08313955863316853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,51200,0.20116355684068468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,8192,0.03160711129506429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,65536,0.2039146688249376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,64,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,2048,0.4360364543067084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,1536,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,1024,0.007349333001507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,51200,0.05277599891026815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,128,0.015871110889646742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,51200,0.11541422208150227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,768,0.02093155516518487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,5120,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,8192,0.011023999916182624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,512,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,256,0.012247111234400006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,1024,0.020287111401557922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,1024,0.01311644415060679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,768,0.017495999733606975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,3072,0.035205334424972534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,16384,0.08336000310050116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,3584,0.025366221865018208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,512,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,10240,0.0665102203687032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,1024,0.12192000283135308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,16384,0.15830311510297987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,1536,0.02408444384733836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,64,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,8192,0.023932443724738225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,512,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,6144,0.03494044476085239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,768,0.0041706665522522396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,16384,0.03454844488037957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,512,0.13468710581461588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,64,0.04530488782458835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,512,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,65536,0.07536888784832425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,3072,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,6144,0.09026844633950128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,768,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,1024,0.0063440000845326324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,65536,0.06405333015653822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,7168,0.04608444372812907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,51200,0.20951822068956164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,51200,0.2751351197560628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,6144,0.06767111354404025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,10240,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,3072,0.049792889091703624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,5120,0.014314666390419006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,1536,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,32,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,768,0.025176000263955858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,12288,0.10526044501198663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,1536,0.0125608891248703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,10240,1.3152231640285916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,16384,0.03020711077584161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,12288,0.04831377665201823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,16384,0.050550222396850586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,512,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,768,0.01570133368174235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,12288,0.004879111217127906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,768,0.005872000008821487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,7168,0.5257280137803819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,3584,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,256,0.004899555610285865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,5120,0.035242666800816856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,128,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,16384,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,12288,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,3072,0.8260560035705566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,7168,0.03111999895837572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,1024,0.024166221419970196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,16384,0.15142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,65536,0.1835928890440199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,3072,0.016678222351604037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,51200,0.05099822084108988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,7168,0.6255288653903538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,8192,0.05739733245637682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,768,0.017844445175594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,12288,0.0424124432934655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,4096,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,1024,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,2560,0.007034666836261749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,1536,0.02093244426780277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,4096,0.10770399702919854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,5120,0.014364444547229342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,8192,0.015439111325475903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,65536,0.6166115336947972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,7168,0.05568355321884155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,51200,0.14673599931928846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,768,0.006498666687144174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,3072,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,12288,0.3291751013861762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,3072,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,51200,0.31832178433736164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,64,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,3072,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,2560,0.05815644396675957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,2560,0.018188445104493033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,4096,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,3072,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,3072,0.08406755659315322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,512,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,256,0.009953777823183272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,7168,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,512,0.004095111042261124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,2560,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,51200,0.10298844178517659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,1024,0.0034577778230110803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,5120,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,128,0.023989333046807185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,65536,0.0809991094801161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,2048,0.11650044388241237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,1536,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,7168,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,51200,2.131321801079644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,3072,0.012229333321253458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,1024,0.005852444304360284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,51200,0.585977766248915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,1024,0.010558221903112201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,128,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,8192,0.019911110401153564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,3584,0.017224889662530687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,768,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,64,0.0022595555831988654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,4096,0.05375644564628601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,4096,0.010618666807810465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,256,0.02199911077817281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,3584,0.008642666869693333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,5120,0.08480088578330146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,6144,0.6044506496853298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,4096,0.008832000195980072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,4096,0.010658666491508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,8192,0.04073866539531284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,8192,0.2486035558912489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,1024,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,2048,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,12288,0.08041066593594022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,256,0.04349244303173489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,10240,0.0503715541627672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,768,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,2048,0.032857778999540545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,10240,0.2945751084221734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,10240,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,2560,0.06643644306394789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,6144,0.018915555543369718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,256,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,10240,0.01846844454606374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,4096,0.020430222153663635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,2560,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,8192,0.06131377485063341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,16384,0.1443262232674493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,5120,0.019504000743230183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,1536,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,128,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,32,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,3072,0.17354844676123726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,5120,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,2560,0.06235466400782267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,256,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,1024,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,1536,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,4096,0.028061333629820082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,10240,0.027460444304678176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,1024,0.009424000150627559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,32,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,10240,0.027365333504146997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,256,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,3072,0.018939556346999276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,10240,0.004607111215591431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,128,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,3072,0.040412445863087974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,64,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,128,0.007015110717879401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,7168,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,1024,0.2022871176401774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,10240,0.1426746712790595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,65536,0.31355733341640896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,16384,0.021399110555648804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,3072,0.01349422252840466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,1536,0.03716622127426995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,51200,0.05325599842601352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,6144,0.0427653325928582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,1536,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,256,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,3584,0.023665777511066858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,16384,0.4762061966790094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,10240,0.017804443836212158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,32,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,32,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,6144,0.01052177780204349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,7168,0.012540444731712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,2048,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,16384,0.23315021726820204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,5120,2.0181448194715714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,1024,0.03387199838956197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,512,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,768,0.023656888140572443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,2048,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,512,0.013224889006879596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,3072,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,65536,5.010320875379774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,3584,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,256,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,256,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,12288,0.02400622268517812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,32,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,16384,0.7984391318427192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,2048,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,2048,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,7168,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,1536,0.012888000243239932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,51200,0.7919182247585721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,3584,0.027049778236283198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,6144,0.030232889784706965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,256,0.01792355544037289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,7168,0.012858666479587555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,2048,0.016202666693263583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,3584,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,16384,0.05219733383920458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,256,0.0064897777305708984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,8192,0.055514666769239635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,12288,0.8777599864535861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,51200,0.0786693361070421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,7168,0.18782933553059897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,8192,0.0407680008146498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,12288,0.11878577868143718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,16384,0.018010665973027546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,3072,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,512,0.022638221581776936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,65536,0.5797893206278483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,10240,0.030135999123255413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,3584,0.026399999856948853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,16384,0.08949155277676052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,3072,0.015098666151364645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,256,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,1536,0.00813333359029558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,65536,2.0862178802490234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,3584,0.02642755541536543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,768,0.006130666782458623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,65536,0.06020977762010363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,1536,0.01626666718059116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,6144,0.15339199701944986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,7168,0.025506666964954797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,16384,0.23113155364990234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,128,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,2048,0.02208799951606327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,128,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,12288,0.04078311059210036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,12288,0.019558222757445443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,256,0.08830310901006062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,128,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,51200,0.13184089130825466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,7168,0.04305422306060791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,7168,0.03533333208825853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,8192,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,7168,0.009917333722114563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,65536,1.2908657921685112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,256,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,2048,0.03509777784347534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,6144,0.028146667612923518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,6144,0.10713155402077569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,2560,0.022355554832352534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,3072,0.1335688829421997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,3584,0.007843555675612556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,768,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,65536,0.38834844695197207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,2048,0.041145778364605375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,256,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,10240,0.029462221595976088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,4096,0.03357955482270982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,16384,0.06213511361016167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,10240,0.15086844232347277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,51200,0.04583911101023356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,1536,0.008934222161769867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,3072,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,3584,0.007780444290902879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,2048,0.01644711030854119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,4096,0.041850666205088295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,4096,0.021615111165576514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,1536,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,2048,0.007772444850868649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,512,0.16413954893747965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,256,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,1536,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,768,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,5120,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,5120,0.09790311257044475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,256,0.0053324442770746015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,2048,0.004868444469239977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,1024,0.07499377595053779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,128,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,768,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,2048,0.07679733302858141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,512,0.0034835553831524322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,128,0.0057911111248864075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,2048,0.03860622313287523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,7168,0.04956533180342781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,1536,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,5120,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,2560,0.017250667015711468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,128,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,768,0.029778665966457788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,256,0.015210666590266757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,10240,0.02532977859179179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,3584,0.007592000067234039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,6144,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,128,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,4096,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,128,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,64,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,10240,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,12288,0.04459822177886963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,51200,0.10642400052812363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,51200,0.857059531741672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,64,0.011346666349305047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,64,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,1024,0.006017777654859755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,8192,0.027405333187845018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,16384,0.04438044296370613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,3584,0.028368890285491943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,51200,3.691351996527778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,768,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,512,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,256,0.015504888362354703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,1536,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,7168,0.07788978020350139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,512,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,32,0.0021199999998013177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,10240,0.2902346716986762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,2048,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,1536,0.004093333250946469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,6144,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,512,0.0031351111829280853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,512,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,128,0.0053288886944452924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,7168,0.02178222272131178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,512,0.006151999864313338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,7168,0.04495288928349813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,12288,0.01661600006951226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,256,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,3072,0.1507591141594781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,3072,0.008184000021881526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,1024,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,16384,0.04765422145525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,4096,0.09102577633327907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,4096,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,12288,0.03363110952907138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,64,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,1536,0.012877333495352002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,2048,0.02385422256257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,1536,0.01603555513752831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,4096,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,16384,0.20696000258127847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,128,0.004933333231343163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,1024,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,256,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,512,0.012034667034943899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,8192,0.18729866875542533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,51200,0.5300889015197754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,10240,0.03949688871701559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,4096,0.160608000225491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,3584,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,51200,1.6791031095716689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,5120,0.04627822173966301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,256,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,3584,0.07129777802361383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,65536,0.17316088411543104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,3584,0.03115288747681512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,8192,0.007583111524581909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,5120,0.021485333641370136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,8192,0.01665422154797448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,16384,0.7608026928371854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,3584,0.11594843864440918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,256,0.10638133022520278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,2560,0.00850755555762185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,128,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,3584,0.049060444037119545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,7168,0.01756888959142897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,3584,0.022962666220135156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,2048,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,5120,0.021935110290845234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,2560,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,1536,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,5120,0.02479733361138238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,1024,0.2058124409781562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,65536,0.1297102239396837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,1024,0.008076444268226624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,51200,0.030844443374209937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,10240,0.17932355403900146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,51200,0.12317689259847005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,12288,0.36563732888963485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,16384,0.14001154899597168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,10240,0.03566933340496487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,3072,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,768,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,7168,0.04218666752179464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,6144,0.020619556307792664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,7168,0.033490667740503945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,4096,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,16384,0.04302933481004503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,2048,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,3584,0.04156088829040527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,1536,0.006882666713661618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,16384,0.2504257890913221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,2048,0.029727111260096233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,4096,0.030466665824254353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,8192,0.038650665018293597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,1024,0.028647998968760174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,12288,0.05939910809199015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,6144,0.013376889129479727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,4096,0.06855999761157565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,7168,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,512,0.006190222170617845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,10240,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,4096,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,3072,0.004256000121434529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,512,0.02117066747612423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,6144,0.010499555203649731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,65536,0.07807555463578966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,7168,0.021390222840838965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,1536,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,256,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,7168,0.012925333446926542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,7168,0.02108000053299798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,4096,0.1051866677072313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,1536,0.025303110480308533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,4096,0.03621244430541992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,6144,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,128,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,768,0.0064026667839950984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,1024,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,12288,0.013423110875818463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,8192,1.2314844131469727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,51200,0.14068355825212267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,256,0.017247110605239868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,2048,0.01367200000418557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,1536,0.02940888868437873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,1536,0.051966223451826304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,1536,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,768,0.012588444683286877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,128,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,10240,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,16384,0.024283554818895128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,3072,0.02584444483121236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,5120,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,1024,0.0483511123392317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,3584,0.027671999401516382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,3584,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,256,0.06826221942901611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,65536,0.06292444467544556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,8192,0.041441778341929116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,2560,0.15727021959092882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,1536,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,768,0.03576711151334975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,64,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,512,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,3072,0.020389333367347717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,3584,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,5120,0.037485334608289934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,1536,0.08365688721338908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,65536,0.13202222188313803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,256,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,64,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,12288,0.019882667395803664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,8192,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,3072,0.024286222126748826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,1536,0.06446755594677396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,768,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,16384,0.05439022183418274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,12288,0.11395111348893906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,1536,0.017808000246683758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,51200,1.3068435457017686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,3072,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,512,0.018166222506099276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,1536,0.018764444523387484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,16384,0.15189599990844727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,128,0.010116444693671333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,512,0.0032266666077905228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,1536,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,16384,0.4709102312723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,2048,0.029142220815022785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,16384,0.021035555336210463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,768,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,3072,0.401875548892551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,51200,0.0986586676703559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,2560,0.0756124456723531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,32,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,3072,0.02742133372359806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,4096,0.10402844349543254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,6144,0.009952888720565373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,12288,0.19230577680799696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,3072,0.05683466460969713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,7168,0.10284355613920425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,512,0.016410667035314772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,8192,0.006801777829726537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,2048,0.006690666493442323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,8192,0.14891289340125188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,7168,0.030817776918411255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,1024,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,1024,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,12288,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,3584,0.007807999849319458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,1024,0.01035644445154402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,2560,0.02094488839308421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,256,0.0031271111220121384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,128,0.00564177789621883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,768,0.02180533276663886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,512,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,1024,0.009309333231714036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,16384,0.15130933125813803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,5120,0.016373333003785875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,12288,0.07629777987798055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,51200,0.09148533476723565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,8192,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,2560,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,16384,0.01870844430393643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,10240,0.4934426413642035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,7168,0.02462133268515269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,1024,0.018004443910386827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,12288,0.38645243644714355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,128,0.012772444221708508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,8192,0.02455377744303809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,16384,0.24967379040188262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,3584,0.06476089027192858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,12288,0.013820444544156393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,51200,7.2161204020182295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,10240,0.08681155575646295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,6144,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,1536,0.1572728951772054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,8192,0.8369786474439832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,128,0.007153777612580194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,8192,0.02830933199988471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,3584,0.05289599961704678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,32,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,2560,0.014666666587193808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,256,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,1024,0.011341333389282227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,4096,0.3264124393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,3072,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,7168,0.04107644491725498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,65536,0.17209066285027397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,16384,0.04754222101635403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,5120,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,32,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,2560,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,8192,0.005624889085690181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,3584,0.020676443974177044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,3072,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,7168,0.34587823020087344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,65536,2.45670043097602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,4096,0.008550222549173567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,3584,0.07551466756396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,512,0.014859555496109856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,10240,0.04659733176231384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,4096,0.03329955538113912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,7168,0.006544888847404056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,512,0.03261955579121908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,6144,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,768,0.023581332630581323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,768,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,10240,0.017830222845077515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,51200,0.05099733339415657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,3584,0.08328800068961249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,256,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,512,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,12288,0.027614222632514104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,6144,0.019825778073734708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,8192,0.04283644424544441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,7168,0.08153777652316622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,4096,0.12912889321645102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,6144,0.03228088882234361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,16384,0.4908755620320638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,51200,0.08380711078643799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,8192,0.03047910994953579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,12288,0.06541066699557833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,256,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,1024,0.038112001286612615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,64,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,512,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,256,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,51200,0.09156533082326253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,16384,0.020238222347365487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,768,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,1536,0.023456000619464453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,128,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,512,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,64,0.0020088888704776764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,1024,0.032237334383858576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,65536,0.05957421991560194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,256,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,256,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,512,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,3072,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,51200,0.4064568943447537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,5120,0.017321777012613084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,6144,0.008453333543406593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,5120,0.028839111328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,768,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,5120,0.062012447251213916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,6144,0.09678755866156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,512,0.012487110992272695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,256,0.009264888862768808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,5120,0.36156000031365293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,65536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,7168,0.04408444298638237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,5120,0.03509599963823954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,256,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,16384,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,16384,0.747919135623508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,12288,0.03393955694304572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,1024,0.007040888898902469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,512,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,3072,0.05592799848980374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,1536,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,2048,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,64,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,3584,0.027447111076778833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,32,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,768,0.017679111825095285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,8192,0.01278222186697854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,5120,0.009756444229020013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,64,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,512,0.005415110952324337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,4096,0.007679111427730984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,65536,1.0045288933648002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,512,0.05098755492104424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,6144,0.02254666719171736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,1024,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,3584,0.014767110347747803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,2560,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,128,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,12288,0.195889777607388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,10240,0.4484586715698242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,32,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,65536,0.40235021379258895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,512,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,3584,0.0244835548930698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,16384,0.02826133370399475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,128,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,3584,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,3072,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,65536,0.01670488880740272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,5120,0.7127297719319662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,12288,0.04361600014898512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,1024,0.04679022232691447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,65536,0.5811528629726833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,10240,0.008519111408127679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,2560,0.028756442997190688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,64,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,3072,0.024900444679790076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,6144,0.14868888590070936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,1024,0.020346666375796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,65536,0.08779999944898818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,512,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,5120,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,512,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,10240,0.034280889564090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,65536,0.05490844448407491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,10240,0.323983112970988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,1024,0.004868444469239977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,12288,0.08134844568040636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,3072,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,128,0.02052088909678989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,5120,0.01421333352724711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,8192,0.023271110322740342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,3584,0.03423466616206699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,512,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,2560,0.012903999951150684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,10240,0.3298364480336507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,7168,0.010421333213647207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,2048,0.004387555436955558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,512,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,4096,0.008090666598743862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,3584,0.0628897746404012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,10240,0.7122622066073947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,768,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,1024,0.01851288808716668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,768,0.008159110943476358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,3584,0.008911111288600499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,2560,0.01645777788427141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,1024,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,2048,0.08522844314575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,1536,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,4096,0.010964444114102257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,2048,0.04962844318813748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,32,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,16384,0.03181511163711548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,768,0.0027884443600972495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,256,0.021565333008766174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,3584,0.031552000178231135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,7168,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,128,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,768,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,6144,0.026102221674389307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,768,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,5120,0.00905866672595342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,2048,0.026011554731263056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,6144,0.5659182336595323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,256,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,3584,0.03826311230659485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,128,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,7168,0.03207733233769735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,2048,0.047092444366878934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,3584,0.08867555856704712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,16384,0.8660906685723199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,1536,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,3584,0.020279111133681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,6144,0.01961866683430142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,10240,3.205990261501736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,16384,0.09494488769107395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,8192,0.014712888333532544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,65536,0.14064978228674993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,2560,0.08381511105431451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,768,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,16384,0.0457839998934004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,8192,0.024111110303137038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,3584,0.08128444353739421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,2048,0.011689778003427716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,768,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,65536,0.16707555452982584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,3584,0.011198221809334226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,256,0.028639998700883653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,2560,0.019492444064882066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,8192,0.03335466649797227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,256,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,32,0.04826844400829739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,51200,0.07666844129562378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,3072,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,1536,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,2048,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,3072,0.023745778534147475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,128,0.018147556318177115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,51200,0.276472012201945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,2048,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,2560,0.03171555532349481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,10240,0.09968711270226373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,8192,0.021270222134060327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,5120,0.09558044539557563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,2048,0.0684666633605957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,2048,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,256,0.019552888141738046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,128,0.003830222205983268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,512,0.014138667119873894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,3584,0.017111111018392775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,768,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,8192,0.03838933176464505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,51200,0.7865777545505099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,3072,0.017488888568348356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,768,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,16384,0.6984773212009006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,128,0.08885688914193048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,16384,0.10832622316148545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,16384,0.3331831031375461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,256,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,256,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,16384,0.11186844772762722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,12288,0.19473600387573242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,512,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,8192,0.06450666983922322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,51200,0.3340746561686198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,512,0.015572445260153877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,7168,0.016328889462682936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,2560,0.07694488763809204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,768,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,6144,0.027338667048348322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,2048,0.007053333024183909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,256,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,1024,0.0037173334923055437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,3584,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,2048,0.071743110815684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,6144,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,6144,0.05277599891026815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,32,0.008536888493431939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,6144,0.01574488812022739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,12288,0.02576711111598545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,7168,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,1536,0.027260444230503503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,512,0.025190222594473097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,2560,0.04288622074657016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,32,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,2560,0.007473777565691207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,5120,0.009704888694816166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,256,0.027598222096761067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,3584,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,16384,3.0792596605088978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,128,0.022796443767017786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,256,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,2560,0.18518133958180746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,64,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,128,0.021014221840434607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,10240,0.4335057735443115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,2048,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,10240,0.22998311784532335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,3072,0.03623466690381368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,5120,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,6144,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,768,0.027263111538357202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,51200,0.04342577854792277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,6144,0.017694221602545846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,1024,0.0034471110751231513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,7168,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,5120,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,256,0.006437333507670297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,10240,0.037823110818862915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,51200,0.047918223672442965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,12288,0.015856888559129503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,3584,0.008480889101823172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,5120,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,8192,0.02598222262329525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,51200,0.17250488864050972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,256,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,6144,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,4096,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,2560,0.028167112006081477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,512,0.006147555592987273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,16384,0.08025688595241971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,16384,0.5986142158508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,51200,0.9422124226888021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,65536,0.2943884531656901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,2048,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,2048,0.004940444396601783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,65536,0.1552693314022488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,5120,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,8192,0.04208799865510729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,3072,0.033569776349597506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,3072,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,3072,0.04168622361289131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,10240,0.11936622195773655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,10240,0.042803555727005005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,128,0.04071733355522156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,768,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,12288,0.12458666165669759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,12288,0.04254666633076138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,4096,0.17830399672190347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,128,0.014856888188256158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,10240,0.0688026679886712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,64,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,7168,0.14248177740308973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,5120,0.05956088834338718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,128,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,51200,0.049535112248526685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,16384,0.021873777111371357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,256,0.006341333190600078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,2048,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,5120,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,16384,0.04972711205482483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,1024,0.7570240232679578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,2560,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,3072,0.009395555489593083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,5120,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,1536,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,2560,0.006204444501135085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,7168,0.05134133497873942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,16384,0.06454488966200086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,6144,0.04669955703947279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,256,0.016619554824299283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,512,0.012336889074908363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,32,0.00453599997692638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,2048,0.00498488876554701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,1536,0.154003553920322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,8192,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,6144,0.04409511221779717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,768,0.023028444912698533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,65536,0.062223109934065074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,10240,0.3607395490010579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,2560,0.09320177634557088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,768,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,768,0.03646488984425863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,16384,0.06264444192250569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,10240,0.08836978011661106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,7168,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,65536,4.50648456149631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,3584,0.06666844420962863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,12288,0.1032186680369907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,2560,0.01016444464524587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,128,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,10240,0.03591466612286038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,5120,0.01810933318403032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,7168,0.009874666730562845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,16384,0.029613332615958318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,4096,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,10240,0.04351199997795952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,2560,0.04322133461634318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,3584,0.0395342210928599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,128,0.020121778051058452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,5120,0.10579733053843181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,10240,0.21736533111996123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,3072,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,3072,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,10240,0.03212799959712558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,51200,0.07018755541907416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,2560,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,65536,0.18601777818467882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,8192,0.15814933511945936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,2560,0.0469724436601003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,3584,0.056445333692762584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,256,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,10240,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,16384,0.03874489002757602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,4096,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,3072,0.007188444336255391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,12288,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,7168,0.03758755657407973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,6144,0.011143999795118967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,3584,0.21616355578104654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,65536,0.07790578073925443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,3072,0.2549911075168186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,10240,0.39026665687561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,1024,0.007344889144102733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,5120,0.05650844176610311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,256,0.00962488849957784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,7168,0.10375822252697414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,65536,0.19682133197784424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,256,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,65536,1.799477259318034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,10240,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,2048,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,1024,0.015581332974963717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,1024,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,65536,0.5959822336832682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,16384,0.049441777997546725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,4096,0.026550221774313185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,51200,12.221071031358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,512,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,4096,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,51200,0.07936266395780775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,128,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,10240,0.02808088726467556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,768,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,256,0.015174221661355762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,64,0.2703893449571398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,10240,0.2276853455437554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,16384,0.2120506631003486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,768,0.024882665938801233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,512,0.13446933693355984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,7168,0.06826755735609266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,6144,0.00998933364947637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,7168,0.18928000662061903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,2560,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,51200,0.12780443827311197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,51200,0.1096808910369873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,6144,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,65536,3.85933346218533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,6144,0.1425564421547784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,2560,0.04117066661516825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,1536,0.0321342216597663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,12288,0.11508177386389838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,1536,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,65536,0.022986667023764715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,2560,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,128,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,512,0.055120001236597695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,128,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,10240,0.018165333403481376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,51200,0.06735199689865112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,16384,0.01553600033124288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,2560,0.019334221879641216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,2560,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,10240,0.019357333580652874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,64,0.0074311114019817775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,16384,0.02551288902759552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,6144,0.12224711312188043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,768,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,12288,0.01817244456874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,32,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,256,0.16234578026665583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,128,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,3584,0.10731555355919732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,65536,0.14014221562279597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,8192,0.24663111898634169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,6144,0.029643555482228596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,10240,0.01628800067636702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,2560,0.0957991083463033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,32,0.0018053334206342697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,65536,1.2785057491726346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,32,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,12288,0.07976800203323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,512,0.04528977804713779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,1024,0.03261422117551168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,2048,0.05519644419352213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,3584,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,65536,0.32899022102355957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,2048,0.12738845083448622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,51200,0.2314000129699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,2560,0.022278222772810195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,51200,0.0862648884455363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,16384,0.02090311050415039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,2048,0.01590577761332194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,65536,0.10596177975336711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,2048,0.008725333544943068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,256,0.18190756109025744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,3584,0.03408533334732056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,16384,0.2028426726659139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,4096,0.032646222247017756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,3584,0.07541955841912164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,3072,0.016555555992656283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,64,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,128,0.02386222283045451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,3072,0.020023999942673575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,5120,0.008801778157552084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,128,0.01868711080816057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,1536,0.007364444434642792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,7168,0.0381368895371755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,12288,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,10240,0.02123822271823883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,16384,0.1435368855794271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,8192,0.16650844944847956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,4096,0.008031110796663497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,65536,0.06222222248713175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,16384,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,16384,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,5120,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,8192,0.019315555691719055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,512,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,12288,0.018233777748213876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,1024,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,7168,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,3072,0.6547270880805122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,128,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,1024,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,8192,0.01573600040541755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,51200,0.30754844347635907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,768,0.01941422290272183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,5120,0.058917330371008984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,4096,0.3692471186319987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,10240,2.502650578816732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,10240,0.017163554827372234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,10240,0.02199733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,7168,0.011804444094498953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,12288,0.1060640017191569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,12288,0.19927999708387586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,64,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,768,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,512,0.0038737778862317405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,51200,6.5006256103515625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,10240,0.031890667147106595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,4096,0.07603022125032213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,2560,0.0063440000845326324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,8192,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,8192,0.03942222065395779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,16384,0.03797066542837355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,1536,0.04067111015319824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,128,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,2560,0.23942576514350045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,10240,0.11631377538045247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,2560,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,256,0.008012444608741337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,16384,0.06207555532455444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,5120,0.02779200010829502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,32,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,6144,0.038896001047558255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,7168,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,3072,0.05472444494565328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,512,0.0028782222006056043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,3072,0.006695111178689533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,768,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,12288,0.11388088597191705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,2560,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,10240,0.06514400243759155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,51200,0.31600088543362087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,2048,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,3584,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,10240,0.023002665903833177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,8192,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,12288,0.8491324318779839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,3584,0.06095288859473335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,3072,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,12288,0.04494400156868828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,7168,0.019528888993793063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,16384,0.016555555992656283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,3584,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,16384,0.10326933198504978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,32,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,4096,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,3584,0.007491555478837755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,3584,0.69723023308648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,16384,0.03925333420435587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,4096,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,12288,0.24800178739759657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,128,0.021747555997636583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,2560,0.007816889219813878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,256,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,7168,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,65536,0.5492231051127116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,2560,0.11140355798933242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,128,0.00942666663063897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,5120,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,64,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,2048,0.06542311112085979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,128,0.015791111522250705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,5120,0.15810399585300022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,5120,0.008813333180215623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,8192,0.027765333652496338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,10240,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,32,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,2560,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,3072,0.009782222410043081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,65536,0.41051467259724933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,2048,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,8192,0.08156000243292914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,512,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,1536,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,8192,0.9368559519449869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,1536,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,64,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,7168,0.10604266987906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,1536,0.13579644097222224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,65536,0.1525084442562527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,1536,0.008410666551854875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,12288,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,2048,1.5649680031670465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,65536,2.077249738905165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,128,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,65536,21.456732855902775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,12288,0.018915555543369718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,16384,0.05835022528966268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,7168,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,8192,0.01276622215906779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,12288,0.3710542254977756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,768,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,256,0.01200177768866221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,4096,0.014210666219393412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,7168,0.03314844436115689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,64,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,32,0.00564444437623024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,4096,6.501766204833984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,6144,0.009529777699046666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,32,0.00996977753109402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,768,0.05759466356701321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,1024,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,1536,0.023843554986847773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,2560,0.03646755549642775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,1536,0.014853333433469137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,3584,0.012707555459605323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,768,0.13389421833886042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,12288,0.02306933369901445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,10240,0.036454220612843834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,7168,0.19233066505855986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,128,0.01902577777703603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,768,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,51200,0.40714044041103786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,7168,0.028029332558314007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,512,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,7168,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,3584,1.0472497940063477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,1536,0.04240444302558899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,1024,0.01647644407219357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,16384,0.05973066886266073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,6144,0.13397244612375894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,12288,0.028757333755493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,51200,0.04619733492533366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,2048,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,256,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,3072,0.007043555378913879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,3072,0.03246577911906772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,512,0.011371555427710215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,2048,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,8192,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,8192,0.15924710697597927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,3584,0.8925351036919488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,32,0.002550222186578645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,7168,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,256,0.00870400004916721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,2048,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,2560,0.01645955608950721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,8192,0.06807021962271796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,6144,0.05780444542566935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,32,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,8192,0.044863111442989774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,3072,0.014706666270891825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,51200,0.2744266721937391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,6144,0.04850488901138306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,1024,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,10240,0.027818666564093694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,51200,0.05972177452511258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,512,0.012787555654843649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,16384,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,1536,0.1506506601969401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,3072,0.08858578072653876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,768,0.030221333106358845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,12288,0.3901306788126628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,10240,0.29195555051167804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,7168,0.5041555298699273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,64,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,16384,0.021985777550273474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,4096,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,4096,0.028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,51200,0.8711137771606445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,65536,0.06634489032957289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,65536,0.08303822411431207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,10240,0.03162666824128892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,65536,0.22976356082492402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,4096,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,128,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,51200,0.6758871078491211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,256,0.05723377731111315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,512,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,128,0.020280889338917203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,4096,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,7168,0.030174222257402208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,2048,0.20043555895487467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,10240,0.07251377900441487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,10240,0.012595555848545499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,65536,3.4733857048882375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,32,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,6144,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,512,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,8192,0.11097777552074856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,2560,0.01718488832314809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,128,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,10240,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,10240,0.07404889000786676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,2048,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,16384,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,7168,0.3063528802659777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,1024,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,768,0.0058017778727743365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,51200,0.116848005188836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,12288,0.03464533223046197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,512,0.007159999675220913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,2048,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,32,0.009918221996890174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,5120,0.00851022203763326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,10240,0.10289777649773492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,16384,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,768,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,6144,0.10456266668107773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,51200,0.050550222396850586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,1536,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,1024,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,3072,0.011857777833938599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,12288,0.09548177984025742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,5120,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,12288,0.059608002503712974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,3584,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,512,0.004784000002675586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,1024,0.014039999908871122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,51200,0.6451422373453776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,51200,0.25619199540879994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,10240,0.02051377793153127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,6144,0.03503111004829407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,1024,0.018414222531848483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,4096,0.07698044512006971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,8192,0.12445955806308323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,512,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,2048,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,64,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,16384,0.2313244342803955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,4096,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,7168,0.00453599997692638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,256,0.030164443784289893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,7168,0.012806221842765808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,2048,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,12288,0.01816177699300978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,12288,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,12288,0.021930666433440313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,10240,0.06297866503397624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,5120,0.030235555436876085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,10240,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,6144,0.014948444234000312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,1024,0.03514666689766778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,64,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,256,0.009152000149091085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,6144,0.021243555678261652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,1024,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,16384,0.018256000346607633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,256,0.01789422167672051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,512,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,10240,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,768,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,51200,0.41986576716105145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,32,0.014894222219785055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,51200,1.0249546898735893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,12288,0.8168560134039985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,2048,0.040256887674331665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,2048,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,6144,0.04036444425582886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,6144,0.6591555807325575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,1536,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,32,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,6144,0.09838933414883083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,256,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,2048,0.007680888805124495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,3072,0.05374755462010702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,5120,0.025967111190160114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,4096,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,32,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,8192,0.029508444997999404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,1536,0.06490933232837252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,8192,2.3980541229248047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,51200,0.1773688925637139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,7168,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,64,0.0070364442136552595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,6144,0.03386844529045953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,10240,0.028981334633297388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,4096,0.10497155454423691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,768,0.017310221989949543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,2048,0.03124177787039015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,7168,0.014123555686738757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,1024,0.004024888906213972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,4096,0.026350221700138513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,768,0.004477333277463913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,512,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,32,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,51200,0.12515021695031062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,64,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,32,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,768,0.020405333903100755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,7168,0.026087111896938745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,4096,0.07146755854288737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,7168,0.0122079998254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,16384,0.013719999955760108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,128,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,3584,0.08226755592558119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,10240,0.1202204492357042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,1024,0.12782222694820827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,1536,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,512,0.017087111870447796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,4096,0.0086986662613021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,8192,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,128,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,10240,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,4096,0.017661333084106445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,8192,0.08644177516301473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,3584,0.008845333423879411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,51200,1.1983475155300565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,10240,0.03299377693070306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,16384,0.02179199953873952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,10240,0.2176897790696886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,10240,0.023344889283180237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,128,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,2560,0.08079822195900811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,10240,0.012273777690198688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,3584,0.009771555662155151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,2560,0.39682843950059676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,6144,0.2974453237321642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,8192,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,128,0.008915555146005418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,2560,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,10240,0.031529777579837374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,2560,0.00489688871635331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,12288,0.06464711162779066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,64,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,2048,0.01757688820362091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,3584,0.12090222040812175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,512,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,5120,0.02979555394914415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,65536,7.759311252170139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,8192,3.5302808549669056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,2048,0.02676444417900509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,16384,0.8484133084615072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,8192,0.08346311251322429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,64,0.0017866666118303935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,1536,0.029616001579496596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,128,0.024030221833123103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,6144,0.04486844605869717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,3072,0.038330667548709445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,3072,0.16488088501824275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,12288,0.029271999994913738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,10240,0.025810667210155066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,1536,0.031317333380381264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,7168,0.027152000202072993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,2560,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,10240,0.033751110235850014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,768,0.018378666705555387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,7168,0.032050665881898664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,1536,0.02995377779006958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,32,0.06811200247870551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,512,0.013890667094124688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,2560,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,2560,0.017893332574102614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,51200,1.2341537475585938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,768,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,64,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,65536,0.06047022342681885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,64,0.004931555440028508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,5120,0.035374220874574445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,7168,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,2560,0.37147021293640137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,64,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,16384,0.11532266934712727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,51200,1.9988142649332683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,5120,0.07475466860665216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,10240,0.2782257662879096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,128,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,5120,0.010287111004193624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,128,0.03883822096718682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,12288,0.018406222263971966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,128,0.03025777803526984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,65536,0.07721155881881714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,16384,0.029116445117526587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,1536,0.0053013331360287136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,1024,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,512,0.00425866660144594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,256,0.02516533268822564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,512,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,2048,0.00666933341158761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,3072,0.007876444194051955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,8192,0.04394133223427666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,4096,0.01538133290078905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,2560,0.025060445070266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,51200,0.06999199920230441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,3584,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,1024,0.015871110889646742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,2560,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,3072,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,7168,0.012000888586044312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,51200,0.1539306640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,7168,0.42412355211046004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,512,0.037110222710503474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,768,0.014440889159838358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,512,0.012872888810104795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,7168,0.022375999225510493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,7168,0.04787733488612705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,1536,0.22242667939927843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,1536,0.007721777591440413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,3584,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,512,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,8192,0.03036622206370036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,1024,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,3584,0.04335022303793165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,3584,0.06674489047792223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,7168,0.1538497739368015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,1024,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,2048,0.03382488754060533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,6144,0.03554133243030972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,16384,0.0199635558658176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,4096,0.11227822303771973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,128,0.0243857784403695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,3072,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,51200,0.12465777662065293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,1536,0.0181377778450648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,12288,0.11340444617801243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,6144,0.022302221920755174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,6144,0.019559999306996662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,1024,0.0064533332155810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,2048,0.0041804446114434134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,4096,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,64,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,65536,0.5436515808105469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,3584,0.017161778277821012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,5120,0.013245333400037555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,8192,0.19852977328830293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,65536,0.7519466612074109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,256,0.022648889157507155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,512,0.012773333324326409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,4096,0.04231466518508064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,8192,0.44374932183159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,12288,0.04455733299255371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,16384,0.14178133010864258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,51200,0.9019680023193359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,3072,0.01578044394652049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,64,0.0027591110103660156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,51200,0.055032001601325146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,768,0.019835554891162448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,51200,0.06300710969501071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,5120,0.04930666751331753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,512,0.0503404438495636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,64,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,1536,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,2560,0.004210666649871402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,6144,0.05881244606441922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,512,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,512,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,256,0.024028443627887305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,256,0.010594666832023196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,65536,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,10240,0.2536088890499539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,12288,0.013298666311634911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,64,0.08298844761318631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,128,0.01995111174053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,512,0.11029689179526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,10240,0.10203911198510064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,512,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,10240,0.7590053346421984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,51200,0.5872809092203776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,65536,0.6416906780666775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,4096,0.04797066582573784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,768,0.043888890080981784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,5120,0.02182577715979682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,128,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,3584,0.462158203125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,4096,0.19897154966990152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,256,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,7168,0.019909333851602342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,7168,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,8192,0.4704355663723416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,51200,2.241546630859375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,12288,0.0199297782447603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,128,0.021943110558721755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,16384,0.07360888852013482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,768,0.0910337766011556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,5120,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,6144,0.013175110850069257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,16384,0.03279822071393331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,3072,0.09917600287331475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,3072,0.18341688315073648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,5120,0.1792257759306166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,2560,0.007078222102589077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,128,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,128,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,3584,0.01292799992693795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,512,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,6144,0.013176889055305056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,51200,0.23652532365587023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,768,0.004506666627195147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,65536,0.06875999768575032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,10240,0.2010666661792331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,768,0.023359111613697473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,32,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,65536,0.05713511175579495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,128,0.0711688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,5120,0.10763111379411484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,7168,0.010738666686746808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,2560,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,512,0.016201777590645682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,12288,0.024476443727811176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,1024,0.2662088871002197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,256,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,4096,0.021330666210916307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,2048,0.015259555644459195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,64,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,10240,0.02817866537306044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,1024,0.016586666305859882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,3584,0.21243377526601157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,4096,0.14845421579149035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,1024,0.02769511110252804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,2560,0.0471955570909712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,16384,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,65536,0.6438942485385472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,5120,0.08847466442320083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,2048,0.40465688705444336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,6144,0.06549777587254842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,51200,0.06740355491638184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,512,0.05201333430078295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,3584,0.019833778341611225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,10240,0.42941421932644314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,128,0.011343999869293638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,4096,0.5602542029486762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,65536,0.06951289044486152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,2048,0.44802400800916886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,1536,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,51200,0.28938576910230845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,6144,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,1024,0.0030097777230872046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,512,0.030576000610987347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,4096,0.14860888322194418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,8192,0.1301057736078898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,16384,4.234881083170573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,7168,0.11726221773359512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,65536,0.03968088825543722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,6144,0.022261333134439256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,2048,0.011767110890812345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,2048,0.006734222173690796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,2048,0.012482666307025485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,10240,0.028974221812354192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,128,0.07065688901477389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,512,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,65536,1.483481725056966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,256,0.06507822540071276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,64,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,16384,0.055784887737698026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,16384,0.058245334360334605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,51200,0.4382471243540446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,1536,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,2560,0.15075377623240152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,3072,0.022474666436513264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,512,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,3072,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,128,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,10240,0.016201777590645682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,128,0.02436888880199856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,4096,0.0927857756614685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,7168,0.023425777753194172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,16384,0.11001866393619114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,3584,0.017934223016103108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,8192,0.10828089051776463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,8192,0.04836800032191806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,16384,0.021111999948819477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,768,0.00811288919713762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,768,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,3584,0.012910222013791403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,5120,0.012323555847009024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,3072,0.11810755729675293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,1536,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,2048,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,16384,0.07827999856736925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,1536,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,65536,0.06265511115392049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,1536,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,768,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,3584,0.024795555406146582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,12288,0.017310221989949543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,7168,0.020896888441509668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,51200,0.6014284557766384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,3584,0.0837271081076728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,2560,0.017695999807781644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,5120,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,768,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,3584,0.09389511081907485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,12288,0.0420702232254876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,1024,0.008907555705971187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,12288,0.03068888849682278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,7168,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,1024,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,8192,0.022091555926534865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,6144,0.02793777651256985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,2048,0.007103111180994246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,32,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,1536,0.011010666688283285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,7168,0.025462221768167283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,768,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,1024,0.007816889219813878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,6144,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,12288,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,32,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,32,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,3584,0.005324444423119227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,10240,0.03159822358025445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,32,0.015198222464985318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,2048,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,8192,0.1461093293295966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,3584,0.03351377778583103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,2560,0.01426488823360867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,5120,0.008799999952316284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,512,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,6144,0.10640088717142741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,8192,1.9354578653971355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,16384,1.3843297958374023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,5120,0.020928889513015747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,512,0.005731555736727185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,12288,0.07547022236718072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,2560,0.008178667061858708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,32,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,65536,0.07652888695398967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,768,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,3584,0.3768577840593126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,51200,0.839487976498074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,3584,0.034224887688954674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,64,0.0021973333011070886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,8192,0.013915555344687568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,512,0.006360000206364526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,10240,0.03271555569436815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,1024,0.045624888605541654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,51200,0.23262400097317168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,1536,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,1536,0.023736889163653057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,2048,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,65536,0.06405866808361478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,3584,0.042230221960279674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,2048,0.2467342217763265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,3072,0.007327111230956183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,128,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,2560,0.022632888621754114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,128,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,768,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,128,0.012785777449607849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,3072,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,16384,0.41355111863878036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,8192,0.020058666666348774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,8192,0.048716445763905845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,512,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,1536,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,2048,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,7168,0.2360880109998915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,6144,0.03709866603215536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,10240,0.032925334241655134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,7168,0.01943822205066681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,1024,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,3072,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,1024,0.0034906665484110513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,1536,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,10240,0.04303377866744995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,51200,0.09208799733055963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,51200,0.46425241894192165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,3072,0.01866044435236189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,2048,0.03658755620320638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,12288,2.9752000172932944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,32,0.005433777968088786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,7168,0.02737599942419264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,7168,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,3072,0.23373778661092123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,6144,0.08284355534447564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,768,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,3584,0.015623110863897534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,256,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,7168,0.03583200110329522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,3072,0.018605333235528734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,1536,0.09903644190894233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,5120,0.020920889245139226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,1536,0.19105422496795654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,3072,0.0860533316930135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,768,0.017312000195185345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,8192,0.023413333627912734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,10240,0.1719519959555732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,5120,0.021958221991856892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,1024,0.050607111718919545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,512,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,5120,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,12288,0.017295999659432303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,51200,0.09630578094058567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,3072,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,2560,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,7168,0.023095111052195232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,7168,0.22446934382120767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,5120,0.028950221008724634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,10240,0.07110577821731567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,12288,0.05095555716090732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,512,0.01351999988158544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,4096,0.009886222581068674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,8192,0.08133066362804837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,64,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,768,0.0029306666304667792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,32,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,256,0.0021199999998013177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,3584,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,1536,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,10240,0.04075288772583008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,3072,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,768,0.017870222528775532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,6144,0.02534488836924235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,128,0.006066666709052191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,4096,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,7168,0.04659733176231384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,3072,0.018562666243977018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,256,0.008368000388145447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,12288,0.005384888913896348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,51200,0.060356444782680936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,1536,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,768,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,7168,0.04502755403518677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,32,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,1024,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,128,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,4096,0.01718488832314809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,10240,0.013357333011097379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,65536,0.07710133658515082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,4096,0.11917866600884332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,12288,0.04798044429885017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,2560,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,16384,0.05722666449016995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,16384,0.04050666756100125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,5120,0.00925866680012809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,768,0.011942221886581846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,7168,12.899530198838976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,1536,0.021963554951879714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,2048,0.00778399987353219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,2560,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,1024,0.013013333082199097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,51200,0.11636622746785481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,256,0.02236533330546485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,768,0.017678222722477384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,8192,0.049674666590160794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,256,0.011117333339320289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,4096,0.029248889949586656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,65536,0.1519626643922594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,256,0.09290399816301133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,7168,1.240579605102539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,1024,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,8192,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,1024,0.011558222273985544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,1536,0.04279288980695936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,2560,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,7168,0.11450133058759902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,65536,0.26542843712700737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,5120,0.09249689181645711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,128,0.009760888914267222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,16384,0.021457778082953557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,4096,0.008083555433485243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,512,0.01535288823975457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,12288,1.0350720087687175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,256,0.021427555216683283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,512,0.05515377720197042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,3072,0.01347555551264021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,16384,0.36978665987650555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,2048,0.015876443849669564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,5120,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,5120,0.08204355504777697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,1536,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,8192,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,6144,0.18616978327433267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,64,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,16384,0.22404000494215223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,16384,0.12447467115190293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,768,0.021448888712459143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,2048,0.01310222182008955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,768,0.026144888665941026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,1024,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,65536,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,7168,0.02469066613250309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,3072,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,1024,0.02299022177855174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,1536,0.008524444368150499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,3072,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,4096,0.02642755541536543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,4096,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,65536,1.2436710993448894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,1536,0.019461333751678467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,3072,0.03536533315976461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,65536,0.9815422693888346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,51200,0.14920622772640652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,51200,2.3047262827555337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,5120,0.03831111060248481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,51200,0.13992088370853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,128,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,12288,0.01829155617290073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,16384,0.9867661794026693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,5120,0.009932444327407414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,1024,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,4096,0.09137511253356934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,51200,0.11309422387017144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,51200,0.06294577651553683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,6144,0.009151111046473185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,768,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,1536,0.006089777582221561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,4096,0.02093244426780277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,4096,0.5119351281060113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,16384,0.02323377794689602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,2048,0.020583111378881667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,16384,0.01093155559566286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,128,0.025302223033375208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,512,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,16384,0.023288889063729182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,51200,0.5351849132113987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,16384,0.08168355623881023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,1536,0.004551110996140374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,7168,0.35005421108669704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,256,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,5120,0.011381333072980246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,10240,0.029719998439153034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,10240,0.014528888795110913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,8192,0.2798808945549859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,1024,0.011971555650234222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,1024,0.0525857773092058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,128,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,6144,1.115089734395345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,3072,0.3972497781117757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,256,0.022336888644430373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,512,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,6144,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,256,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,4096,1.2534915076361763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,10240,0.044459554884168834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,12288,0.21527822812398276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,768,0.05527377790874905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,65536,0.1489324437247382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,3072,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,12288,0.01849422189924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,2560,0.014035556051466199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,3072,0.007702222300900354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,10240,0.0188711103465822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,10240,0.03165955675972833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,2048,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,64,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,2048,0.0251920007997089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,768,0.007240888973077138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,3072,0.016207999653286405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,1024,0.032241778241263494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,256,0.006351999938488007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,5120,0.0063440000845326324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,51200,0.07087466451856825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,5120,0.12556533018747965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,4096,0.02384000023206075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,32,0.005698666804366642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,12288,0.01457155578666263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,6144,0.028520888752407495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,3584,0.0768337779574924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,16384,0.14542578326331243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,3072,0.014092443717850579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,2560,0.004987555659479565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,2048,0.03846044341723124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,1536,0.01963200006220076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,1536,0.032778667079077825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,256,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,3584,0.012469333079126147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,128,0.00655644428398874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,2048,0.04256177941958109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,32,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,256,0.020709334148301017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,4096,0.004256000121434529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,3072,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,65536,0.569992012447781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,512,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,3072,0.07731111182106866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,768,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,12288,0.20379911528693306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,512,0.019573332534896005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,12288,0.019739554988013375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,51200,0.10390400224261814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,6144,0.027296889159414504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,10240,0.031488888793521456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,7168,0.02366133365366194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,128,0.001967111188504431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,51200,0.03545866741074456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,16384,0.0670293304655287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,2560,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,256,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,256,0.005349333501524395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,5120,0.09178933170106675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,32,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,1024,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,3072,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,51200,1.8924176957872179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,3072,0.009295110901196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,1024,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,2560,0.031160887744691636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,2560,0.02057422200838725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,8192,0.012452444268597497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,128,0.020434666011068556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,64,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,256,0.018601778480741713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,1024,0.026894221703211468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,65536,0.14638755056593153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,2560,0.017120889491505094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,512,0.024502222736676533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,1536,0.02446577780776554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,51200,0.4019004503885905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,12288,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,12288,0.3058399889204237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,256,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,32,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,512,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,6144,0.01569688816865285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,7168,0.004594666676388847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,1024,0.019480889042218525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,3584,0.014281777871979607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,5120,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,1024,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,4096,0.056448002656300865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,65536,1.0953262117173936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,2560,0.056417776478661426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,3072,0.06641599867078993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,2048,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,2560,0.037252446015675865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,6144,0.2395413451724582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,51200,0.13057511382632786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,768,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,512,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,3072,0.02603200078010559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,4096,0.01812977757718828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,256,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,2048,0.0295724438296424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,2560,0.021145777569876775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,5120,0.0814337796635098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,65536,0.06247644291983711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,51200,0.13414488898383245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,2560,0.029270221789677937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,2048,0.026037333740128413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,7168,0.08044888575871785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,128,0.02010311186313629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,6144,0.05401600069469876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,4096,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,5120,2.1361431545681424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,1536,0.007490666376219855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,51200,0.12292710940043132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,16384,0.06547644403245714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,65536,0.4496622085571289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,7168,0.08300977945327759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,1536,0.006800000038411882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,768,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,4096,0.010979555547237396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,2048,0.02760533326201969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,65536,0.23675555653042266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,2048,0.02056799994574653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,1024,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,6144,0.019127999742825825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,1024,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,8192,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,16384,0.03847288754251268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,6144,0.1633519993888007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,4096,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,6144,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,5120,0.021885333789719477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,4096,0.1372986634572347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,12288,0.14478666252560085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,5120,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,1024,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,1024,0.04688266581959195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,768,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,4096,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,51200,0.32011644045511883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,51200,0.15290399392445883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,1536,0.04217955470085144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,1024,0.0035439998739295532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,3584,0.03505955470932855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,64,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,1024,0.05805777841144138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,3584,0.010577778021494547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,51200,0.06865244441562228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,6144,0.025082666012975905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,16384,0.22257865799797905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,1536,0.009249777429633671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,7168,0.030155556069480047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,256,0.013631111217869652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,7168,0.036483556032180786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,768,0.00757244477669398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,2560,0.10344622532526652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,7168,0.01976799964904785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,32,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,128,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,6144,0.03685599896642897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,4096,0.0086986662613021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,12288,0.022348443667093914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,2560,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,2048,0.0064106666379504735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,51200,0.015251555376582675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,768,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,51200,0.052863111098607383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,128,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,1536,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,16384,0.11329244242774116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,5120,0.018424888451894123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,16384,0.7901324696011014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,5120,0.0881991121504042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,12288,0.0535662207338545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,12288,0.0901048911942376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,32,0.019627556204795837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,3584,0.015184889237085978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,768,0.015450666348139444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,51200,0.2608666684892443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,10240,0.053231110175450645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,1024,0.009485333330101436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,65536,0.14482755131191677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,32,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,16384,0.11552355024549697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,51200,0.05162133442031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,8192,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,7168,0.012803555362754397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,2048,0.07280177540249295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,12288,0.062414222293429904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,4096,0.030692444907294378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,4096,0.047374222013685435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,512,0.056461334228515625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,12288,0.09202488925721909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,1024,0.022679110368092854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,12288,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,6144,0.25863999790615505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,16384,0.04624355501598782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,8192,0.0292195545302497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,128,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,2560,0.02131555477778117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,16384,1.2286835776435006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,8192,0.018276444739765592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,4096,0.041611555549833514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,8192,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,10240,0.012113778127564324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,8192,0.008719999757077958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,3072,0.0613937775293986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,65536,0.08787910805808173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,5120,0.03171555532349481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,1536,0.013632888595263163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,10240,0.0354053344991472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,1024,0.007355555891990662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,65536,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,256,0.020661332541041903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,128,0.006657777975002925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,6144,0.06243289179272122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,6144,0.08726577626334296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,6144,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,3072,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,2048,0.13965511322021484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,5120,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,768,0.025096888343493145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,16384,0.16889511214362252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,512,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,5120,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,8192,0.015404444601800708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,65536,0.10498222377565171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,2048,0.007568000091446771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,10240,0.051923556460274585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,1024,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,1024,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,128,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,8192,0.03870044483078851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,7168,0.039800001515282526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,12288,0.43137865596347386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,16384,0.11271911197238499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,3072,0.04438755578464932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,2560,0.06740178002251519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,12288,0.01961777773168352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,256,0.11917777856190999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,64,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,8192,0.021684444612926904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,10240,1.9164729648166234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,7168,0.028360890017615423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,65536,0.1010133359167311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,128,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,3584,0.03704977697796292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,16384,0.0336480008231269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,1536,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,7168,0.061012446880340576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,4096,0.017264000243610807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,256,0.0158942225906584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,8192,0.04873422119352552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,12288,0.055191109577814736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,1024,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,1024,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,12288,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,1536,0.5766960250006782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,64,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,3584,0.04316533274120755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,512,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,3072,0.007110222346252865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,51200,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,8192,0.1788346634970771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,3584,0.03590666585498386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,1024,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,3072,0.009447111023796929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,3072,0.04155733187993368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,7168,0.26544533835517037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,2048,0.01759822169939677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,768,0.015001777145597668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,2048,0.040672888358434044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,2560,0.027848889430363972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,3584,0.027831110689375136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,1536,0.014697778556081982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,7168,0.0799822211265564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,7168,0.009728888670603434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,128,0.06856622298558553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,2048,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,5120,0.030181331766976252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,65536,0.05350666575961643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,5120,0.02753066685464647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,16384,0.03059911065631443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,10240,0.04303200046221415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,128,0.015392889579137167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,3584,0.03488889005449083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,10240,0.697689798143175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,4096,0.027462222509913977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,10240,0.06776977909935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,12288,2.476015090942383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,1024,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,256,0.10228444470299615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,2560,0.012560000022252401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,6144,0.01401511165830824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,12288,0.4467217657301161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,256,0.008600888980759515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,65536,0.06637777884801228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,2048,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,2560,0.009507555928495195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,512,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,12288,0.8022506501939562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,10240,0.01564711166752709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,4096,0.1374382177988688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,256,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,1536,0.025085333320829604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,128,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,3072,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,5120,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,5120,0.01534844438234965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,1536,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,512,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,3072,0.015954666667514376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,7168,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,10240,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,4096,0.0639680027961731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,1536,0.0047608887155850725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,1024,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,2560,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,16384,0.04338755541377597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,768,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,5120,0.03390489021937052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,3072,0.025964443882306416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,3072,0.04864711231655545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,1024,0.04434222314092848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,7168,0.038359112209743924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,7168,0.037073777781592474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,256,0.03625955515437656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,16384,0.038655112187067665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,1024,0.03741777936617533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,65536,1.096138636271159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,65536,0.07359821928872003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,16384,0.01888088881969452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,8192,0.03401155604256524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,16384,0.05954222546683418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,7168,0.07814577553007337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,1536,0.04731022318204244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,12288,0.046875556310017906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,51200,0.17665955755445692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,32,0.001997333433892992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,32,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,2048,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,4096,0.00869688888390859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,8192,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,4096,0.015885333220163982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,2048,0.058796445528666176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,3072,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,1536,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,16384,0.1433911058637831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,64,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,512,0.017229333519935608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,256,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,8192,0.013391111459996967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,10240,0.30770667394002277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,51200,0.11101333300272624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,3584,0.3579537868499756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,10240,0.03631733192337884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,65536,1.25544982486301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,1024,0.02095199955834283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,256,0.013336000343163809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,512,0.0033297776761982176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,5120,0.05979377693600125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,768,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,256,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,8192,0.04774844315316942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,1536,0.0064791109826829694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,2560,0.1581208838356866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,3072,0.02016444504261017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,128,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,32,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,512,0.017257778180970084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,6144,0.050500445895724826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,7168,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,7168,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,4096,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,768,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,64,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,7168,0.1211902168061998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,256,0.12592266665564641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,10240,0.3463475439283583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,12288,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,5120,0.06053600046369764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,3584,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,6144,0.009126221968068017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,65536,0.1386559936735365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,3584,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,64,0.017070222232076857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,1024,0.018914666440751817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,1536,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,6144,0.05364355444908142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,12288,0.015340444114473132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,5120,0.04856977860132853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,2048,0.005072000126043956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,2048,0.031601776679356895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,10240,0.025246222813924152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,2048,0.07781689034567939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,1536,0.03260711166593764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,256,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,2048,0.004670222186379963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,7168,0.03647199935383267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,8192,0.011117333339320289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,2048,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,3072,0.05027733246485392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,6144,0.01330488920211792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,10240,0.25403555234273273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,6144,0.08265688684251574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,10240,0.23778843879699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,65536,1.1852266523573134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,7168,0.06006933583153618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,768,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,4096,0.42766933970981175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,512,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,2560,0.093158225218455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,7168,0.015576889117558798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,2048,0.035585777627097234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,128,0.02835644284884135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,10240,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,51200,0.05234577920701769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,16384,0.41201334529452854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,1024,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,7168,0.004242666479614046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,12288,0.20760977268218994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,64,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,1024,0.0413333343134986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,3584,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,10240,0.046113779147466026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,32,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,10240,0.046776887443330556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,2048,0.026597332623269822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,3072,0.011146666275130378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,512,0.022838221655951604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,12288,0.00775822252035141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,65536,0.7679804695977105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,16384,0.10322488678826226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,2560,0.02454400062561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,5120,0.0678888890478346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,3584,0.15504444970024958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,5120,0.022502221994929846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,3072,0.040566222535239324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,512,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,768,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,256,0.02502222193611993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,8192,0.033315555916892156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,2560,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,16384,0.026766222384240892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,4096,0.02367466688156128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,512,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,3072,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,1024,0.13314933247036403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,3584,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,256,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,1024,0.009859555297427708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,128,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,12288,0.13831733332739934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,2048,0.004568888909286923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,3584,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,51200,0.3300951057010227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,2048,0.04991199904017978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,10240,0.29545778698391384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,65536,0.16119466887580022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,6144,0.04731466703944736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,512,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,1536,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,8192,0.03135733471976386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,3584,0.007239110767841339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,128,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,8192,0.1781680054134793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,2560,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,4096,0.05714222457673815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,8192,0.03973689013057285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,7168,0.107859558529324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,1536,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,3584,0.019871110717455547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,10240,0.012420444852775998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,1536,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,512,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,12288,0.06191111273235745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,128,0.009887110855844285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,3072,0.04397155510054695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,3584,0.007671111159854465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,3584,0.029249777396519978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,2560,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,3584,0.013367111484209696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,3072,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,12288,0.013385777672131857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,1536,0.007698666718271043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,16384,0.01758577757411533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,6144,0.019323555959595572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,51200,0.20097955067952475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,3584,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,2048,0.018575999471876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,10240,0.01665422154797448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,256,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,12288,0.03510311245918274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,512,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,65536,0.41017511155870223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,3072,0.007754666937722101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,3072,0.046799110041724316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,128,0.01054933336046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,7168,0.14003911283281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,12288,0.017733333839310538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,64,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,12288,0.056779556804233126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,6144,0.020505777663654752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,2560,0.3993519941965739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,10240,0.01219555570019616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,8192,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,65536,0.08312533299128215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,2560,0.01812177730931176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,10240,0.08737422360314263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,51200,1.2566177580091689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,7168,1.7450328403049047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,7168,0.08366222514046563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,8192,0.5423075358072916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,768,0.026782222919993933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,2048,0.04288977715704176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,512,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,512,0.017457778255144756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,16384,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,5120,0.026749332745869953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,10240,0.01262399968173769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,51200,0.04014222158326043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,256,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,65536,0.04055733482042948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,8192,0.14128177695804173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,65536,0.16469866699642605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,3584,0.046783109505971275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,7168,0.019854222734769184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,256,0.00686488880051507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,16384,0.042929778496424355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,768,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,6144,0.012549333274364471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,65536,1.018664042154948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,6144,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,4096,0.11929155720604791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,1536,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,2048,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,4096,1.2865173551771376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,256,0.03244533472590976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,64,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,4096,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,2048,0.019369777705934312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,16384,1.5653902689615886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,10240,0.03352800011634827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,3584,0.038578665918774076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,51200,0.14649777942233616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,3584,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,7168,0.1649057732688056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,2560,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,5120,0.009783999787436591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,3072,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,65536,0.0908764468299018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,768,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,7168,0.03496799866358439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,51200,0.4105973243713379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,51200,0.17127645015716553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,768,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,2048,0.022044445077578228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,3584,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,7168,0.08262666728761461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,65536,33.71307712131077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,6144,0.03365777929623922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,3072,0.05500266618198819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,51200,0.13113333119286433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,12288,0.02257866660753886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,5120,0.02741333345572154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,6144,0.10744355784522162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,6144,1.0047857496473525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,2048,0.009806222385830348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,1536,0.017785777648289997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,512,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,1024,0.008739555875460306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,3072,0.09686044189665054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,7168,0.010767999622556897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,3584,0.026570666167471144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,3584,0.04368533359633552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,2560,0.03476622369554307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,51200,0.027662222584088642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,10240,0.017877333694034152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,5120,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,512,0.0029004443850782183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,1024,0.006512889017661412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,8192,0.09219911363389756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,16384,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,5120,0.5674693319532607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,256,0.0026373333401150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,5120,0.02861600120862325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,16384,0.11269156138102214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,16384,2.573301315307617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,65536,0.7064773241678873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,1024,0.00655644428398874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,3584,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,7168,0.013927110367351107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,2560,0.0672888888253106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,3072,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,3072,0.02328088879585266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,3584,0.012175111307038201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,768,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,2048,0.019256888164414298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,128,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,256,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,65536,0.44863377677069766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,32,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,512,0.011736888852384357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,3584,0.1132426659266154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,256,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,768,0.012448888685968188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,7168,0.010324444207880232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,768,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,1024,0.047338667843076915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,2560,0.012263110942310758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,51200,0.038014223178227745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,512,0.014127110441525778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,6144,0.021505778034528095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,16384,0.05943999687830607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,7168,0.1344346602757772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,3072,0.14371021588643393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,4096,0.01681866745154063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,2560,0.01773599949147966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,12288,0.06064266628689236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,1024,0.020601777566803824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,512,0.177802668677436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,51200,0.6082604196336534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,16384,0.03966044386227926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,2560,0.012272888587580787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,8192,0.02555199960867564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,10240,0.034400890270868935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,128,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,3584,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,8192,0.02586666742960612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,10240,0.011870221959220039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,8192,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,51200,0.15985333919525146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,65536,1.9540568457709417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,51200,0.3129342132144504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,8192,0.030378666188981798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,8192,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,12288,0.08952533536487156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,64,0.012148444851239523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,768,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,4096,0.03926577832963731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,1024,0.017924444542990792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,51200,1.8533653683132594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,4096,0.050550222396850586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,512,0.0028773333049482773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,10240,0.0816497802734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,2560,0.02091377807988061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,6144,0.3158622317843967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,3584,0.017863111363516915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,6144,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,12288,0.26922310723198783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,768,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,4096,0.024360888534122046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,768,0.024840889705551997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,128,0.004242666479614046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,2560,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,6144,0.009631111390060848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,7168,0.13774222797817653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,1024,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,1024,0.003866666720973121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,64,0.004204444587230682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,128,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,7168,0.010695999695195092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,51200,1.085784912109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,10240,0.0325617790222168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,256,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,65536,0.29949866400824654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,4096,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,512,0.017291555802027386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,6144,1.3116497463650174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,6144,0.052574220630857676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,512,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,512,0.02399377855989668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,51200,0.07815821965535481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,8192,0.0451182226339976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,1536,0.020573332905769348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,512,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,2048,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,32,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,1536,0.004313777718279096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,2560,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,65536,0.15882578161027697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,256,0.018956444329685636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,128,0.01016355554262797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,3072,0.0384497774971856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,3072,0.026100445124838088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,4096,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,51200,0.12026311291588677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,256,0.015127999915017022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,256,0.018889778190188937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,6144,0.03478577733039856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,768,0.031143112315071955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,512,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,512,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,128,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,1536,0.0158906661801868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,3584,0.012168889244397482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,8192,0.027425777581002977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,8192,0.04920444554752774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,7168,0.010233778092596266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,3584,0.017255999975734286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,51200,0.24193689558241102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,16384,1.7442364162868922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,5120,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,51200,0.08473689026302761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,10240,0.049633777803844877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,256,0.012030222349696688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,1024,0.03431466552946303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,16384,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,16384,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,2048,0.02077155477470822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,1024,0.019493333167499967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,6144,0.019322666856977675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,768,0.005704888867007361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,65536,0.17451111475626627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,51200,4.119494120279948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,512,0.15056710773044163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,1536,0.07322221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,7168,0.05192000004980299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,10240,0.019316444794336956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,8192,0.12146400080786811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,3072,0.028613332245084975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,128,0.0022195554855797025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,3584,0.011648889217111798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,2560,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,2560,0.0063591111037466265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,256,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,65536,0.6015511088901097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,3584,0.023338667220539514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,12288,0.04993866549597847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,8192,0.02325600054528978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,5120,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,128,0.013619555367363824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,512,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,7168,0.0520684454176161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,256,0.006826666494210561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,16384,0.017532444662517972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,65536,0.1673253377278646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,128,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,7168,0.020185778538386028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,3584,0.026036444637510512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,768,0.027282666828897264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,6144,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,256,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,2048,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,3584,0.2987271149953206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,4096,0.11057600047853257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,3072,0.025234666135576036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,8192,0.0705946683883667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,6144,0.3830720053778754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,3072,0.04956000049908956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,32,0.011914666328165265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,256,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,51200,0.07675555679533216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,12288,0.20403644773695204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,7168,0.035587555832333036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,6144,0.022859555151727464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,3584,0.09268177880181207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,3072,0.01202755586968528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,10240,0.6618089146084255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,128,0.015184889237085978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,1536,0.4374195469750299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,3072,0.007461333440409766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,16384,0.27424799071417916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,3584,0.30789955457051593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,1536,0.012718222207493253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,1536,0.012322666744391123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,32,0.026484444737434387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,3584,0.026543999711672466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,8192,0.01886044442653656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,51200,0.21026933193206787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,256,0.015108444624476962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,8192,0.040594667196273804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,512,0.018358222312397428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,51200,0.16079910596211752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,3072,0.0058648888435628675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,4096,0.0407626661989424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,4096,0.0925439993540446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,1536,0.015519110692871941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,2560,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,6144,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,10240,0.016902221573723685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,65536,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,128,0.01870044403605991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,10240,0.01631733278433482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,128,0.025676444172859192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,128,0.013824889229403602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,10240,0.07538666990068224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,7168,0.020250666472646926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,64,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,768,0.014916444818178812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,12288,0.026377777258555096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,256,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,256,0.04650222261746725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,16384,0.06530400117238362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,2560,0.027131555808915034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,2048,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,12288,0.03530844383769565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,3584,0.007824888659848107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,1024,0.01980622278319465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,16384,0.2511306603749593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,1024,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,6144,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,7168,0.012129777835475074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,256,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,6144,0.04022488991419474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,7168,0.012925333446926542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,4096,0.017528888252046373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,16384,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,65536,0.09419200155470107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,1024,0.010642666783597736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,5120,0.22375467088487414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,7168,0.0376160012351142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,5120,0.05589689148796929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,10240,0.13305777973598903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,10240,0.05157777667045593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,3072,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,1536,0.0439591109752655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,4096,0.02047644390000237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,7168,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,65536,4.849032931857638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,51200,0.08164799875683255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,5120,0.1090666651725769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,4096,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,3072,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,512,0.018886221779717337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,512,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,512,0.014836443795098199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,768,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,10240,0.02674666709370083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,32,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,16384,0.07050133413738675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,5120,0.0339680016040802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,256,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,10240,0.1597555610868666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,128,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,7168,0.0248364441924625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,16384,0.017924444542990792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,512,0.011809777882364062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,1024,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,3072,0.04043022129270766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,12288,0.030078222354253132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,1024,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,768,0.006497777584526274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,7168,1.527043554517958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,5120,0.021967111362351313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,32,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,3072,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,3584,0.07514488697052002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,1536,0.0244862222009235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,768,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,65536,0.11291910542382134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,3584,0.0424213343196445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,1024,0.013410666750537025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,4096,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,51200,0.05979022052552965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,8192,0.21406400203704834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,256,0.016523554921150208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,3584,0.029510223203235205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,51200,2.4813003540039062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,16384,0.030211554633246526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,2560,0.043613334496816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,8192,0.01440177857875824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,65536,0.18160177601708305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,7168,1.0377368927001953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,64,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,3072,0.12097333537207709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,4096,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,256,0.007829333345095316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,6144,1.029621336195204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,65536,0.10985955264833237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,768,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,51200,0.35940000745985246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,6144,0.18519377708435059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,1536,0.07111910978953044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,3584,0.03195200032658047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,4096,0.07663644684685601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,768,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,2048,0.12601244449615479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,3072,0.00903733323017756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,5120,0.26364533106486004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,10240,0.03852622376547919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,3072,0.02306488818592495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,1024,0.003950222084919612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,6144,0.0217875556813346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,4096,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,12288,0.18479643927680123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,128,0.010442666709423065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,51200,0.773936006757948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,6144,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,32,0.006043555421961679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,10240,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,1024,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,768,0.013153778182135688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,8192,0.014658666319317289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,3072,0.008105778031879002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,32,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,16384,0.03154755632082621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,6144,0.08045511113272773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,8192,0.027228444814682007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,51200,0.050219555695851646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,51200,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,6144,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,6144,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,512,0.009829333258999718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,5120,0.008063111040327284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,128,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,10240,0.16492977407243517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,128,0.017232000827789307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,4096,0.008022222254011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,65536,0.714080015818278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,8192,0.04453866680463155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,16384,0.02364088926050398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,128,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,32,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,1536,0.08505422539181179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,2048,0.041718221373028226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,32,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,256,0.0929751131269667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,65536,0.10224711232715183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,65536,0.0665608909395006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,5120,0.056223114331563316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,12288,0.03031911121474372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,256,0.006816000160243776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,10240,0.06275466415617201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,5120,0.01585155559910668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,51200,0.02034222251839108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,3072,0.013412444127930535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,51200,1.4632381863064234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,2560,0.05525955557823181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,768,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,768,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,6144,0.0075395554304122925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,6144,0.03116977877087063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,512,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,12288,0.2599546644422743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,768,0.004505777938498391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,16384,0.02204177776972453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,12288,0.05248444279034933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,51200,0.052016887399885386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,2048,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,1024,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,32,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,3072,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,65536,0.3379342291090224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,2560,0.029735111527972754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,3072,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,65536,0.06508621904585096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,65536,0.061491555637783475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,5120,0.01961777773168352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,7168,0.07723999685711332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,3072,0.13179022736019558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,7168,0.027640889088312786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,7168,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,128,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,10240,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,10240,0.04183822207980686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,65536,0.09147377808888753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,65536,0.0976604421933492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,12288,0.08393422100279067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,12288,0.05075200067626106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,3072,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,51200,0.35964356528388125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,5120,0.07987288633982341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,3072,0.004585777719815572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,2048,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,12288,0.10326755709118313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,65536,0.07172977924346924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,768,0.007376000285148621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,51200,0.4427359898885091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,7168,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,128,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,10240,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,128,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,1536,0.025176889366573755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,3072,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,2048,0.2888515525394016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,256,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,128,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,128,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,7168,0.022208889325459797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,256,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,4096,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,1024,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,512,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,4096,0.019647111495335896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,65536,0.013535999589496188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,1024,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,2560,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,16384,0.07449866665734185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,16384,0.020627554919984605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,1024,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,2560,0.3397253354390462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,7168,0.17332800229390463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,16384,0.17961866325802275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,10240,0.012375999655988483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,6144,0.025064888927671645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,512,0.01309333327743742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,5120,0.021459556288189355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,128,0.01312800000111262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,4096,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,6144,0.08932799763149685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,256,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,512,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,32,0.0075866662793689305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,3072,0.008448888858159384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,8192,0.011680000358157687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,256,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,6144,0.15199466546376547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,10240,0.057512885994381376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,128,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,5120,0.008648888932334052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,256,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,512,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,8192,0.021702221698231165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,3072,0.03965866565704346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,1536,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,8192,0.02404444416364034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,12288,0.016435555285877652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,2560,0.009295110901196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,1536,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,10240,0.12389955255720352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,12288,0.019653333557976615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,12288,0.6853617562188042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,64,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,4096,0.09377955728107029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,8192,0.13938755459255642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,768,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,1024,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,10240,0.018793778287039865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,8192,0.05769333574506971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,2048,0.03995022177696228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,768,0.02234399980968899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,2048,0.13928977648417154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,51200,0.06517066558202107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,10240,0.057151112291547984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,12288,0.029189334975348577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,2560,0.015524443652894763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,12288,0.04908355408244663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,3584,0.1649893389807807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,65536,0.1521315574645996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,8192,0.02299822204642826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,7168,0.031361778577168785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,7168,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,1024,0.006047111004590988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,6144,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,2048,0.01149511088927587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,256,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,10240,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,3584,0.08166488673951891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,6144,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,3072,0.046506666474872164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,16384,0.050945778687795006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,768,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,10240,0.17268533176845977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,3072,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,768,0.017788444956143696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,65536,0.04866755670971341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,2560,0.1231057776345147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,3072,0.10870577891667683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,6144,0.8701430956522623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,4096,0.0484444432788425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,2560,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,3584,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,5120,0.022732444935374792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,65536,0.13306221697065565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,7168,0.3107164435916477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,7168,0.06231111288070679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,5120,0.010079111489984725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,3072,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,64,0.0022764443937275144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,1024,0.006017777654859755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,512,0.033415112230512835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,5120,0.04700444473160637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,8192,0.09763555394278632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,3072,0.037131554550594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,65536,0.4348897669050429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,65536,0.3479146692487929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,512,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,4096,0.00816888858874639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,51200,0.1922586626476712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,2048,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,8192,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,5120,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,12288,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,10240,0.026366222235891554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,3072,0.039833777480655246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,7168,0.021295999487241108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,1536,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,6144,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,3584,0.19727822144826254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,3584,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,2560,0.03946222199334039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,128,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,51200,0.6355688836839464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,10240,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,256,0.017055999901559617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,512,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,256,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,512,0.020251555575264823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,256,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,12288,0.02511466708448198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,2560,0.019675556156370375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,2560,0.03120000163714091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,128,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,1024,0.027929776244693335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,4096,0.023360000716315374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,5120,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,5120,0.017109332813156974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,12288,0.06244444184833103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,16384,0.021597334080272253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,2048,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,3072,0.05508444375462002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,65536,0.12685599591996935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,2560,0.08437244759665595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,1024,0.019536000159051683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,1024,0.02831022275818719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,2048,0.026055110825432673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,2560,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,51200,0.05234400100178189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,1024,0.02068533334467146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,512,0.015860444969601102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,6144,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,7168,1.0538817511664498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,128,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,1536,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,32,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,51200,0.43599645296732586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,64,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,256,0.011704888608720569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,4096,0.13048088550567627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,12288,0.0295724438296424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,256,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,51200,0.06311200062433879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,256,0.005365333209435145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,4096,0.009673777553770278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,7168,0.06500177913241915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,768,0.025241777300834656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,8192,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,4096,0.61371824476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,1536,0.006211555666393704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,3584,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,6144,0.4047253396775987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,65536,0.25351643562316895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,10240,0.024496000674035814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,2560,0.030416889323128596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,3072,0.01680355601840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,2560,0.027911110056771174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,2048,0.040455109543270536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,51200,0.1509057813220554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,768,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,512,0.011931555966536203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,12288,0.05298666821585762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,6144,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,128,0.002452444492114915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,51200,0.129077328575982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,128,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,2560,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,2048,0.016935111747847665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,2048,0.047806223233540855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,10240,0.03274310959710015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,256,0.015154666370815702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,1536,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,65536,0.22387467490302193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,128,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,65536,0.8859724468655057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,2048,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,3584,0.020145777199003432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,512,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,768,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,3072,0.026327111654811438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,3584,0.28287021319071454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,8192,0.20913155873616537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,256,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,32,0.004536889079544279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,4096,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,3584,0.01887999971707662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,1024,0.031552000178231135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,3072,0.017845332622528076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,10240,0.09093244208229913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,768,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,768,0.02702933384312524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,512,0.006002666635645761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,10240,0.31829955842759877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,6144,0.1928631067276001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,128,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,3584,0.011112000379297467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,3072,0.0073724447025193115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,16384,0.039855112632115684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,8192,0.08490044540829128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,12288,0.037909332248899676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,128,0.015171556009186639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,7168,0.03421866562631395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,51200,0.03773244553142124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,128,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,3072,0.06292444467544556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,2048,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,4096,0.46174224217732746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,8192,0.02718577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,2560,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,12288,0.3116462230682373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,4096,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,5120,0.12444267008039687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,65536,0.6397795677185059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,768,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,51200,0.6564053429497613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,4096,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,3584,0.01610844499535031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,3584,0.017643555998802185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,2048,0.01405066748460134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,6144,0.020463999774720933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,128,0.014867555763986377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,12288,3.5717875162760415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,51200,0.08166311184565227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,3584,0.4217848777770996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,2048,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,7168,0.05088000165091621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,128,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,12288,0.3901662296719021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,1024,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,1536,0.015417777829700045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,1024,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,4096,0.028390222125583228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,6144,0.00943377779589759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,2048,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,7168,0.00958577791849772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,7168,0.05297599898444282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,3584,0.017242666747834947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,32,0.021661332911915247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,12288,0.1273075607087877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,3072,0.15054578251308864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,16384,0.6655484305487739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,512,0.007489778101444244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,12288,0.24814756711324057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,8192,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,7168,0.08064355452855428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,10240,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,8192,0.011949333051840464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,65536,0.24827912118699816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,2048,0.02943999899758233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,32,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,51200,6.621946546766494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,768,0.01625066664483812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,32,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,6144,0.09955910841623943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,65536,0.1866968870162964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,3584,0.007735110819339752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,16384,0.0732568899790446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,512,0.018274666534529794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,512,0.006086222413513396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,256,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,2560,0.0070497774415545994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,1536,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,4096,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,512,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,64,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,65536,0.04713066750102573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,2560,0.012358222570684222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,768,0.028127110666698877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,2560,0.08920711278915405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,10240,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,2048,0.006728000111050076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,7168,0.03900444507598877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,12288,0.04450577828619215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,1024,0.28984088367886013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,2048,0.03745066788461473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,51200,0.06142489115397135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,1536,0.024337778488794964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,1536,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,2048,0.017687999539905123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,256,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,3072,0.012236444486512078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,65536,3.1377875010172525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,51200,0.312537776099311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,65536,0.2563546763526069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,1024,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,10240,0.004631999880075455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,256,0.0019004444281260173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,16384,0.12838489479488796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,5120,0.00889955543809467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,51200,0.06017333269119263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,65536,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,32,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,512,0.022631999519136217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,256,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,6144,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,51200,0.06061244673199124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,8192,0.02952622373898824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,4096,0.11117955048878987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,6144,0.012789333032237159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,2560,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,256,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,64,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,2560,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,2560,0.007366221812036302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,4096,0.007687999970383114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,1024,0.02167911165290409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,1536,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,128,0.01923377811908722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,256,0.0029022221763928733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,5120,0.07314399878184001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,4096,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,3072,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,128,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,256,0.012500444220172035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,7168,0.007076444725195567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,8192,0.025416889124446448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,5120,0.02524799936347538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,7168,1.645090632968479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,51200,0.1546728875901964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,65536,0.027863999207814533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,512,0.018515555395020377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,7168,0.09726844231287639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,1536,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,3584,0.030858665704727173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,256,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,4096,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,51200,0.017848889032999676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,2048,0.1945199966430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,5120,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,3584,0.10439466767840916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,65536,0.12132533391316731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,1024,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,51200,0.05013066530227661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,7168,0.013739555246300168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,32,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,10240,0.15510311391618517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,10240,0.021008888880411785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,64,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,7168,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,5120,5.094140370686849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,12288,0.024007999234729346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,8192,0.022408000297016565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,32,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,4096,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,256,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,512,0.01715555621518029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,512,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,12288,0.22638487815856934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,10240,0.0166666673289405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,768,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,51200,5.835618760850694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,128,0.0471111105548011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,65536,0.12349066469404434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,51200,0.07420533233218722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,12288,0.020390222469965618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,7168,0.061943113803863525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,256,0.018929777873886954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,3584,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,128,0.012146666646003723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,64,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,1536,0.04227377639876472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,10240,0.029829333225886028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,256,0.005375999957323074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,65536,0.07599111398061116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,512,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,1536,0.007743111087216272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,256,0.022317333353890315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,768,0.003253333270549774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,768,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,1536,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,7168,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,1024,0.019554666346973844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,2048,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,8192,0.026572444372706946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,768,0.004239999999602635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,3584,0.07194311088985868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,2560,0.05159022079573738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,65536,0.06583822435802884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,8192,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,4096,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,128,0.006207110981146495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,32,0.043566223647859365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,7168,0.020043555233213637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,768,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,2560,0.22737956047058105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,16384,0.030208888981077407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,1024,0.03762133253945245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,8192,0.02237155536810557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,65536,0.08939288722144233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,51200,0.054530666934119336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,51200,0.1599866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,7168,0.02110933264096578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,10240,0.015576889117558798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,1024,0.05039822061856588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,65536,0.034689777427249484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,16384,0.18188267283969453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,512,0.014044443766276041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,768,0.005647111270162795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,32,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,128,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,3584,0.01568800045384301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,2048,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,10240,0.038951999611324735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,1536,0.04881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,2560,0.739191108279758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,32,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,5120,0.060028447045220264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,10240,0.016972444123691983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,128,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,32,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,64,0.008105778031879002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,3072,0.0344142218430837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,2560,0.019692444139056735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,2560,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,7168,0.040622221099005804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,16384,0.025034666061401367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,7168,0.03154666556252374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,256,0.028759111960728962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,2048,0.3679368760850694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,512,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,16384,0.4649119906955295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,12288,0.013472889032628802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,1536,0.015034667319721647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,1536,0.09178577529059516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,3072,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,768,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,768,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,64,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,16384,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,10240,33.69013129340278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,3072,0.022249778111775715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,2048,0.00433955548538102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,3584,0.007391999993059371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,128,0.007490666376219855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,10240,0.37436887953016496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,5120,1.290779537624783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,4096,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,1536,0.02868266569243537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,51200,0.07343200180265638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,12288,0.030625777112113103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,3072,0.4063875675201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,4096,0.2466648949517144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,256,0.02589599953757392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,1024,0.07077511151631673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,768,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,3584,0.10619822475645278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,5120,0.03106488784154256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,6144,0.025549333956506517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,6144,0.028575999869240656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,512,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,4096,0.19340266121758354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,4096,0.04056888818740845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,51200,0.05807199743058947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,64,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,128,0.013591110706329346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,16384,0.04895377821392483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,256,0.010746666954623329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,2560,0.564375983344184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,12288,0.15765244430965847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,12288,0.2167724503411187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,51200,0.2749120129479302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,65536,0.134533339076572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,1024,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,2560,0.013728888498412238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,2560,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,8192,0.05839822027418348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,768,0.3859279950459798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,10240,0.044143110513687134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,512,0.003825777934657203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,3584,0.06655289067162408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,32,0.006076444354322221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,512,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,8192,0.6966524653964572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,512,0.01795377830664317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,6144,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,2560,0.007559110720952352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,8192,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,51200,0.05415466758939955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,16384,0.025789333714379206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,51200,0.1964426702923245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,4096,0.01828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,65536,0.0829555524720086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,65536,0.4020675553215875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,7168,0.3004515435960558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,10240,0.08632089032067193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,8192,0.011186666786670685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,6144,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,51200,0.023404444257418316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,768,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,8192,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,1024,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,512,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,512,0.0495377779006958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,8192,1.579413308037652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,3584,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,1536,0.0233288887474272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,8192,0.2758088906606038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,256,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,16384,0.23552799224853516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,768,0.009966221948464712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,2048,0.019681778219011094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,2048,0.027826666831970215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,1536,0.020516445239384968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,2048,0.0159262220064799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,64,0.0018986666368113623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,32,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,3584,0.11644800504048665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,768,0.10220355457729763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,65536,0.11881421671973334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,5120,0.017258667283587985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,12288,0.012267555627557965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,2560,0.03261955579121908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,4096,0.029476443926493328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,8192,0.031826668315463595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,256,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,8192,0.0498542222711775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,3584,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,51200,0.047916445467207164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,1536,0.006326222171386083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,12288,0.021294222937689886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,2560,0.07402222024069892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,256,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,256,1.0673786799112956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,768,0.032772445016437106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,8192,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,2560,0.026473777161704168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,3072,0.019987555013762582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,65536,18.18508063422309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,128,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,12288,0.023631999890009563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,12288,0.02198666665289137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,10240,0.024307555622524683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,256,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,2048,0.015271999769740634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,16384,0.057778669728173145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,65536,0.06623022423850165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,4096,0.020452444752057392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,4096,0.07286844650904338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,65536,0.24174488915337455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,3072,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,5120,0.004584889031118817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,512,0.017128888103697035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,16384,0.03068355719248454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,12288,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,3584,0.04984622200330099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,768,0.06064889166090223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,256,0.02500088844034407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,7168,0.047500444783104785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,12288,0.03131644593344794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,6144,0.05510666635301378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,4096,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,5120,0.7968799802992078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,3584,0.03832799858517117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,512,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,16384,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,2048,0.016474665866957772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,8192,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,3584,0.006021333237489064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,8192,0.02293066680431366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,8192,0.07130133443408541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,51200,1.3515973620944555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,768,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,128,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,4096,0.03919733232922024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,6144,0.015612444943851896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,1536,0.005416888743638992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,1536,0.009959999885823991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,3584,0.06834399700164795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,768,0.0029493332323100832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,3584,0.03120533294147915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,512,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,5120,0.06877155436409844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,2048,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,3072,0.016599999533759225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,256,0.01716088917520311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,65536,6.461613125271267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,3584,0.059119999408721924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,16384,0.04411555661095513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,3584,0.02906844351026747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,1024,0.016914667354689706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,5120,0.09018311235639785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,51200,5.960279252794053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,256,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,2560,0.014534221755133735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,7168,0.02495288848876953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,1536,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,8192,0.056286222404903836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,4096,0.01817244456874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,128,0.015085332923465304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,64,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,2048,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,16384,0.01640977793269687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,7168,0.05184977915551928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,3584,0.007919111185603673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,8192,0.050006220738093056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,128,0.025095999240875244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,12288,0.052833778990639575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,1024,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,1024,0.03637422124544779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,768,0.01585155559910668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,2560,0.020750222934616935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,768,0.0035777779089079965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,6144,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,3072,0.12706400288475886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,10240,0.024333332975705464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,8192,0.04294666647911072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,51200,0.48829242918226456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,65536,0.07243822018305461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,12288,0.16890133751763237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,5120,0.3923635482788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,256,0.01625422139962514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,1536,0.018049778209792245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,2560,0.045904888047112354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,3584,0.1093555556403266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,32,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,8192,0.026289777623282537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,512,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,7168,0.013075555364290873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,256,0.025702221526039973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,4096,0.014134221606784396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,3584,0.09065422084596421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,512,0.05790311098098755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,7168,0.14177244239383274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,3072,0.03774844606717428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,128,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,8192,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,4096,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,256,0.0642719997300042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,6144,0.01995022263791826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,4096,0.03129155437151591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,4096,0.030440890126758154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,2048,0.011360000405046674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,12288,0.08761599991056655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,5120,0.034128000338872276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,2560,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,6144,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,768,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,1536,0.11499555905659993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,1024,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,2560,0.01817599932352702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,128,0.023736000061035156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,16384,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,3584,0.04749511016739739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,2048,0.05226133267084757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,128,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,7168,0.010103111465771994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,7168,0.01866933372285631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,10240,0.09585866663191055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,65536,0.14250132772657606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,8192,0.16308621565500894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,768,0.0036595554815398324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,1024,0.2601484457651774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,2048,0.06181955337524414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,2560,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,4096,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,65536,1.0850035349527996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,5120,0.5578533278571235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,32,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,10240,0.5665351019965278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,6144,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,16384,0.03372088736957974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,16384,0.024105777343114216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,2560,0.013600000076823764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,65536,0.1744257741504245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,3584,0.9645688798692492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,128,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,12288,0.06925955745908949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,65536,0.01744888888465034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,128,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,51200,0.06318844689263238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,64,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,65536,0.5597350862291124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,51200,0.8494328922695584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,128,0.0021111110432280437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,7168,0.0177528891298506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,6144,0.1587795548968845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,5120,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,1024,0.0032151111712058387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,10240,0.17746310763888887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,2048,0.015895111693276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,12288,0.01681066718366411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,3584,0.007804444266690149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,8192,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,5120,0.018779555956522625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,256,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,3584,0.010871111518806882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,5120,0.23911023139953613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,768,0.017239999439981248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,12288,0.1953937742445204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,12288,0.04615999923812019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,1536,0.010992888775136737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,768,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,2560,0.47168175379435223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,51200,1.470950232611762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,8192,0.01572444372706943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,2560,0.030250665214326646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,256,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,5120,0.01633155511485206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,6144,0.6187271012200249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,4096,0.046675556235843234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,51200,0.5300097995334202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,6144,0.04296000136269463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,128,0.00964355551534229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,6144,0.1317253377702501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,1024,0.0671164459652371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,51200,0.3791111045413547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,10240,0.036941332949532404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,4096,0.025282666087150574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,10240,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,5120,0.08020355304082234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,10240,0.021925333473417494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,512,0.002565333412753211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,65536,0.048800001541773476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,51200,0.051868445343441434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,768,0.012420444852775998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,768,0.011373333632946014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,3072,0.007376888559924231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,12288,0.054526223076714404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,2560,0.027436445156733196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,12288,0.3062444527943929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,256,0.10804977681901719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,8192,0.16886488596598306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,4096,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,256,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,768,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,512,0.007888889147175683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,16384,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,3072,0.03711644477314419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,3584,0.017264889346228707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,1024,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,3584,0.007524444825119442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,3584,0.074182218975491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,5120,0.114555557568868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,128,0.03596266773011949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,768,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,3072,0.01936444474591149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,65536,0.12657511234283447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,51200,0.019363555643293593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,10240,0.02852177619934082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,7168,0.019479999939600628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,64,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,10240,0.06130399968889025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,768,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,10240,0.011764444410800934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,5120,0.8864711125691732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,65536,0.577023983001709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,12288,1.579401757982042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,256,0.014379555980364481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,65536,0.08609066406885783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,3584,0.48923465940687394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,32,0.009406222237481011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,5120,0.30170223448011607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,2560,0.12160710493723552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,16384,0.0229759994480345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,3584,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,65536,5.160845438639323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,1024,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,7168,0.11872800191243489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,12288,0.041511999236212835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,4096,0.12331999672783746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,51200,1.1754080454508464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,1024,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,256,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,4096,0.041278223196665444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,12288,0.08772089083989461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,10240,0.09112444188859727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,32,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,51200,0.14183644453684488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,768,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,7168,0.023354666100607976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,512,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,1024,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,32,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,5120,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,10240,0.05955288807551066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,128,0.09184710847006904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,3584,0.007673777639865875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,7168,0.031629333893458046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,1536,0.02000266644689772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,1024,0.010668444136778513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,4096,0.03990844554371304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,2560,0.034295999341540866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,12288,0.23460621303982207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,256,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,4096,0.21973866886562773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,512,0.0469688872496287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,5120,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,256,0.014477334088749357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,2560,0.011033777561452655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,8192,0.014645333091417948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,51200,0.11481066544850667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,32,0.006001777946949005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,51200,0.6012861993577745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,12288,0.04336977667278714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,32,0.023312888211674158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,512,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,16384,0.13719199763403997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,512,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,128,0.047353777620527476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,7168,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,2048,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,51200,0.06324888600243463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,51200,0.0521964430809021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,5120,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,7168,0.18703822294871011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,512,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,256,0.05239377750290764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,7168,0.0373253325621287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,3072,0.15766489505767822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,3072,0.015659555792808533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,7168,0.019596444235907663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,3072,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,51200,1.5434808731079102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,64,0.021540444758203294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,65536,0.09285422166188557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,1536,0.05488533443874783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,4096,0.0600222216712104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,128,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,256,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,128,0.020453333854675293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,512,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,7168,0.1845617824130588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,7168,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,51200,0.5735226737128364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,2560,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,2048,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,256,0.0373982224199507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,256,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,16384,0.044312887721591525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,64,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,768,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,512,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,16384,0.03140177660518222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,3584,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,10240,0.01699555582470364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,128,0.011529777612951068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,128,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,51200,6.059083726671006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,4096,0.07573955588870578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,64,0.013367111484209696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,51200,0.10603466961118911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,3072,0.1936213307910495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,51200,0.16421866416931152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,1024,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,512,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,256,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,7168,0.027662222584088642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,512,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,3072,0.007409777906205919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,12288,0.021477333373493616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,128,0.0024542222834295696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,4096,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,6144,0.017126222451527912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,65536,0.0771262248357137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,12288,0.01723911033736335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,8192,0.10050844483905369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,7168,0.047651555803087026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,32,0.007687110867765214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,3072,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,16384,0.04599555664592319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,12288,0.0356391105386946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,2048,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,7168,0.07581955856747098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,1536,0.006203555398517185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,512,0.015508444772826301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,12288,0.046911110480626426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,2048,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,5120,0.021583999196688335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,4096,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,2560,0.01165511127975252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,7168,0.05504711137877571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,4096,0.01776622235774994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,768,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,512,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,51200,3.0705945756700306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,8192,0.03333777851528592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,65536,0.6383982234530979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,256,0.026218665970696345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,5120,0.302511109246148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,8192,0.039691554175482854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,5120,0.03361422154638503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,3072,0.047916445467207164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,65536,0.15433955192565918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,6144,0.04053777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,8192,0.031028442912631567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,4096,0.01425866617096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,7168,0.015219555960761176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,4096,0.010789333118332757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,6144,0.027456889549891155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,256,0.06158755885230171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,5120,0.8952222400241427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,6144,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,64,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,1536,0.028600000672870215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,5120,0.1608017815483941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,8192,0.06680533621046278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,65536,0.06753689050674438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,65536,0.4963582356770833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,512,0.017271111408869427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,6144,0.33992799123128253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,51200,0.09969866938061184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,64,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,768,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,7168,0.012719111310111152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,256,0.013638222383128272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,10240,0.49131377538045246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,128,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,16384,0.05599821938408745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,64,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,12288,0.059530668788486056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,128,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,16384,0.2507786750793457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,2560,0.01826400061448415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,7168,0.44828446706136066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,128,0.017172444197866652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,1536,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,64,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,8192,0.013635555903116861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,2048,0.07884533537758721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,5120,0.034916443957222834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,10240,0.055366221401426524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,7168,0.1569653352101644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,2048,0.005643555687533484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,10240,0.055461333857642285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,4096,0.02439733346303304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,10240,0.11491910616556804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,6144,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,1536,0.02163022259871165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,6144,0.011583999627166323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,512,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,2048,0.013733333183659447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,2560,0.0244151105483373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,6144,0.011976888610257043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,2048,0.04281333420011732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,7168,0.030644443300035264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,65536,0.16242755783928767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,65536,0.05705866548750135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,1536,0.011878222227096558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,1024,0.01568888955646091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,2048,0.005208888815508948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,3584,0.35233955913119847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,8192,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,2560,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,10240,0.02849066588613722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,768,0.005624889085690181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,256,0.0035991109907627106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,4096,0.08362578021155463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,8192,1.4647173351711695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,2560,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,6144,0.16941511631011963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,65536,2.065185758802626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,2048,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,768,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,7168,0.014519110321998596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,3584,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,65536,0.12742222679985896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,768,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,10240,0.21256444189283583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,7168,0.021307556165589228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,256,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,6144,0.0491484436723921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,65536,0.3032044569651286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,1024,0.015559110376569959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,4096,0.03212177753448486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,2560,0.05951733059353299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,2048,0.3298124472300212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,6144,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,512,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,64,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,51200,0.05363466673427158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,2048,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,5120,0.14396266142527261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,6144,0.017664889494578045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,12288,0.017848889032999676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,7168,0.01236177815331353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,16384,0.02401688860522376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,512,0.031930665175120033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,4096,0.017849778135617573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,768,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,256,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,512,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,3584,0.021307556165589228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,768,0.06668266985151503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,768,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,768,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,8192,0.07766666677263048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,768,0.01739555597305298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,3072,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,65536,0.7007395426432291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,1536,0.005444444302055571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,3072,0.027965333726671007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,3072,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,7168,0.004253333227501975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,768,0.030191110240088567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,2560,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,2560,0.7105049027336968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,64,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,6144,0.05079911152521769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,512,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,5120,0.01570933394961887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,1536,0.018633777896563213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,16384,0.01593244406912062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,16384,0.10113244586520725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,2048,0.028113779094484117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,4096,0.15493422084384495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,1024,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,3072,0.04647022154596117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,16384,0.038718223571777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,1024,0.047510223256217114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,128,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,16384,0.04266311062706841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,1024,0.016193777322769165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,3072,0.15478932857513428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,32,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,2048,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,7168,0.03221244282192654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,51200,1.30403380923801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,3072,0.5499075783623589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,5120,0.07875644498401217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,32,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,3584,0.015967110792795818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,3072,0.00756444450881746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,512,0.01370844410525428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,7168,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,3072,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,3584,0.010195555786291758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,768,0.02310133311483595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,8192,0.05044800043106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,6144,0.0454346670044793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,512,0.1643999947441949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,51200,0.4552488856845432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,32,0.011751111182901593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,51200,0.5508986579047309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,7168,0.06007555458280775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,1536,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,3584,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,12288,0.03568266828854879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,3072,0.028573334217071533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,5120,0.015260444747077094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,7168,0.01424533294306861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,10240,1.6932284037272136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,2048,0.02830844455295139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,3072,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,2560,0.02737155556678772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,65536,0.11155200004577637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,5120,0.44000177913241917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,16384,0.2025066614151001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,2048,0.02602488961484697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,65536,0.4410257869296604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,7168,0.06049599912431505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,512,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,128,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,6144,0.09685689210891724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,7168,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,51200,0.48580000135633683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,8192,0.04364711046218872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,1536,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,51200,0.046866665283838906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,51200,0.1732177734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,32,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,3584,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,512,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,10240,0.016889777448442247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,10240,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,512,0.04347911145952013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,6144,0.028233776489893597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,51200,0.24957333670722115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,5120,0.03119644522666931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,32,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,1024,0.16121778223249647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,768,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,2560,0.028368890285491943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,10240,0.0766400032573276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,1536,0.009631111390060848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,1536,0.01572444372706943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,3072,0.10171466403537327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,64,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,1024,0.011007111105653973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,7168,0.034635557068718806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,12288,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,768,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,4096,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,3584,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,7168,0.08135822084214953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,64,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,32,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,128,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,2048,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,3072,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,16384,0.44027288754781085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,256,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,2560,0.014787554740905762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,256,0.031376888354619346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,10240,0.02662311163213518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,65536,0.7866773075527616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,10240,0.029912001556820337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,1536,0.023050665855407715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,2048,0.047621332936816745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,1536,0.016672000288963318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,16384,0.08365777466032241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,5120,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,2560,0.06960533062616985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,5120,0.020561777883105807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,768,0.02896266513400608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,64,0.007521777517265744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,1024,0.009568889107969072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,3584,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,12288,0.03597155544492934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,8192,0.015288889408111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,2048,0.008033778104517195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,5120,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,65536,0.23881154590182832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,6144,0.01960355540116628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,768,0.11500622166527642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,2048,0.004940444396601783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,1024,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,2048,0.004510222209824456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,32,0.004609777695602841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,12288,0.12208710776435004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,256,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,3072,0.005378666851255629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,5120,0.0285502208603753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,8192,0.018167111608717177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,12288,0.036898665957980685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,5120,0.036722666687435575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,16384,0.03583644496070014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,3072,1.1204168531629775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,51200,0.0284862220287323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,2560,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,5120,0.08042133516735502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,51200,0.0631822215186225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,10240,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,2048,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,16384,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,5120,0.11548799938625759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,64,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,8192,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,2560,0.09646133581797282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,51200,5.027335272894965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,12288,0.07629422346750896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,65536,0.30637065569559735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,65536,0.06332888868119982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,6144,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,65536,1.2749190860324435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,32,0.008059555457697975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,8192,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,64,0.0026088888860411118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,12288,0.13931822776794434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,65536,0.07693422502941556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,3072,0.01350400017367469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,128,0.017166222135225933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,768,0.01644977761639489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,6144,0.012712000144852532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,2560,0.031720889939202204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,128,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,128,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,6144,0.010487111078368293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,16384,0.05405066741837395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,8192,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,16384,0.03613777624236213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,5120,0.1280462212032742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,8192,0.11003377702501084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,16384,0.17463378111521402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,16384,0.19922844568888345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,4096,0.012488000094890594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,2560,0.011179555621412067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,768,0.03656977746221755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,3584,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,4096,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,3584,0.0635991096496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,2560,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,51200,0.050591998630099826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,1024,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,64,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,12288,0.036280890305837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,12288,0.1804542276594374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,256,0.04256355431344774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,16384,0.1118311087290446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,128,0.012479110724396177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,12288,0.03380089004834493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,10240,0.016552888684802584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,65536,0.07088266478644477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,51200,0.06939022408591376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,16384,0.03073955575625102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,6144,0.027464888162083093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,3584,0.03548977772394816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,7168,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,5120,0.03978044456905789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,2048,0.03475822342766656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,1536,0.011994666523403592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,5120,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,1024,0.05191555619239807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,768,0.19666311475965711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,4096,0.03974399964014689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,16384,0.42691556612650555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,7168,0.15023110972510442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,1536,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,51200,0.19764355818430582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,4096,0.07591733005311754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,7168,0.0195768889453676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,5120,0.0634675555759006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,5120,0.017488888568348356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,16384,0.04732444551255968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,7168,0.08558399809731378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,51200,0.11230488618214925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,16384,0.24666044447157118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,10240,0.040856887896855675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,512,0.004918222212129169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,64,0.03646577729119195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,12288,0.027926223145590887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,32,0.0017164444757832421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,3072,0.011323555476135679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,2048,0.20824710528055826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,1024,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,512,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,256,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,1024,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,2048,0.019472888774342008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,3584,0.01572088897228241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,128,0.020255110330051847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,3584,0.008079110748238033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,1024,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,2560,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,4096,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,128,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,2560,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,128,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,512,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,16384,0.10739644368489583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,4096,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,65536,0.9248453776041666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,3072,0.00924711094962226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,3584,0.008596444295512306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,768,0.003140444349911478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,8192,0.13554044564565024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,65536,10.30664316813151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,8192,0.26637066735161674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,7168,0.05916000074810452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,512,0.019215111931165058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,16384,0.02737866673204634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,65536,0.1600115564134386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,768,0.09990844461652969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,3584,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,12288,0.09821866618262397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,12288,0.021976000732845728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,5120,0.008346666892369589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,32,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,3584,0.06281866629918416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,1024,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,10240,0.10792977942360772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,51200,0.19241510497199166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,6144,0.07450489203135173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,3584,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,65536,0.2509760061899821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,3584,0.09094755517111884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,1536,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,8192,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,10240,0.036017778846952654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,6144,0.10703377591239081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,1024,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,256,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,2048,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,10240,0.01755288905567593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,8192,0.011051555474599203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,512,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,2560,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,512,0.011774222056070963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,3584,0.01053955571519004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,12288,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,12288,0.1200595564312405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,5120,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,512,0.00572533326016532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,128,0.025571554899215698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,6144,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,51200,0.05927733580271403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,8192,0.019935111204783123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,16384,0.0759813321961297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,1024,0.028530667225519817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,16384,0.011386666860845355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,10240,0.0336444444126553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,1024,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,1536,0.02722399930159251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,3584,0.014017777310477363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,3584,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,6144,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,3072,0.0119991112086508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,12288,0.07811999983257718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,256,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,256,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,256,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,64,0.0017760000709030365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,5120,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,3584,0.02853333287768894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,1024,0.016584888100624084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,3072,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,256,0.018631110588709515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,1536,0.16237955623202854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,2048,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,12288,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,32,0.004808889081080754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,1024,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,2048,0.01239999963177575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,4096,0.1016426682472229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,3072,0.031168000565634832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,1024,0.01517688896920946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,12288,0.19621600045098198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,12288,0.04914311236805386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,4096,0.02414400047726101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,4096,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,5120,0.03353955679469638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,768,0.013231111069520315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,32,0.028625776370366413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,1536,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,768,0.002895111011134254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,3584,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,5120,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,12288,0.019814221395386588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,4096,0.017472889688279893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,7168,0.021967111362351313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,7168,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,1536,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,2048,0.009684444301658207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,4096,0.0561715563138326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,16384,0.364757325914171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,8192,0.21528266535864937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,64,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,512,0.002581333327624533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,2048,0.010899555351999072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,51200,0.42975733015272355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,4096,0.017101332545280457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,1024,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,6144,0.6920871204800076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,128,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,1024,0.03487111131350199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,256,0.018236445056067575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,51200,12.131595187717013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,16384,0.0244186669588089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,768,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,8192,0.018379555808173288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,12288,0.21669156021542021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,12288,0.05837777588102552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,12288,0.0299582216474745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,2560,0.02652799917591943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,7168,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,1024,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,5120,0.019544000426928203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,8192,0.019506666395399306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,512,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,8192,0.11929599444071452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,1536,0.17863021956549752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,512,0.005509333478079901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,12288,0.24951023525661895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,1536,0.058874666690826416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,2048,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,4096,0.02347999976740943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,4096,0.10609155231051975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,256,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,5120,0.00489688871635331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,6144,0.04108533263206482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,1536,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,128,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,512,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,5120,0.034994665119383074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,512,0.005429333282841577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,1536,0.1479920016394721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,2560,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,1536,0.021956443786621094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,2560,0.035624000761244036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,3072,0.1684622234768338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,3072,0.07550311088562012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,768,0.015815110670195687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,4096,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,1024,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,2560,0.028232001596026953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,16384,0.10148355695936416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,2560,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,5120,0.016140444411171805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,7168,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,7168,0.7038559913635254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,1024,0.004100444416205089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,256,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,12288,1.8745394812689886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,128,0.022963555322753057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,4096,0.027923554182052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,65536,0.0746977792845832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,65536,7.736360337999131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,65536,0.07980799674987793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,64,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,16384,0.6643759939405653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,32,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,6144,0.442488882276747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,8192,0.018198221921920776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,6144,0.17992799811893037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,6144,0.06482222345140246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,5120,1.566250695122613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,51200,0.4943057696024577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,768,0.0169493340783649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,1536,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,4096,0.008030222521887885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,768,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,7168,0.1381031142340766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,3584,0.05408622158898247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,3072,0.033377776543299355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,16384,0.1852355533175998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,2048,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,65536,0.15628088845147028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,128,0.03818488783306546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,7168,0.017384888397322763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,12288,0.01850488947497474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,6144,0.1170568863550822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,16384,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,2560,0.568261358473036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,5120,0.10355644755893284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,128,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,4096,0.25647022989061147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,1536,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,51200,0.05173600051138136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,12288,0.05795733133951823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,3584,0.017079111602571275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,1024,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,12288,0.005008000052637524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,65536,0.20525688595241973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,768,0.11383377181159125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,512,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,51200,1.0078257454766166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,768,0.013635555903116861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,3072,0.010975999964608086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,768,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,16384,0.024685333172480266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,3584,0.06178222099939982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,16384,0.014166222678290473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,4096,0.03015822172164917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,3584,0.007241778075695038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,51200,1.8498783111572266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,3584,0.09859466552734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,1536,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,5120,0.012610666453838348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,1536,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,1536,0.06290844413969252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,65536,2.694779502020942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,65536,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,1536,0.022982221510675218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,4096,0.05295111073387993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,128,0.007376000285148621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,32,0.028194665908813477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,8192,0.022592888938056097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,8192,0.05125777920087179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,8192,0.08223822381761339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,3584,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,5120,0.016972444123691983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,256,0.023416888382699754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,16384,0.08811110920376247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,256,0.013953778478834363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,12288,0.04821244544453091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,10240,0.017691555950376723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,8192,0.03494666682349311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,2560,0.04124088750945197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,51200,0.10051200124952529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,2048,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,8192,0.169615109761556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,768,0.006702222343948152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,4096,0.5310728814866807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,256,0.28589510917663574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,65536,0.6952960226270887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,65536,1.0517280366685655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,256,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,64,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,1024,0.023345778385798138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,10240,0.0225600004196167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,512,0.05567466550403171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,8192,0.0214373336897956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,768,0.008237332933478886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,3584,0.026991999811596338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,512,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,16384,0.10397777954737346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,1536,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,8192,0.05567644702063667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,6144,0.07101422548294067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,3072,0.010810666614108615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,10240,0.06767821974224515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,2048,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,8192,0.14035733540852866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,32,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,1024,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,2560,0.06871377759509616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,3072,0.0137155552705129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,768,0.030751999881532457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,5120,0.04639288783073425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,1536,0.030878222650951807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,2048,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,768,0.014141332772043018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,16384,0.1085217793782552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,6144,0.042740444342295326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,128,0.003580444388919406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,12288,0.15009155538347033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,8192,0.019122666782803006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,768,0.004587555511130227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,16384,0.11378222041659886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,7168,0.2677084339989556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,51200,0.27200267050001353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,128,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,5120,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,12288,0.01145155562294854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,51200,0.12487999598185222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,5120,0.014306666122542487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,7168,0.20073510540856254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,3072,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,3072,0.021744888689782884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,3072,0.03113244308365716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,1536,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,2048,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,51200,0.05687644746568468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,1024,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,6144,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,16384,0.05985155370500353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,8192,0.11964089340633816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,2560,0.035164445638656616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,3584,0.06365066766738892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,3072,0.023592000206311543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,2048,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,3072,0.008603555460770925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,3072,0.021351999706692163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,5120,1.9570070902506511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,2048,0.008788444101810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,2048,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,1024,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,256,0.01752088963985443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,1024,0.14617155657874212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,65536,1.055472903781467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,65536,0.13983466890123156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,32,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,8192,0.04826400015089247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,2560,0.019294222195943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,768,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,64,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,6144,0.034633778863483004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,10240,0.020104888412687514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,2560,0.027673777606752183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,12288,0.021311110920376245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,128,0.01349066694577535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,3072,0.08440533611509536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,128,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,16384,0.9457946353488498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,32,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,32,0.003573333223660787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,4096,0.00851644410027398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,512,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,5120,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,3072,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,8192,10.196349249945746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,128,0.036460445986853704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,12288,0.027412444353103638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,3072,0.025992888543340895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,768,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,3072,0.012853333519564735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,2048,0.01223288890388277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,65536,0.4953102005852594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,51200,0.05166133244832357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,1536,0.007391999993059371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,10240,0.13257243898179796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,768,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,51200,0.10561688741048177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,10240,0.04414044486151802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,128,0.015999999311235216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,1024,0.006065777606434292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,7168,0.06916889217164782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,4096,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,51200,0.050718221399519176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,3584,0.023366222778956097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,12288,0.036794665786955096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,4096,0.01721333298418257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,8192,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,5120,0.009720889230569204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,8192,0.03130755490726895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,5120,0.20083022117614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,51200,0.0709279974301656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,6144,0.03529511226548089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,5120,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,7168,0.02951288885540432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,3584,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,4096,0.10072533289591472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,10240,0.052829331821865506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,5120,0.06425511174731784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,3072,0.1715164449479845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,10240,0.035776889986462064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,51200,0.05790044201744927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,768,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,10240,0.2859644360012478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,65536,0.3189146783616808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,10240,0.011782222323947482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,16384,0.4890524546305339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,51200,0.046282665597067944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,7168,0.02546399997340308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,512,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,5120,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,8192,0.0507919987042745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,128,0.006108444597986009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,7168,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,1536,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,6144,0.06650400161743164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,1536,0.023774221539497375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,51200,0.1551084385977851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,512,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,16384,0.06608977582719591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,64,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,8192,0.02972088919745551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,51200,0.33648888270060223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,6144,0.32351554764641655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,5120,0.03439555565516154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,10240,0.019701333509551156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,65536,0.19075111548105875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,12288,0.5381644566853842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,768,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,7168,0.031000001562966242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,12288,0.37636176745096844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,65536,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,6144,0.022848000129063923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,2560,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,256,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,5120,0.059821334150102407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,51200,2.208340538872613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,1536,0.030833777454164293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,4096,0.024159111910396155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,65536,0.4377466572655572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,10240,3.150592803955078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,8192,0.08672889073689778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,2048,0.024158222807778254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,4096,0.040531555811564125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,2560,0.009082666701740688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,5120,0.04127466678619385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,768,0.018093332648277283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,8192,0.01795111099878947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,12288,0.2080542246500651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,2048,0.027108444107903376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,7168,0.07674666908052233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,10240,0.033377776543299355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,8192,0.014225777652528552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,1024,0.07806489202711317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,12288,0.021969777014520433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,5120,0.10715288586086696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,6144,0.05610577927695381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,5120,1.2336532804701064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,5120,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,128,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,3584,0.1128871175977919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,65536,0.1617866622077094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,32,0.0022239999638663395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,3072,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,5120,0.14764177799224854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,128,0.01019377758105596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,7168,0.061186664634280734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,7168,0.012541333006487953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,2048,0.031403554810418025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,4096,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,12288,0.019987555013762582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,12288,0.03619111246532864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,5120,0.4037297831641303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,65536,0.1294071144527859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,8192,0.022278222772810195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,3584,0.025739555557568867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,3584,0.33857154846191406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,5120,0.25664620929294163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,16384,0.042867554558648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,2048,0.5042319827609593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,512,0.002611555573013094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,5120,0.9071182674831815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,6144,0.3958088821834988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,6144,0.022971555590629578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,5120,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,65536,0.14206133948432073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,5120,0.05851111147138807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,8192,0.1210284498002794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,65536,0.45643112394544816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,1536,0.007177777588367462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,512,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,51200,0.05594755543602837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,51200,3.2348577711317272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,512,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,256,0.028117332193586562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,7168,0.4809608989291721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,256,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,5120,0.025461332665549383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,6144,0.013041777743233575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,5120,0.02304711110062069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,256,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,8192,0.02086488902568817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,5120,0.4406435489654541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,51200,10.904937744140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,6144,0.009165333376990424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,1024,0.005984000033802456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,10240,0.01891288823551602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,128,0.0018826667219400406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,32,0.01741066740618812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,7168,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,2048,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,4096,0.059803558720482715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,65536,0.011360888679822287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,1536,0.020042666130595736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,8192,0.027882665395736694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,128,0.01477688882086012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,1024,0.11204444037543403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,512,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,128,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,6144,0.013257777525318993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,12288,0.5879635810852051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,512,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,256,0.0057964444988303715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,5120,0.062308443917168506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,1024,0.01161688897344801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,10240,0.02075466679202186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,12288,0.013783999615245394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,4096,0.48556089401245117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,7168,0.06437155273225573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,10240,0.29203467898898655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,768,0.024688000480333965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,512,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,16384,0.2960142294565837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,2560,0.027448889282014634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,768,0.030768887864218816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,7168,0.01237777786122428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,5120,0.04776888754632738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,3072,0.004515555583768421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,4096,0.015049777097172208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,32,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,256,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,12288,0.05308444301287333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,4096,0.008065777520338694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,10240,0.07687466674380831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,51200,1.135800043741862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,6144,0.015307555596033732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,65536,0.05894133117463854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,1024,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,512,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,65536,0.1263937817679511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,6144,0.030012445317374334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,512,0.028933333026038274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,8192,0.038442667987611555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,51200,0.12080710464053684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,65536,0.7721004486083984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,768,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,256,0.0155404441886478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,2048,0.030761778354644775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,3072,0.013199999928474426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,3072,0.08456710974375407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,2048,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,3584,0.03234044379658169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,12288,0.17769243982103136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,65536,0.08529422018263076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,1536,0.019712888532214694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,7168,0.06891910897360908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,10240,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,12288,0.20719288455115425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,65536,8.032771640353733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,256,0.06891110870573255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,1536,0.06258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,1024,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,512,0.016924444172117446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,1024,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,12288,0.08416889111200969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,512,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,3584,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,256,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,2560,0.007062222394678328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,8192,0.01167555567291048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,2560,0.04227377639876472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,8192,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,12288,0.06963911321428087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,5120,0.0865315530035231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,768,0.0058764442801475525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,12288,2.348465813530816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,32,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,128,0.022276444567574397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,1536,0.009739555418491364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,4096,0.1913342210981581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,32,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,4096,0.015505777464972602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,1536,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,10240,0.03257511059443156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,16384,1.704212400648329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,2560,0.053095999691221446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,51200,0.05572799841562907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,256,0.022190221481853064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,1536,0.12373510996500652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,5120,0.21668267250061035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,10240,0.0449795557392968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,5120,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,1536,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,128,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,8192,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,1536,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,32,0.00318755561278926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,1536,0.020637333393096924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,6144,0.044249776336881846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,16384,0.023404444257418316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,3584,0.020167110694779288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,12288,0.019588443968031142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,4096,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,128,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,1536,0.22185244825151232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,2048,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,65536,0.07153866688410442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,1536,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,1024,0.743010679880778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,12288,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,51200,0.05236622360017565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,128,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,64,0.2007288932800293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,4096,0.023717333873112995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,7168,0.009193778038024902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,256,0.01553600033124288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,1024,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,7168,0.22614312171936035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,5120,0.036370668146345354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,10240,0.016304888659053378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,16384,0.481661319732666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,2048,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,512,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,51200,0.46973064210679794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,256,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,16384,0.02012266715367635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,512,0.006128888991143968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,65536,0.05784889062245687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,16384,0.025807110799683466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,3584,0.0991048879093594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,3072,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,1536,0.007078222102589077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,512,0.009491555392742157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,1024,0.01444533301724328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,3584,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,2560,0.05537244346406725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,768,0.0420702232254876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,16384,0.1958035495546129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,51200,0.04349866509437561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,51200,0.05244355400403341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,10240,0.0328453348742591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,3584,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,16384,0.03896977835231357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,4096,10.063056945800781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,5120,0.0284186667866177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,128,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,12288,0.0411208901140425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,4096,0.022700443863868713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,6144,0.013063111239009433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,4096,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,5120,0.20075378153059217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,7168,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,8192,0.09363377756542629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,512,0.003248888792263137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,12288,0.10528977711995442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,512,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,768,0.0140684445699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,512,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,6144,0.05866577890184191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,51200,0.6382986704508463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,512,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,12288,0.022689777943823073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,5120,0.027727999620967444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,32,0.0021973333011070886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,64,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,12288,0.03468266791767544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,7168,0.525569756825765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,4096,0.010394666757848527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,8192,0.1793973313437568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,3584,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,6144,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,512,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,65536,0.08732177813847859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,768,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,128,0.007740444607204861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,512,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,32,0.06578755378723145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,10240,0.11059733231862386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,2048,0.07687999804814656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,10240,0.17599733670552573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,512,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,128,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,1536,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,65536,0.11227111021677653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,128,0.005427555491526921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,3072,0.5511822170681423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,16384,0.018792000081804063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,7168,0.041722665230433144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,51200,0.06979111168119642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,16384,0.3096666600969103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,2560,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,64,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,128,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,5120,0.037159999211629234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,5120,0.029688000679016113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,6144,0.226402653588189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,16384,0.021047110358874004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,16384,0.04540888799561394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,2560,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,4096,0.011378666592968835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,65536,2.065237257215712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,256,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,1536,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,32,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,7168,0.0110871113008923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,2560,0.11525689231024848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,51200,0.10556711090935601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,32,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,4096,0.03548889027701484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,1024,0.007687110867765214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,768,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,64,0.01498933302031623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,64,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,8192,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,2560,0.021016000045670405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,4096,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,128,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,16384,0.44303022490607363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,16384,1.6085084279378254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,256,0.021929777330822412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,64,0.04528266522619459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,7168,0.024704888463020325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,12288,0.021528888079855178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,7168,0.01254488858911726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,6144,0.03068000078201294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,7168,0.06563111146291097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,65536,0.2366408771938748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,32,0.006168889088763132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,128,0.006168889088763132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,51200,0.1422613329357571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,2048,0.039814223845799766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,51200,26.442304823133682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,32,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,3584,0.008210666477680206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,51200,0.08092621962229411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,2560,0.03563466668128967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,1024,0.06692977746327718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,65536,0.1293626626332601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,1536,0.030850665436850652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,64,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,512,0.26818667517768013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,512,0.03617599937650893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,5120,1.0380559497409396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,4096,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,7168,0.01276533305644989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,6144,0.05668977896372477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,64,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,12288,0.01517599986659156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,256,0.14420533180236816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,1024,0.024139554964171514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,3072,0.1429191165500217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,5120,0.009860444400045607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,16384,0.07571733660168119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,6144,0.4369537830352783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,7168,0.04285244478119744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,3584,0.022515555222829182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,6144,0.03783822059631348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,2048,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,2048,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,51200,0.1978586647245619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,5120,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,12288,0.09227910968992446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,512,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,12288,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,16384,0.07655733161502414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,51200,0.48467376497056747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,8192,0.013262222210566202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,12288,0.1534257729848226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,5120,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,2048,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,1024,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,1536,0.015863110621770222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,768,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,10240,0.15356266498565674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,512,0.006786666810512543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,16384,0.6324293348524306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,512,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,6144,0.17601688702901205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,12288,0.03366044494840834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,4096,0.020577778418858845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,3072,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,256,0.013750221994188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,6144,1.4261422687106664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,2560,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,1536,0.02865866820017497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,7168,0.12423643800947402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,65536,0.06396533383263482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,10240,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,64,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,5120,0.151311993598938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,6144,0.013648000028398303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,6144,0.026996443669001263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,4096,0.0548000004556444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,6144,0.03984977801640829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,64,0.0898017750846015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,512,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,1536,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,1536,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,5120,0.004883555488453971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,6144,0.08426755666732788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,1536,0.023129777775870428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,64,0.0021342221233579847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,2560,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,1024,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,2560,0.03728355632887946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,128,0.004519111166397731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,7168,0.02232444451914893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,1024,0.011738667057620155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,6144,0.007068444457319047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,2048,0.008407110969225565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,12288,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,8192,0.021111999948819477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,8192,0.04370311233732435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,2048,0.0038782221575578055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,1536,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,12288,0.7894959979587131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,65536,0.10464355680677627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,512,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,2048,0.05502577622731527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,768,0.020595555504163105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,51200,0.05502666698561775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,2048,0.01937244501378801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,1024,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,2048,0.028224888775083754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,256,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,256,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,8192,0.19697599940829805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,512,0.015448888142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,512,0.030671109755833943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,32,0.011352000137170156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,2560,0.021299555897712708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,2560,0.015423110789722867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,6144,0.0719368921385871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,3072,0.40225333637661403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,3072,0.46936798095703125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,768,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,512,0.02734577821360694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,128,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,32,0.001855111163523462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,6144,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,256,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,3072,0.44644620683458114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,3584,2.750014199150933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,128,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,7168,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,6144,0.024665777881940205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,1024,0.003698666476541095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,1024,0.009835555321640438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,10240,0.02068622244728936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,65536,0.09788266817728679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,4096,0.02929333183500502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,4096,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,768,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,256,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,8192,0.032724443409177996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,4096,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,2048,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,6144,0.03189422355757819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,5120,0.06977422369851007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,768,0.014523555835088095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,1024,0.07969688706927829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,32,0.011736888852384357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,64,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,3584,0.01757333344883389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,8192,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,65536,1.4156391355726454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,1536,0.04726577798525492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,64,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,2560,0.015070221490330167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,2048,0.016064888901180692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,2048,0.04233955674701267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,5120,0.010249777800507015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,7168,0.034640000926123724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,128,0.029554665088653564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,256,0.01572355628013611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,12288,0.08188177479637994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,512,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,7168,0.2401937908596463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,6144,0.04548177785343594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,3584,0.020727111233605277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,256,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,7168,0.018868444694413077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,10240,0.0492213335302141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,51200,0.10920888847774929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,512,0.009962666365835402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,10240,0.22095289495256212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,4096,0.03688177797529433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,1536,0.02020088831583659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,7168,0.024309333827760484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,10240,0.0654017792807685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,2560,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,5120,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,7168,0.06481511063045926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,4096,0.015359110302395292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,768,0.0029297777348094513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,7168,0.031128889984554712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,5120,0.058692442046271436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,64,0.0058337777025169795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,6144,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,65536,0.0861591100692749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,4096,0.13131644990709093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,2048,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,32,0.03410844339264764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,32,0.0412986675898234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,10240,0.02068533334467146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,256,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,3584,0.00453955555955569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,2048,0.010681777364677854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,8192,0.010512000156773461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,4096,0.1027031143506368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,3584,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,1536,0.06593422094980876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,8192,0.04941866795221964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,128,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,128,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,512,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,256,0.016809778081046212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,2048,0.10244977474212646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,64,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,3072,0.031231111950344507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,256,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,2560,0.01240888900227017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,7168,0.1629519992404514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,512,0.02075199948416816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,65536,0.23136978679233125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,10240,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,2560,0.3587262100643582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,512,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,7168,0.01977422171168857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,12288,0.01903199983967675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,512,0.0058311112225055695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,6144,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,256,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,4096,0.07865866687562731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,5120,0.4262479941050212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,128,0.011735111474990845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,2560,0.05724355247285631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,2048,0.02312444481584761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,128,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,512,0.042908443344963916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,3584,0.12109243869781494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,16384,0.14593511157565647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,4096,0.10017510917451646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,65536,0.06376800272199842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,768,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,1536,0.07397511270311144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,2560,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,10240,0.022721777359644573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,768,0.04134577843878005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,3584,0.6260088814629449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,51200,0.34714221954345703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,7168,0.5160782072279189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,512,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,32,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,3072,0.007825777762466008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,65536,0.21111822128295898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,3584,0.024667556087176006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,8192,0.014192000031471252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,256,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,1536,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,2560,0.02161955502298143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,128,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,4096,0.04760622315936618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,10240,0.19822311401367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,128,0.009187555147541894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,65536,0.08407733175489639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,7168,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,1536,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,10240,0.10690310928556655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,2560,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,128,0.010111110905806223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,512,0.012286221815480126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,3072,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,10240,0.042450666427612305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,6144,0.009539555344316695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,3072,0.02017511096265581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,3072,0.10181688600116306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,1536,0.01969688965214623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,5120,0.014755555325084262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,7168,0.21059732966952852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,16384,0.045448889334996544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,1536,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,8192,0.024269334144062463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,10240,0.07088266478644477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,3072,0.0979164441426595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,512,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,3584,0.023919110496838886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,256,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,51200,0.8787315686543783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,16384,0.021652445197105408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,3072,0.027288888891537983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,2048,0.08368088801701863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,16384,2.316897710164388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,51200,1.2547235488891602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,10240,0.026241777671708003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,65536,0.2688302199045817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,8192,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,2560,0.090957330332862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,1024,0.030193779203626845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,3072,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,3072,0.007377777662542131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,4096,0.016209777858522203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,7168,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,1536,0.048400001393424145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,5120,0.00960533320903778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,51200,0.08718399869071113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,65536,2.620256847805447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,51200,0.060863110754224986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,3584,0.02716355522473653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,6144,0.5292595757378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,256,0.02163733376397027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,2048,0.03608977794647217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,8192,0.0816026661131117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,128,0.008084444536103142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,1536,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,128,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,128,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,256,0.0647982226477729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,1536,0.03442222211096022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,3584,0.020950223008791607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,512,0.022991999983787537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,1536,0.006133333262470033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,512,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,8192,0.013162666724787818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,12288,0.016795555750528973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,2048,0.44385332531399196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,6144,0.01975555552376641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,3072,0.008779555559158325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,51200,0.05397955576578776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,10240,0.044116444057888456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,768,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,32,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,16384,0.019924443629052904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,256,0.017464000317785475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,128,0.005449777675999536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,51200,0.03193599979082743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,2048,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,3584,0.05779555771085951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,4096,0.03124089042345683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,256,0.03708533445994059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,1024,0.01460622251033783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,16384,0.062373333507113986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,32,0.004547555413511065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,768,0.3686960008409288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,12288,0.14353156089782715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,1024,0.020613332589467365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,2048,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,6144,0.013159111142158508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,64,0.0018053334206342697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,12288,0.1450035572052002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,12288,0.18426933553483751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,8192,0.05490577883190579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,128,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,5120,3.1091253492567272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,51200,0.24176979064941406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,256,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,16384,0.052766223748524986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,7168,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,2048,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,51200,0.9233919779459635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,128,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,6144,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,1536,0.010612444745169746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,12288,0.020055999358495075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,8192,0.04134666588571336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,64,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,3584,0.5243733194139268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,3584,0.4541475507948134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,12288,0.436447991265191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,6144,0.023980443676312763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,512,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,7168,0.14057689242892796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,4096,0.023051554958025616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,64,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,12288,0.010791999598344168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,1024,0.009154666629102495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,32,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,8192,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,7168,0.05439644389682346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,128,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,10240,0.25681066513061523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,1536,0.03873777720663283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,768,0.03792799843682183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,6144,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,65536,0.1515751149919298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,6144,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,10240,0.08733333481682672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,3072,0.08368177546395196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,16384,0.03616177704599168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,256,0.0037333332002162933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,8192,0.02330400049686432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,51200,0.2281857861412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,51200,0.01165600038237042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,32,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,65536,0.4383564525180393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,7168,0.030174222257402208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,65536,0.17816710472106934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,1024,0.01129955550034841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,12288,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,7168,0.08419733577304417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,16384,0.020950223008791607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,51200,0.05424088901943631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,32,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,32,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,1536,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,1536,0.06490400102403429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,768,0.02264444364441766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,1536,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,3584,0.04922399918238322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,64,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,768,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,3072,0.05307111144065857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,5120,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,256,0.005379555539952383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,12288,0.027760000692473516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,65536,0.22775554656982422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,10240,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,65536,3.2130639817979603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,2560,0.01292888902955585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,51200,0.11930399470859104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,51200,0.17031199402279326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,4096,0.008472888833946651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,2560,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,3072,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,7168,0.20084533426496717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,1536,0.0037475555307335327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,8192,0.077565332253774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,3584,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,51200,0.06885155704286364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,1536,0.015953777564896476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,512,0.00814844419558843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,4096,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,512,0.01556622154182858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,7168,0.14334132936265734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,51200,0.10957866907119751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,1024,0.027639999985694885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,1024,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,8192,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,12288,0.0321377780702379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,768,0.15890578428904215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,10240,0.022509333160188463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,1536,0.026712000370025635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,65536,0.07255644268459745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,2560,0.19195111592610678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,7168,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,512,0.009310222334331935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,3072,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,10240,0.04758310980266995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,7168,0.0488684442308214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,256,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,8192,0.051121777958340116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,768,0.017667555146747164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,128,0.02553333342075348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,16384,1.1231742435031467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,8192,0.03784977727466159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,6144,0.02131822208563487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,8192,0.07882755332522921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,51200,0.06063822242948744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,512,0.0053191110491752625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,6144,0.07220800055397882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,128,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,3072,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,12288,0.1818293333053589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,10240,0.019655999210145738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,1536,0.016297777493794758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,4096,0.012637333737479316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,6144,0.040355556541019015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,5120,0.0073315559162033936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,256,0.016175111134847004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,12288,0.06280977858437432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,16384,0.05755822526084053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,51200,0.052910221947564014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,768,0.01973511113060845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,51200,2.114373313056098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,3072,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,12288,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,3584,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,64,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,4096,0.4578248659769694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,65536,0.2457039886050754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,256,0.03503466645876566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,16384,0.024068444967269897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,2560,0.6917484601338705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,2048,0.044575111733542554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,4096,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,1024,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,65536,0.17354399628109404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,512,0.01994755533006456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,12288,0.12012088298797607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,128,0.018636445204416912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,10240,0.015311999453438653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,65536,0.06719555457433064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,12288,0.08234666453467475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,5120,0.09079733159806992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,2560,0.013933334085676404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,32,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,4096,0.025684444440735713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,2560,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,10240,0.021216000119845074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,8192,0.04551022251447042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,8192,0.062485330634646945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,128,0.041045334604051374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,8192,0.01815822223822276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,12288,0.07009866502549914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,64,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,2560,0.027105778455734253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,128,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,128,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,3072,0.023307555251651343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,64,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,3072,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,2048,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,5120,0.02126577827665541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,1536,0.017818666166729398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,65536,0.32076533635457355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,4096,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,16384,0.042857779396904834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,51200,0.0721199976073371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,3072,0.08352000183529323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,3072,0.028107557031843398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,1536,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,2560,0.005177777674463061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,6144,0.012475555141766867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,64,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,2048,0.0042444442709287005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,256,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,6144,0.022467555271254644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,7168,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,128,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,5120,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,12288,0.07825155390633477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,8192,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,10240,0.026400888959566753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,3584,0.037053333388434515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,5120,0.01018666641579734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,2048,0.02516799999607934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,2048,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,12288,0.014073777529928418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,768,0.017872888180944655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,1024,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,4096,0.05254400107595655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,10240,0.0350560016102261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,3584,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,4096,0.0074515557951397365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,768,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,6144,0.038118223349253334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,16384,0.2753804524739583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,6144,0.01613066593805949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,12288,0.014306666122542487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,768,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,256,0.026996443669001263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,8192,0.17652622858683267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,2048,0.004604444321658877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,128,0.09597511423958673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,7168,0.01348888874053955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,65536,1.0862853791978624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,51200,0.10901688867145115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,16384,0.11123022768232559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,32,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,65536,2.0058720906575522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,7168,0.07109689050250582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,2560,0.02323644525474972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,5120,0.019567110472255282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,256,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,512,0.01017066670788659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,51200,0.07058666812049018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,512,0.1853333314259847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,12288,0.5312310854593912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,512,0.015000000596046448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,32,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,1536,0.029321776496039495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,1536,0.012867555850081973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,768,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,3584,0.007367110914654202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,2048,0.021300445000330608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,10240,0.14451111687554255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,16384,0.12075289090474446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,12288,1.7017528745863173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,51200,0.06401155392328899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,65536,2.535596423678928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,128,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,7168,0.018153778380817838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,3072,0.016586666305859882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,4096,0.022335999541812476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,16384,0.026903111073705886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,8192,0.8250008689032661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,12288,0.031855109665128924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,8192,0.012616000241703458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,512,0.004971555537647671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,1024,0.006080889039569431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,3584,0.0340666671593984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,768,0.04232355621125963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,4096,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,3072,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,8192,0.052058666944503784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,2560,0.015863110621770222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,4096,0.0329111119111379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,64,0.014396443963050842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,128,0.006411555740568373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,2048,0.07425422138637967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,10240,0.012630222572220696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,12288,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,2560,0.0148062225845125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,65536,0.2416764365302192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,128,0.018464888135592144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,2560,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,12288,0.035319109757741295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,512,0.015858666764365304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,10240,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,256,0.003370666669474708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,768,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,6144,0.12549066543579102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,128,0.006324444380071428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,7168,0.012495111260149213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,4096,0.035262223747041486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,5120,0.09048266543282403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,768,0.009732444253232744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,128,0.011991110940774282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,2048,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,768,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,3072,0.028880000114440918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,32,0.004176889028814104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,256,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,128,0.014092443717850579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,65536,0.43765865431891543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,4096,0.03514044483502706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,4096,0.025506666964954797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,10240,0.04577599962552389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,256,0.009835555321640438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,5120,0.06759111086527507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,768,0.016140444411171805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,51200,0.02200177808602651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,256,0.016896888613700867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,2048,0.00959911114639706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,8192,0.027818666564093694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,3072,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,12288,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,1024,0.027084444959958393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,6144,0.03184266553984748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,10240,0.05311555663744608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,8192,1.3692515691121419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,768,0.046647111574808754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,8192,0.029301332102881536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,6144,0.026100445124838088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,3072,0.03881422347492642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,2560,0.9990604188707141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,4096,0.047084444099002414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,3584,0.08322044213612874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,1024,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,7168,0.010334221853150262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,256,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,7168,0.01367377738157908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,256,0.011236444115638733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,12288,0.020854221449957952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,4096,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,768,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,256,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,32,0.001960888918903139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,3072,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,2560,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,12288,0.07561955849329631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,3072,0.015986666083335876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,10240,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,6144,0.07990133100085788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,1536,0.01888533267709944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,16384,0.02539644473128849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,5120,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,64,0.01613066593805949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,16384,0.05607110924190945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,5120,0.012461333639091916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,1024,0.017500445246696472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,768,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,51200,0.06030133035447863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,10240,0.05646577808592054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,6144,0.010205333431561789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,3584,0.024099555280473497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,3072,0.028566221396128338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,512,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,7168,0.03472266594568888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,7168,0.031207111146714952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,3072,0.0169840008020401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,5120,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,10240,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,10240,0.23398134443495008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,512,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,1536,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,2560,0.2918204466501872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,51200,0.059664891825781934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,512,0.013146667016877068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,5120,0.009959110783206092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,4096,0.03462310963206821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,5120,0.9666035970052084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,32,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,2560,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,2048,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,51200,0.05128799875577291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,2560,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,256,0.013045333325862885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,128,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,7168,0.036955555280049644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,65536,0.027792887555228338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,768,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,768,0.018182223041852314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,12288,0.03369155526161194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,12288,0.03308799862861633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,256,0.02557955516709222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,3072,0.08871999714109634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,16384,0.031823111904992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,6144,0.06280977858437432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,512,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,32,0.024264888630972967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,16384,0.030326220724317763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,10240,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,51200,0.05608533488379585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,128,0.01035288886891471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,51200,0.27281777064005536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,65536,0.08182755443784925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,32,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,51200,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,512,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,65536,0.056833777162763804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,1536,0.016347555650605094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,65536,0.05001155535380045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,12288,0.018941332896550495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,6144,0.022095110681321885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,51200,3.0895911322699656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,6144,0.2767244444953071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,8192,0.16220000055101183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,4096,0.7515706486172147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,2560,0.009475555684831407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,32,0.010603555374675326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,1024,0.0037831109431054858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,4096,0.09468888574176365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,1024,0.006199999815887875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,12288,0.5213644239637587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,1536,0.016991110311614144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,3072,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,6144,0.10301244258880615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,1536,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,1536,0.009461333354314169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,5120,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,256,0.016904888881577384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,7168,0.01420088940196567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,2560,0.1665279997719659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,4096,0.4805057843526204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,51200,0.06168355544408163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,64,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,5120,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,10240,1.3783999549018011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,16384,0.03902133305867513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,7168,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,3584,0.012539555629094442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,6144,0.012452444268597497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,51200,10.981879340277779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,4096,0.05001599921120537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,768,0.052079998784595065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,3072,0.07360355721579658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,768,0.21333689159817168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,12288,0.02968888812594944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,8192,0.027099554737408955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,3072,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,3072,0.02772444486618042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,512,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,12288,0.06022310919231839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,65536,0.04486844605869717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,16384,0.016695999436908297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,16384,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,7168,0.03237333231502109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,6144,0.026143110460705225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,65536,0.033145778708987765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,2048,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,3072,0.017383111847771537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,3072,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,8192,0.045069333579805165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,1536,0.0114702218108707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,128,0.04639110962549845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,8192,0.09328977929221259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,256,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,256,0.0344817770851983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,1024,0.018953777021831937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,16384,0.37051110797458225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,3584,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,2560,0.022318222456508215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,65536,0.14586222171783447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,10240,0.4167342185974121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,256,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,256,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,51200,0.17090045081244576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,512,0.029321776496039495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,2560,0.007908444437715743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,65536,2.1441805097791886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,256,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,1024,0.007387555307812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,65536,0.057277335060967334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,10240,0.19594132900238037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,2560,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,1536,0.018732445107565988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,768,0.03057510985268487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,12288,0.06554577747980754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,768,0.14971021811167398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,7168,0.08477599753273858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,3584,0.013246222502655454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,16384,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,3072,0.0539973345067766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,65536,0.22032533751593697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,3072,0.021712000171343487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,128,0.015519110692871941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,2048,0.03141422073046366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,1536,0.0053288886944452924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,4096,0.5331608984205458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,6144,0.2927866776784261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,768,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,128,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,7168,0.022107554806603327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,6144,0.024040000306235418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,51200,0.051903999514049955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,256,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,2560,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,6144,0.12342666255103217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,128,0.01510933372709486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,512,0.007371555599901412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,3072,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,7168,0.19951911767323813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,12288,0.02908888790342543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,16384,0.043637331989076406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,12288,0.01885511146651374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,5120,0.04857422245873345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,256,0.013429332938459186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,512,0.014123555686738757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,2048,0.02068000038464864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,51200,0.06174221966001722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,2048,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,4096,0.0284186667866177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,256,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,65536,0.0714320010609097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,256,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,512,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,51200,0.1808373398251004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,7168,0.013243555194801755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,256,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,16384,0.15516711605919734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,2048,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,1536,0.030064000023735896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,128,0.013367999758985309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,6144,0.01867111192809211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,512,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,5120,0.08870933453241985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,768,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,3072,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,7168,0.2163564496570163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,2048,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,2048,0.004876444323195351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,6144,0.018781332506073844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,3072,0.047284444173177086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,1536,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,128,0.10735822386211819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,2048,0.034471111165152654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,64,0.001960888918903139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,8192,0.030213332838482324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,2560,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,512,0.02276355524857839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,16384,2.493493398030599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,65536,0.11527110470665826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,256,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,2560,0.038498666551378034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,12288,0.08648977677027385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,3584,0.047008001142077975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,512,0.01484444406297472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,1024,0.019553777244355943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,768,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,1024,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,3584,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,3584,0.07024621963500977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,16384,0.026497777965333726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,768,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,1536,0.08990044726265801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,7168,0.012500444220172035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,65536,0.08639999892976548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,12288,0.0924524466196696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,16384,0.15975733598073324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,3584,0.010689777632554373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,6144,0.02973777718014187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,1024,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,768,0.012822222378518848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,1024,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,5120,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,16384,0.04011999898486667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,51200,1.8015698326958551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,65536,3.1169200473361545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,3584,0.019156444403860304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,3072,0.046131554577085704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,4096,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,12288,0.017619555195172627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,8192,0.038021332687801786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,10240,0.06146044201321072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,12288,0.09165866507424249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,8192,0.007531555162535773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,512,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,1024,0.0029271110478374693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,768,0.017661333084106445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,8192,0.011720889144473605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,65536,0.06146488587061564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,2560,0.049457778533299766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,16384,0.6662462022569444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,65536,0.16744800408681235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,768,0.003830222205983268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,7168,0.05643910831875271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,2560,0.017504000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,64,0.004380444271696938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,10240,0.02257599929968516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,2560,0.121960891617669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,12288,0.01635199950801002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,64,0.007748444047239091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,512,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,2048,0.025068443682458665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,7168,0.02107644412252638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,6144,0.07094844182332356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,4096,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,1536,0.003882666842805015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,128,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,2560,0.016117332710160148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,2560,0.1904888947804769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,3072,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,1536,0.04668800036112467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,2048,0.009280000295903947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,768,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,256,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,2560,0.25465067227681476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,2048,0.03626755542225308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,3584,0.022720000810093347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,4096,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,12288,0.02252799934811062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,7168,0.06996622350480822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,1024,0.21808711687723795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,51200,3.072506586710612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,6144,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,6144,0.01382844481203291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,16384,0.057005332575904004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,768,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,12288,0.10573777887556289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,1536,0.023197333017985027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,64,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,8192,0.17550844616360137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,512,0.01278222186697854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,6144,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,65536,0.275528007083469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,65536,0.7275964419047037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,3584,0.3141102261013455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,4096,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,8192,3.364335166083442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,1024,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,1024,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,16384,0.008647110727098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,2048,0.05093866586685181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,2048,0.05120799938837687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,4096,0.026081777281231348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,10240,2.000025855170356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,65536,1.3743253284030492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,7168,0.019928889142142404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,6144,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,3072,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,1536,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,7168,0.01937333411640591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,3584,0.01926666663752662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,4096,0.18713955084482828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,3584,0.007792000141408708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,512,0.006004444426960415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,3072,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,16384,0.08225599924723308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,128,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,2560,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,5120,0.019479110836982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,1024,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,3584,0.08718133634991115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,2560,0.09328355391820271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,64,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,768,0.014306666122542487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,65536,0.043455110655890576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,3584,0.009714666340086196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,5120,0.024872889121373493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,512,0.027059555053710938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,256,0.459278212653266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,7168,0.008395555946562026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,512,0.01566133399804433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,65536,0.10832711060841878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,768,0.015792888071801927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,12288,0.01816177699300978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,12288,0.2795635594262017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,64,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,256,0.02014133334159851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,2560,0.014709333578745524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,2560,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,7168,0.11306311024559869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,51200,0.1426737838321262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,768,0.018541332748201158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,2560,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,2560,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,7168,0.015619556109110514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,8192,0.02365155518054962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,8192,0.05083910955323113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,8192,6.334078047010634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,10240,0.02860266632503933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,1024,0.02161600026819441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,768,0.01702844434314304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,6144,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,3584,0.029692444536421034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,65536,0.6494231224060059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,12288,0.8683377901713053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,3584,0.20533511373731825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,8192,0.2172551155090332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,768,0.029487109846538965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,256,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,5120,0.03165155649185181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,5120,0.0120328888297081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,3584,0.032264888286590576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,128,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,2048,0.0210506667693456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,4096,0.09978577825758193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,5120,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,256,0.007028444773621029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,16384,0.45184532801310223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,1024,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,768,0.017200888858901132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,1536,0.03198577629195319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,6144,0.10718933078977798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,3584,0.009398221969604492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,1024,0.0480657782819536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,512,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,256,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,16384,0.04351733459366692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,7168,0.0178195552693473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,1024,0.0075777777367168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,512,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,1536,0.019899555378490023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,8192,0.11875643995073105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,3584,0.02443466583887736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,1536,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,256,0.008539555801285638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,1536,0.0034648889882696998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,8192,0.048262221945656665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,7168,0.07639911439683703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,512,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,4096,0.00775111135509279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,512,0.049229333798090615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,512,0.005692444327804778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,768,0.06171288755204943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,2048,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,1536,0.031501332918802895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,10240,0.03696977761056688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,8192,0.025662221842341956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,4096,0.10063199864493476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,10240,0.1754924456278483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,65536,0.16988889376322427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,12288,0.02402577797571818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,256,0.07852266894446479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,7168,0.06725955671734281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,3072,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,1536,0.005676444619894028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,8192,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,2048,0.029527111185921565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,10240,0.019430221782790292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,8192,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,2560,0.021199110481474135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,4096,0.010626667075686984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,1536,0.018890667292806838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,768,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,7168,0.03356533249219259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,128,0.14763022793663874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,256,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,7168,0.010315555665228102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,8192,0.010631110933091907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,7168,0.06887466377682157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,3072,0.009194666312800514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,6144,0.0711582236819797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,768,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,64,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,3072,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,1024,0.04056355688307021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,3584,0.024744888146718342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,2048,0.019225777851210702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,2560,0.05769689215554131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,1536,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,1024,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,128,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,3584,0.0840559999148051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,16384,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,10240,0.018070222602950204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,4096,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,1536,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,51200,0.1389440033170912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,6144,0.014083556003040738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,256,0.0028853333658642242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,2048,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,1536,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,6144,0.021292444732454088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,2048,0.011624889241324531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,51200,0.2530826727549235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,4096,0.08144622378879122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,10240,0.17861244413587782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,6144,0.0604942242304484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,256,0.0041146667467223275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,768,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,3584,0.025881777207056682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,7168,0.029733333322736952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,256,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,1024,0.020085333122147452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,128,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,1536,0.022639999787012737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,65536,0.16979200310177275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,12288,0.008351999852392409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,512,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,768,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,1536,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,2560,0.208040886455112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,2048,0.05199199914932251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,10240,0.05204533206091987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,3584,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,768,0.01498933302031623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,3584,0.03371822171741062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,4096,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,64,0.0018764444523387486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,65536,0.18124800258212617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,768,0.003527111063400904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,8192,0.008202667037645975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,256,0.013266666895813413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,512,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,128,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,6144,0.01723022262255351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,256,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,7168,0.5313484403822157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,10240,0.0669155584441291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,32,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,2560,0.04053066505326165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,10240,0.019168888529141743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,3072,0.035960889524883695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,16384,0.0864115556081136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,10240,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,2560,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,10240,0.09207199679480658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,65536,0.3134995566474067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,16384,0.14082133769989014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,2560,0.024515555964575872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,768,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,16384,0.3372773329416911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,3072,0.03135644396146139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,2048,0.10755555497275458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,6144,0.012994666894276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,3072,0.014191110928853353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,4096,0.01772711177666982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,256,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,256,0.005976000179847081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,7168,0.10944977733823989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,128,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,12288,0.030682666434182063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,2560,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,5120,0.012636444634861417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,16384,0.03768710957633124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,512,0.025490666429201763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,128,0.015079110860824585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,51200,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,32,0.01538044379817115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,2048,0.011522666447692446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,10240,0.025592888394991558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,4096,0.05721244547102186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,65536,0.6376373502943251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,7168,0.01758666667673323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,65536,1.439716445075141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,2560,0.027895109521018133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,128,0.0064586665895250105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,8192,1.2306932873196073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,7168,0.09803111023373073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,2560,0.02544266647762722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,1024,0.007725333174069722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,10240,0.2348844475216336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,1024,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,12288,0.017844445175594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,12288,0.0505600008699629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,8192,0.039919998910692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,5120,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,5120,0.04289066791534424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,512,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,65536,0.6747288703918457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,256,0.030285331938001845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,512,0.01311288856797748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,3072,0.037696888049443565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,4096,0.5076559914482964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,6144,0.10934311151504517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,8192,0.19444621933831108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,6144,0.011760888828171624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,256,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,2560,0.02509244448608822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,1536,0.05702577696906196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,128,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,5120,0.13401956028408474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,10240,0.0832586685816447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,256,0.022978666755888198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,512,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,16384,0.1943324406941732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,12288,0.024360888534122046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,6144,0.040940443674723305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,4096,0.26527200804816353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,10240,0.11641155348883735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,6144,0.018311111463440787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,3584,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,7168,22.504709879557293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,65536,0.06667733192443848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,4096,0.007789333661397298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,1024,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,1536,0.007000889215204451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,1024,0.02274311085542043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,1536,0.08899822500016953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,2048,0.051479111115137734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,512,0.014808000789748298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,51200,0.06929955879847209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,65536,0.04772888951831394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,6144,0.09115555551317002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,1024,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,2560,0.008380444513426887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,1536,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,3072,0.035359111097123884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,51200,0.30582666397094727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,1536,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,2048,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,7168,0.16537155045403376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,7168,0.031114667654037476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,3584,0.01314311143424776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,7168,0.01610844499535031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,768,0.005762666463851929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,10240,0.012331555287043253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,16384,0.019913777709007263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,51200,0.4884595341152615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,2560,0.015101333459218344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,768,0.012798222402731577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,8192,0.12328444586859809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,768,0.12054577138688828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,768,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,6144,0.012496888637542725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,512,0.02407288882467482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,2048,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,12288,0.12349600262112087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,32,0.0025439999169773525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,128,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,512,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,16384,0.031896889209747314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,64,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,128,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,3072,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,32,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,512,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,3072,0.021552888883484736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,3072,0.020929776959949072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,1024,0.008723555339707268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,64,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,7168,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,3072,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,7168,0.04025955663786994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,1536,0.013175110850069257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,64,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,12288,0.04541688826349047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,64,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,3584,0.14611910449133977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,32,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,12288,0.027506666051016912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,1536,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,1024,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,256,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,64,0.002537777854336633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,7168,0.03067555692460802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,256,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,64,0.026491555902693007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,8192,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,16384,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,64,0.0019013333237833446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,768,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,8192,0.07247377766503228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,8192,0.8110933303833008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,2048,0.07190399699740939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,512,0.03607822126812405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,8192,0.02444355520937178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,16384,0.6575297779507107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,8192,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,64,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,2560,0.1311404440138075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,51200,0.23312799135843912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,65536,0.19643821981218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,512,0.07776800129148695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,64,0.019813333948453266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,4096,0.4605555534362793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,8192,0.04341600007481045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,1536,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,8192,0.0369466675652398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,16384,0.013567999833159976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,6144,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,4096,0.008488888541857401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,2560,0.04037066631846958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,3072,0.013825777504179211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,51200,0.1025173332956102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,12288,0.016162667009565566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,3072,0.026202667090627883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,65536,0.06176977687411838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,2560,0.0064488889442549805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,5120,0.011912888950771756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,6144,0.1865288946363661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,3584,0.27444177203708225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,1024,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,512,0.05644444624582926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,3072,0.02148888839615716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,51200,0.05825777848561605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,51200,0.09617333279715644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,12288,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,65536,0.06411288844214545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,10240,0.13245066006978354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,6144,0.011953777737087674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,16384,0.05423022309939066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,51200,0.10515199767218696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,128,0.015544000599119397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,1536,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,2560,0.03172889020707872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,16384,0.8094106780158149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,6144,0.019564444820086162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,12288,0.10815111133787368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,3072,0.007391999993059371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,512,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,64,0.1318097776836819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,2048,0.0388062232070499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,16384,0.1277351114484999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,1024,0.0070257774657673305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,65536,0.06888444556130303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,4096,0.010888888604111142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,5120,0.01200799975130293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,2560,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,3584,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,6144,0.017761778500345018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,7168,0.07138755586412218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,3072,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,3584,2.1968328687879777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,768,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,5120,0.044608000251981944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,7168,0.058748443921407066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,1024,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,6144,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,5120,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,5120,0.023815999428431194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,7168,0.09453511238098145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,3072,0.011310222248236338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,6144,8.350944519042969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,7168,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,8192,0.04206755426194933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,10240,0.037440889411502414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,2048,0.007188444336255391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,512,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,768,0.005646222167544895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,10240,0.4818719757927789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,7168,0.037900444534089826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,12288,0.046410666571723096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,12288,0.058423108524746366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,6144,0.09959733486175537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,8192,0.0802897744708591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,128,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,64,0.005741333381997214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,3584,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,12288,0.036207109689712524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,512,0.015899555550681222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,64,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,768,0.03491111265288459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,768,0.007048000064161088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,768,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,12288,0.013573333621025085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,51200,0.04638133446375529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,8192,0.12482221921284993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,768,0.021486222743988037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,4096,0.8829822010464139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,16384,0.06490844488143921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,4096,0.22171555625067818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,10240,0.05109688970777723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,128,0.019736889335844252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,768,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,2560,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,16384,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,256,0.007147555549939473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,5120,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,128,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,1536,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,2560,0.020784000555674236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,256,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,512,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,51200,0.0195768889453676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,2048,0.013288000391589271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,12288,0.06338044669893053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,6144,0.0881262222925822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,12288,0.06339644723468356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,2048,0.023353776997990076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,10240,0.11176089445749919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,16384,0.04120355513360765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,51200,0.022622222701708477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,6144,0.11892799536387126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,4096,0.13959377341800266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,32,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,256,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,7168,0.06282933553059895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,3584,0.03517511155870225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,256,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,8192,0.14464799563090006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,64,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,1024,0.026371555195914373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,16384,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,4096,0.1625315613216824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,2560,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,256,0.008571555217107138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,6144,0.15991022851732042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,16384,0.021739555729760066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,3072,0.009543110926946005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,3072,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,10240,0.189465774430169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,3584,0.0474275549252828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,65536,0.2429715527428521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,128,0.013210666676362356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,4096,0.17038933436075845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,256,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,128,0.01851466629240248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,256,0.015632889337009855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,256,0.02420444455411699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,51200,0.05679555733998617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,3584,0.03715466790729099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,10240,0.0952613353729248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,10240,0.021012443635198805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,256,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,3072,0.0074577778577804565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,5120,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,3072,0.2872702280680339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,7168,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,5120,0.04690044456058078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,5120,0.08071288797590467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,3584,0.028823110792371962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,768,0.029793779055277508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,512,0.013891556196742587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,1536,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,16384,0.016187555260128446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,5120,0.14833333757188585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,3072,0.011703111231327057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,768,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,65536,0.12971111138661703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,8192,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,10240,0.01711466742886437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,3072,0.0170168893204795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,256,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,1536,0.03255911005867852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,7168,0.03845599955982632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,6144,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,1536,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,8192,0.044144888718922935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,1024,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,128,0.009454222189055549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,1024,0.012773333324326409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,2560,0.029762665430704754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,256,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,6144,0.026740445031060114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,65536,0.3111502329508464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,51200,0.04610222246911791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,5120,0.09992711411582099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,2560,0.017147555947303772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,65536,4.231753879123263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,768,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,256,0.014505777094099255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,128,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,10240,0.02678933408525255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,10240,0.011361777782440186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,5120,0.03196444445186191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,65536,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,6144,0.01239466667175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,51200,0.07218399975034925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,12288,0.05800444549984402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,5120,0.027425777581002977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,65536,0.030679110023710463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,3072,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,2048,0.028151111470328435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,65536,0.07360266976886325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,8192,0.005030222237110138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,8192,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,1536,0.016823111308945548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,7168,0.026151999831199646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,6144,0.056474665800730385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,6144,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,128,0.015261333849694995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,5120,0.008346666892369589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,8192,0.04901777704556783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,1536,0.04186400108867221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,768,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,4096,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,7168,0.010919111470381418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,10240,0.08999466896057129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,51200,0.8725635740492078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,51200,0.7757066620720757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,768,0.06955199771457247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,2048,0.1733706659740872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,65536,0.1861599948671129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,6144,1.645708401997884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,2560,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,65536,0.13383466667599148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,65536,1.000328911675347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,768,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,10240,0.16227822833591038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,7168,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,5120,0.11674755149417454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,8192,0.011128889189826118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,128,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,8192,0.08896978033913507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,3584,0.08860355615615845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,2048,0.014740443891949125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,2048,0.02095111045572493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,4096,0.04914044340451559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,1536,0.006867555694447623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,8192,0.01108444482088089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,2048,0.02843466732237074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,256,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,10240,0.0642355548010932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,12288,2.2816507551405167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,8192,1.6582541995578344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,10240,0.0311155551009708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,64,0.03307111064592997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,12288,0.28802845213148326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,1024,0.011015999648306104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,4096,0.007191110816266801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,512,0.025519111090236243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,4096,0.025623111261261836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,32,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,2560,0.02476444509294298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,8192,2.429628372192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,3584,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,64,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,3072,0.0950488911734687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,65536,0.1902151107788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,1024,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,10240,0.16740443971421984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,7168,0.013075555364290873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,7168,0.052432888083987765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,1024,0.31750578350490993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,16384,0.11242932743496364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,3072,0.17546489503648546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,65536,0.563610659705268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,65536,0.06425866815778944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,3072,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,16384,0.04904266529613071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,5120,0.220104005601671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,4096,0.017252445220947266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,768,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,512,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,2560,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,512,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,8192,0.05927199787563748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,1024,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,2560,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,128,0.005296000176005893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,512,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,768,0.020236444142129686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,64,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,12288,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,65536,0.19308088885413277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,7168,0.024719999896155462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,3584,0.035734222994910345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,7168,0.013915555344687568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,2560,0.0370417767100864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,65536,0.19919378227657744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,65536,0.5129751099480523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,7168,0.11321422788831924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,64,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,51200,0.23532266087002227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,5120,0.016701334052615695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,3584,0.037286221981048584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,512,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,1536,0.029363556040657893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,16384,0.2175911135143704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,4096,0.08184533649020724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,4096,0.014755555325084262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,2560,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,16384,0.018198221921920776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,5120,0.7252177662319608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,1536,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,12288,0.10693689187367757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,512,0.007167999943097432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,1024,0.018577777677112155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,8192,0.09479200177722508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,4096,0.14198933707343206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,1536,0.013584000368913015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,2560,0.06455110841327243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,512,0.005530666559934616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,3072,0.01830755505296919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,16384,0.049399111005995006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,768,0.009242666264375051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,2560,0.012667555775907306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,51200,0.6781235800849067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,3584,1.0041955312093098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,65536,0.10881866349114312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,1536,0.02453777856296963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,3584,0.010273777776294285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,512,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,3072,0.010664888554149203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,1536,0.005873777800136142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,10240,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,128,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,8192,0.12479466862148708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,1024,0.003866666720973121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,12288,0.04727910955746969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,2560,0.05447377761205038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,768,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,4096,0.017779555585649278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,128,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,3584,0.015040889382362366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,2048,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,6144,0.02176711128817664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,128,0.019139554765489366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,768,0.006330666856633292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,3584,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,512,0.013306666579511432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,1536,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,256,0.024343111448817786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,10240,0.016136000553766888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,2560,0.06325333648257785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,16384,0.8883733219570584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,5120,0.04318311148219638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,4096,0.03507288959291246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,51200,0.059545775254567467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,65536,0.44613154729207355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,64,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,1024,0.03867910967932807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,4096,0.055268443293041654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,768,0.015247111519177755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,3584,0.00853066643079122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,1536,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,51200,0.04914755622545878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,4096,0.04902044600910611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,128,0.005510222166776657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,16384,0.0203493336836497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,51200,4.3901244269476996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,256,0.03306844499376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,12288,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,64,0.0026453334010309647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,768,0.04418755571047465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,7168,0.012947555217478009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,2048,0.19024977419111463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,256,0.0024497778051429326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,2560,0.005994666781690385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,32,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,1536,0.0038408889538711975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,1024,0.02291822267903222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,32,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,8192,0.10184178087446426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,64,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,12288,0.4105679988861084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,128,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,3072,0.008936000367005667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,12288,0.6648604604932997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,2560,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,12288,0.3073066605461968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,10240,0.026376888155937195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,4096,0.06833244694603814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,8192,0.010983111129866706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,3584,0.029343999094433252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,16384,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,512,0.0058151111006736755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,2048,0.004543111142185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,16384,0.04108000132772658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,65536,0.06643911202748616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,4096,0.008605333666006723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,2560,0.021321778496106465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,3584,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,64,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,2048,0.017649778061442904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,1024,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,10240,0.09164710839589436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,8192,0.1871039999855889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,1024,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,51200,0.8202355702718099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,1536,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,6144,0.0359191099802653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,8192,0.3331413269042969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,256,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,10240,0.02586488922437032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,16384,0.055916441811455615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,3072,0.004939555707905027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,1536,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,6144,0.02293155590693156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,3072,0.009230222139093611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,512,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,3584,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,128,0.02277244461907281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,128,0.0033555556502607134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,512,0.17139821582370332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,3584,0.043248001072141856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,7168,0.4124568833245172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,32,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,128,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,65536,0.12050755818684895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,2048,0.00888533310757743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,5120,0.47891913519965273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,768,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,5120,0.021254221598307293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,32,0.1335715585284763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,2560,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,512,0.011411555939250521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,7168,0.027678221464157104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,2560,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,6144,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,7168,0.09506044122907852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,16384,0.019698666201697457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,3584,0.07197511196136475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,2560,0.031884445084465876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,16384,0.020228443874253165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,6144,0.00907022257645925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,65536,0.0912897785504659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,65536,0.3720240063137478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,2560,0.018430221411916945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,128,0.020436444216304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,6144,0.016357334123717416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,512,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,768,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,768,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,3072,0.027448889282014634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,51200,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,256,0.01863466699918111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,256,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,5120,0.00923466682434082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,768,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,768,0.0074986666440963745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,51200,1.1005111270480685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,768,0.02410488824049632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,2560,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,51200,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,3072,0.025807999902301367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,12288,0.4586062431335449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,5120,0.022095110681321885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,2048,0.027941332923041448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,768,0.13383644156985813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,2048,0.02331466641690996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,512,0.024494222468800012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,10240,0.03552444444762336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,512,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,7168,0.01498933302031623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,10240,3.604766421847873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,10240,0.016137777103318107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,1536,0.14651021692487928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,16384,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,8192,0.022555554906527202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,7168,0.09341688950856526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,256,0.05515022079149882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,3072,0.20146577888064912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,3072,0.03666844301753574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,10240,0.039184000757005476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,2560,0.02385955552260081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,12288,0.4475884437561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,256,0.013286222186353473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,6144,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,6144,0.31452978981865776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,7168,0.044769777192009814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,3584,0.5493840111626519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,3584,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,8192,0.06616799698935615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,65536,0.07583644655015734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,32,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,6144,0.017851556340853374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,2048,0.0255724440018336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,7168,0.018163555198245578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,128,0.0550764434867435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,2560,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,7168,0.25928621821933323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,128,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,7168,0.05543733305401272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,12288,0.15878132979075113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,2560,0.05049333307478163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,512,0.008932444784376357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,51200,0.21838843822479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,256,0.02123911182085673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,6144,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,3584,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,768,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,2048,0.010319111247857412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,768,0.020301333732075162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,6144,0.01793066660563151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,12288,0.13410754998524985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,51200,0.10414933496051365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,3072,0.2078204419877794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,2048,0.07486844725079007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,2560,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,128,0.004503111044565837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,1536,0.018903111418088276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,1024,0.02527644402450985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,512,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,3584,0.10761421918869019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,32,0.1496995555029975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,512,0.023676445086797077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,10240,0.029775109555986192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,768,0.07718400160471599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,51200,0.07678755786683825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,16384,0.026328000757429335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,5120,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,5120,0.014665777484575907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,12288,0.021272889441914026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,51200,0.05437866515583462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,10240,0.02943377693494161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,65536,3.6904818216959634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,51200,0.4083911048041449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,3072,0.013281777501106262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,6144,0.1253982252544827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,10240,0.026808889375792608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,64,0.015828443898095023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,2048,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,64,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,4096,0.08381599850124782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,4096,0.03563111027081808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,5120,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,1536,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,2560,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,256,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,3072,0.03902933332655165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,65536,0.06441599792904325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,7168,0.12262845039367676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,7168,0.026944888962639704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,5120,0.03134311238924662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,2560,0.04814133379194471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,3584,0.03576088945070902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,12288,0.5016986529032389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,768,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,4096,0.02714755634466807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,1024,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,1536,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,1024,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,51200,0.12089333269331191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,3072,0.02346311178472307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,1024,0.01606044504377577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,5120,0.15432622697618273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,2560,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,1024,0.17772089110480416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,768,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,4096,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,10240,0.06688088840908475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,4096,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,51200,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,2048,0.016900445024172466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,1536,0.023600889576805964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,16384,0.04269955555597941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,768,0.024502222736676533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,768,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,2560,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,1536,0.011415999796655444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,512,0.015998222761683993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,2048,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,65536,0.0625235570801629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,4096,0.09778400262196858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,7168,0.01146399974822998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,32,0.006550222221348021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,4096,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,1536,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,3584,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,64,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,3072,0.024958221448792353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,51200,0.9667156007554797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,4096,0.02762577765517765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,768,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,6144,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,2560,0.16817333963182238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,5120,0.09784622324837579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,1024,0.017588444881969027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,2048,0.04287999868392944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,3584,0.011158222125636207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,12288,0.060567114088270396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,16384,0.03029600116941664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,16384,0.10817955599890815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,1024,0.16033066643608943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,512,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,128,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,6144,0.019260444574885897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,12288,0.039721777041753135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,512,0.014979556202888489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,4096,0.015212444795502557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,65536,0.40180267228020566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,7168,0.1387315591176351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,4096,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,51200,0.039466665850745306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,6144,0.020798222886191476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,512,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,4096,0.024488889508777197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,1024,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,128,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,512,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,4096,0.040800889333089195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,51200,0.6365182134840224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,512,0.015531554818153381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,8192,0.04713777701059977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,256,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,3584,0.031735112269719444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,8192,0.22523821724785698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,6144,0.03252622154023912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,51200,0.17675111028883192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,2560,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,8192,1.0303404066297743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,256,0.015217777755525378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,128,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,7168,0.07983466651704577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,5120,0.028697778781255085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,12288,0.017215111189418368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,16384,0.013735999663670858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,256,0.00608800020482805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,12288,0.11390133698781331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,7168,0.0427706672085656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,8192,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,1024,0.008651555412345463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,512,0.08031111293368869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,6144,0.08550755845175849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,256,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,2048,0.014708444476127625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,2048,0.01145866678820716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,768,0.02165688905451033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,10240,0.7062391175164117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,16384,0.08579377995596991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,12288,0.6271200180053711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,10240,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,5120,0.030966222286224365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,1536,0.0185253338681327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,6144,0.025394666526052687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,128,0.0063288890653186375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,1024,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,2048,0.015056000815497505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,3584,0.01696266730626424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,16384,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,5120,0.03239466746648153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,64,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,512,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,51200,0.06396177742216322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,10240,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,51200,0.05646755297978719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,3072,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,256,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,16384,0.12905866569942898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,32,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,10240,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,10240,0.01575377749072181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,2560,0.02546577817863888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,3584,0.013707555830478668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,512,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,7168,0.04224088788032532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,10240,0.05233155687650045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,4096,0.015223110715548197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,1536,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,3584,0.007568000091446771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,1024,0.018346667289733887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,1536,0.06965777609083387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,768,0.0028853333658642242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,5120,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,3584,0.00813688917292489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,12288,0.021474666065639917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,65536,0.06799911128150092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,512,0.007227555745177799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,128,0.012106666962305704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,128,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,51200,0.32793066236707896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,4096,0.21823022100660536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,128,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,16384,0.019648889700571697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,7168,0.012095111111799875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,4096,0.004242666479614046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,16384,0.4919279946221246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,128,0.12273866600460476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,3584,0.09723822275797527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,2048,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,1536,0.03216266632080078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,3584,0.007915555602974361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,65536,0.08843021922641331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,256,0.027502222193611994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,768,0.005872000008821487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,32,0.01593866613176134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,1536,0.04425155454211765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,2048,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,512,0.008363555702898238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,3584,0.07644710938135783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,65536,3.4865209791395397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,128,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,3584,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,2560,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,768,0.01973777843846215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,6144,0.026760000321600173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,768,0.12464000119103326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,256,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,2048,0.006758222149478064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,2048,0.004882666799757215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,16384,0.02243822150760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,3584,0.06303111049864027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,128,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,8192,0.010672888822025724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,256,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,32,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,4096,0.007879999776681265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,65536,0.08911555343204075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,768,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,8192,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,768,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,6144,0.010261333651012844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,16384,0.023407111565272015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,4096,0.004257777912749184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,256,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,512,0.011777777638700275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,5120,0.09577600161234538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,65536,0.2871573236253527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,2048,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,768,0.006069333189063602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,1536,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,32,0.004521777646409141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,16384,0.14626667234632704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,6144,0.02149244480662876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,1536,0.035736001200146146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,1536,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,128,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,128,0.013162666724787818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,51200,0.13402044773101807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,256,0.1607146660486857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,7168,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,16384,0.0411155554983351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,5120,0.045331554280387036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,768,0.006704888823959563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,3584,0.034421334664026894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,128,0.2186737722820706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,2048,0.0348639984925588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,51200,0.9077101813422309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,16384,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,1024,0.004524444540341695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,65536,0.5491324530707465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,256,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,4096,0.6111742125617133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,128,0.08688888947168986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,64,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,10240,0.03577955563863119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,3584,0.01718933383623759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,2048,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,51200,0.21766222847832573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,3072,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,6144,0.03836177786191305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,2560,0.012440000143316058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,5120,0.026725333597924974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,8192,0.011861333416567909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,8192,0.08364888694551255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,4096,0.01276622215906779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,12288,0.028892444239722356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,51200,0.425559123357137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,2560,0.007048000064161088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,51200,0.07520800166659884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,10240,0.05651466714011299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,768,0.029623998536003962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,256,0.12748000356886122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,10240,0.019882667395803664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,7168,0.019430221782790292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,1024,0.03280444608794318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,6144,0.06491111384497748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,3072,0.014767110347747803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,10240,0.20982666810353598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,7168,0.011140444212489657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,10240,0.03557066784964667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,128,0.06839911142985027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,2048,0.020192000601026747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,768,0.00870666652917862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,3072,0.008720888859695857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,1024,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,32,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,1536,0.041054222318861223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,3072,0.02199466692076789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,2560,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,2048,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,7168,0.011504000259770287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,3072,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,8192,0.029827555020650227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,65536,0.13852266470591226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,256,0.018630221486091614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,1536,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,12288,0.03436977664629618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,5120,0.03424088822470771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,10240,0.0499893327554067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,3584,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,2560,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,12288,0.013753777576817406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,256,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,12288,0.07595733139250013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,65536,0.1726133293575711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,1024,0.014156444205178155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,512,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,128,0.02293866707219018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,5120,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,10240,0.02370400064521366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,1536,0.012210666305489011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,16384,0.1356168852912055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,2048,0.03765155540572272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,2048,0.04718577861785889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,51200,0.03823733329772949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,3584,0.01426755554146237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,16384,4.91410149468316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,65536,0.9464044570922852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,12288,0.8735831048753527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,512,0.010405333505736457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,4096,0.014923555983437432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,16384,0.02537333303027683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,16384,0.24777777989705405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,3072,0.03361777795685662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,2048,0.06873422198825412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,6144,0.11021155781216091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,2048,0.12268622716267903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,10240,0.23344087600708008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,3584,0.010841777755154504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,64,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,1024,0.010094222095277574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,12288,0.017500445246696472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,65536,1.3779662450154622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,4096,0.1263751056459215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,768,0.042520887321896024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,12288,0.8843706448872884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,512,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,128,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,3584,0.018231999542978074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,1024,0.013598221871587964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,128,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,1536,0.02070844504568312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,5120,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,51200,0.11705422401428223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,12288,0.025172443853484258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,51200,0.3517315652635362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,1024,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,8192,0.03609955641958449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,128,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,12288,0.027121777335802715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,7168,0.011301333705584208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,512,0.04202666547563341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,8192,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,1536,0.012333333492279053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,512,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,768,0.0037875556283526947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,1024,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,1024,0.04916266600290934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,12288,0.023408888114823237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,6144,0.010297777752081553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,51200,3.2356241014268665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,10240,0.03930577635765076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,32,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,7168,0.030930668115615845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,51200,0.0654933320151435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,3072,0.035827553934521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,768,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,1024,0.014331556028789945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,65536,0.05854488743676079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,64,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,512,0.03229422370592753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,5120,0.024653333756658766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,256,0.02344888945420583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,512,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,4096,0.03367022342152066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,12288,0.018809777167108323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,7168,1.2654915915595162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,5120,0.012447111308574677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,8192,0.05028444528579712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,51200,0.5008551279703776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,1536,0.003867555409669876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,10240,0.05414399835798475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,128,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,1536,0.07097244262695312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,2560,0.012513777448071374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,2048,0.3208079867892795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,1536,0.01978755493958791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,7168,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,8192,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,5120,0.02160888910293579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,6144,0.013573333621025085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,768,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,4096,0.053470220830705434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,768,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,2560,0.02269777821169959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,128,0.021254221598307293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,65536,7.743717617458767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,6144,0.11700444751315647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,6144,0.017110221915774874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,1024,0.021493333909246657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,256,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,12288,1.1146044201321073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,10240,0.016736888223224215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,6144,0.04342400034268697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,128,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,768,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,16384,0.038204444779290095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,5120,0.01260088880856832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,65536,0.06885511345333524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,32,0.008037333687146505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,8192,0.07962933513853285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,1024,0.019376888871192932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,10240,12.331548055013021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,6144,0.036646223730511136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,10240,0.027007111244731482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,512,0.022150221798155043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,1536,0.03219111098183526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,32,0.022950222094853718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,1024,0.016220443778567843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,3584,0.011733333269755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,256,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,8192,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,7168,0.0632773306634691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,512,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,4096,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,3072,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,65536,0.20246577262878418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,8192,2.7504285176595054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,4096,0.012466666599114736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,2048,0.022051556242836848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,2560,0.06784711281458537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,7168,0.12178488572438557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,32,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,64,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,1024,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,32,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,65536,0.7757244639926486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,3072,0.029521776570214167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,4096,0.007914666500356462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,7168,0.022442667020691767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,51200,0.29803556866115993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,128,0.006663110935025745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,1024,0.018882667024930317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,3584,0.05673066775004069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,768,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,1024,0.003601777884695265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,1024,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,8192,0.022686221533351477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,3072,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,2560,0.0070364442136552595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,7168,0.16172444820404053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,3584,0.03830133212937249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,512,0.01923111081123352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,6144,1.1160684161716037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,4096,0.04458222124311659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,65536,0.6968400213453504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,65536,7.134167989095052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,256,0.02077866593996684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,128,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,16384,0.9150035646226672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,12288,0.03400355577468872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,2560,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,3584,0.027968890137142603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,65536,0.15835732883877224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,4096,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,10240,0.022648000054889258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,3584,0.06390221913655598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,16384,0.3429493374294705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,7168,0.02088800072669983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,51200,0.16275644302368164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,3072,0.02606933315594991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,1024,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,1536,0.023158222436904907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,65536,0.08136266469955444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,768,0.014279110564125909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,2560,0.04774311184883118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,65536,0.06516622172461616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,768,0.018265777164035372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,3072,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,2048,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,768,0.034894221358829074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,32,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,12288,0.19177599747975668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,6144,0.06051911248101128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,512,0.011363555159833698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,65536,0.610954655541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,512,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,2048,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,7168,0.01640977793269687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,512,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,1024,0.029192888074451025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,1536,0.06158044603135851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,3072,0.012680889003806643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,51200,0.45699644088745117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,6144,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,65536,1.4363252851698134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,1024,0.05135822296142578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,2560,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,3584,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,4096,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,8192,0.0581644442346361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,128,0.027944889333513048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,32,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,64,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,7168,0.050070222881105214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,3072,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,512,0.12542933887905544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,3072,0.010155555274751451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,32,0.03379199902216593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,1536,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,128,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,65536,0.23707911703321669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,7168,0.17569866445329455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,7168,0.029506666792763606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,10240,0.047154668304655284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,51200,0.03654488921165466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,65536,0.034210665358437434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,8192,0.02310666607485877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,3072,0.020220445262061227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,5120,0.010951999988820819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,12288,0.394271108839247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,512,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,6144,0.02780088782310486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,2560,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,5120,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,256,0.02299822204642826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,1536,0.37108800146314835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,64,0.00908622228437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,16384,0.04738310972849528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,1024,0.00351555562681622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,3072,0.01351466692156262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,256,0.00452622233165635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,51200,0.18240177631378174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,1536,0.016912000046836007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,10240,0.14062044355604383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,51200,0.13741066720750597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,3584,0.014041778114106921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,1024,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,128,0.025702221526039973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,65536,0.02514311174551646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,10240,0.13232799371083578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,8192,0.3995724519093831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,10240,0.0184871107339859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,1536,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,128,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,12288,0.1482293340894911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,10240,0.07352533605363634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,512,0.0028968888024489083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,4096,0.023012444376945496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,3072,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,2048,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,16384,0.04484711090723673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,65536,0.5581022368537055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,1024,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,6144,0.015959110524919298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,7168,0.020406223005718656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,6144,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,5120,0.016639999217457242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,5120,0.04701688885688782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,8192,0.07335022423002455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,32,0.008541333178679148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,768,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,768,0.035143110487196184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,1024,0.10900888840357463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,12288,0.12923378414577907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,65536,0.0783368878894382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,65536,2.6121946970621743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,10240,0.19174488385518393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,5120,0.13311555650499132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,64,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,5120,0.04974044362703959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,3584,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,10240,0.02888444397184584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,4096,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,65536,13.393196105957031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,512,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,3072,0.08897155523300171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,128,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,768,0.017865777015686035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,51200,0.049238221512900464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,256,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,32,0.02090488870938619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,2048,0.07506489091449313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,6144,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,65536,0.057144886917538114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,3072,0.021710221966107685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,7168,0.028537776735093858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,6144,0.04929244518280029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,3584,0.030635555585225422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,2560,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,5120,0.011109333071443768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,512,0.16813155015309653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,256,0.005697777701748743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,7168,0.09836888975567287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,65536,0.07619466384251912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,16384,0.9526649051242404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,128,0.1614746650060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,6144,0.03974044322967529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,6144,0.4961546791924371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,5120,0.016591999265882704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,12288,0.058442665470970996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,65536,0.16284444597032335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,1024,0.02243822150760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,128,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,1536,0.023549333214759827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,3584,0.004601777841647466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,7168,0.05288533369700114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,8192,0.05826489130655924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,64,0.006674666785531574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,1536,0.015411555767059326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,256,0.0063564446237352155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,2048,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,5120,0.22842221789889863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,1536,0.013786666923099093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,64,0.0019262221952279408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,16384,0.032279110617107816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,6144,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,6144,0.04610222246911791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,128,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,8192,0.4558711051940918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,12288,0.019902222686343722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,3584,0.008828444613350762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,5120,0.04135377870665656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,65536,15.630606757269966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,512,0.015459555718633862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,10240,0.017507554756270517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,1536,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,2560,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,3584,0.011719110939237805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,12288,0.021312889125612047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,6144,0.01886488828394148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,128,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,2560,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,1024,0.017483555608325534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,10240,0.011714666253990598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,2048,0.052946666876475014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,65536,0.08924800157546997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,12288,0.03455377618471781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,2560,0.012505778008037143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,2560,0.016576889488432143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,10240,0.1783813370598687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,51200,0.4330880112118191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,8192,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,768,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,16384,0.09091644154654609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,1536,0.018797333041826885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,6144,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,16384,0.5292488733927408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,6144,0.020036444067955017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,12288,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,2048,0.019135110908084445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,2560,0.04740355412165324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,64,0.006659555352396435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,512,0.10599022441440159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,5120,0.03551200032234192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,3584,0.007792000141408708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,1536,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,7168,0.02458488941192627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,10240,0.025715554753939312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,512,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,1536,0.0041946665280395085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,8192,0.020652444826232064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,256,0.0057502223385704895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,16384,0.24807821379767525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,51200,0.30816711319817436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,512,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,6144,0.030099554194344416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,16384,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,128,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,12288,0.1708106729719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,12288,0.08786844544940525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,6144,0.030511998467975195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,1024,0.0064239998658498125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,512,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,12288,0.03764177693261041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,768,0.0027057776848475137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,3584,0.01425866617096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,5120,0.009151111046473185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,3584,0.010433778166770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,8192,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,128,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,128,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,6144,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,4096,0.01721600029203627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,7168,0.030704001585642498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,32,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,5120,0.10161511103312175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,3584,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,1024,0.0074373334646224976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,3584,0.054416000843048096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,6144,0.028621332512961492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,64,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,16384,0.04131111171510484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,7168,0.015968888998031616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,16384,0.009198221895429824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,4096,0.7824933263990613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,1536,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,3584,0.6917271084255643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,2048,0.017163554827372234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,2048,0.014563555518786112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,4096,0.08450755808088516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,512,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,7168,0.05602133274078369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,3584,0.23042133119371203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,12288,0.0970746676127116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,256,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,1536,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,2048,0.02149066660139296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,32,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,512,0.015095111396577625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,3584,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,128,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,6144,0.044585777653588184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,128,0.021474666065639917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,10240,0.0845644474029541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,3072,0.01658933361371358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,16384,0.3074711163838704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,128,0.003563555578390757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,768,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,5120,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,4096,0.02420177724626329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,512,0.025831111603313025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,65536,1.1166452831692166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,6144,0.0628382232454088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,512,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,4096,0.008740444150235917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,3072,0.023082666926913794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,51200,1.6031182607014973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,8192,0.02254844374126858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,256,0.016558221644825406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,6144,0.01963911122745938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,4096,0.1401040024227566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,10240,0.10635733604431152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,768,0.006842666616042455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,1536,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,10240,0.16608533594343397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,128,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,16384,0.0251511103577084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,12288,0.09961422284444173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,51200,0.16747466723124185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,5120,0.023312888211674158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,1536,0.022078222698635522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,512,0.022272888157102797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,10240,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,1024,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,12288,0.026937777797381084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,12288,0.20829510688781738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,6144,0.015608888533380298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,1536,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,51200,0.04470666580730014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,128,0.00869511150651508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,4096,0.0057964444988303715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,10240,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,768,0.05070933368470934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,3072,0.019443555010689628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,1024,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,256,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,65536,0.09537955787446763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,16384,4.666565365261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,1024,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,64,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,32,0.01608711149957445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,1024,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,64,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,65536,0.14539021915859646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,51200,0.0455057786570655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,65536,2.3578622606065536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,6144,0.01757244434621599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,3584,0.11635466416676839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,51200,0.5174782011244032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,3072,0.032080001301235624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,256,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,128,0.010490666660997601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,256,0.03137955400678847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,6144,0.08528355757395427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,65536,0.47970220777723527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,3072,0.10425421926710342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,16384,0.05369244350327385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,2048,0.036328001154793635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,7168,1.7151963975694444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,128,0.08032266961203681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,6144,0.012274666792816587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,32,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,1536,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,1536,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,16384,0.1542044480641683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,4096,0.01958044370015462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,32,0.006078222145636876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,7168,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,256,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,3584,0.02425244450569153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,12288,0.020610666937298246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,7168,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,3072,0.020463110672103036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,1024,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,3584,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,2560,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,512,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,2560,0.0304115547074212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,5120,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,768,0.027997334798177082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,32,0.012442666623327466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,8192,0.03837688763936361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,7168,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,12288,0.1363475587632921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,128,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,128,0.015617777903874716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,8192,0.039311998420291476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,3584,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,512,0.014154665999942355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,8192,0.01144444445768992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,65536,0.09087110890282525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,10240,0.03291199935807122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,51200,2.07914310031467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,512,0.014198222094111972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,51200,0.6076684527926975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,5120,0.008849778109126622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,4096,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,1536,0.006751110984219446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,8192,0.02444800072246128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,2048,0.017456889152526855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,16384,0.020344000723626878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,6144,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,7168,0.022043555974960327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,5120,0.015563555889659457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,16384,0.03680444426006741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,3072,0.05983733468585544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,3584,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,3584,0.044812444183561534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,6144,0.015417777829700045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,65536,0.13467378086513943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,512,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,256,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,51200,4.378340403238933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,6144,0.056518223550584584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,3072,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,2048,0.014059555199411182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,16384,0.014010666145218743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,1536,0.01798311041461097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,1024,0.15794132815466985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,16384,0.02328266700108846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,16384,0.025416000021828547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,7168,0.012383999923865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,256,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,6144,0.04179733329349094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,1024,0.006064888917737537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,2048,0.004282666577233208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,3072,0.09505955378214519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,32,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,65536,0.06573955880271064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,2048,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,51200,0.13948978318108454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,256,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,512,0.012783110969596438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,3072,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,768,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,3072,0.036320000886917114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,12288,0.01780533293883006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,128,0.11551377508375381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,6144,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,1024,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,768,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,3584,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,256,0.017234666479958426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,2048,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,65536,0.3864595625135634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,4096,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,16384,0.022280888424979314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,7168,1.182692421807183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,8192,0.01936177743805779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,2560,0.03340355555216471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,256,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,768,0.010794666906197866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,6144,0.04759644468625387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,3072,0.05842133363087972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,1024,0.0169813334941864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,7168,0.027816888358857896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,6144,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,7168,0.008577778107590145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,7168,0.057780444622039795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,4096,0.028892444239722356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,2048,0.005504888792832692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,1024,0.011034666664070554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,64,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,5120,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,10240,0.09333955579333836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,4096,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,65536,0.4837653372022841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,3584,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,8192,1.7256986829969618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,6144,0.3030302259657118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,256,0.09591644340091282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,256,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,5120,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,10240,0.21382399400075278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,256,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,10240,0.012482666307025485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,7168,0.025408888856569927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,2560,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,256,0.016532444291644625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,3072,0.01816444430086348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,768,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,2560,0.07343111435572307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,2048,0.0071937781241205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,128,0.013777777552604675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,4096,0.009823111196359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,1024,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,1024,0.0131217779384719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,6144,0.01886133352915446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,65536,0.20174311267005074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,7168,0.086535996860928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,10240,0.029302222861184016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,3072,0.014589332871966891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,7168,0.10817511214150323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,1536,0.0063182223174307085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,7168,0.1558497746785482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,12288,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,6144,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,2048,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,10240,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,512,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,512,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,7168,0.11335466967688666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,5120,0.03519466519355774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,4096,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,5120,0.04404711061053806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,32,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,768,0.010855110983053843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,4096,0.010557333628336588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,128,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,7168,0.013000888956917657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,12288,0.025981333520677354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,3584,0.012277333272827996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,10240,0.022321777211295232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,64,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,3072,0.023936000135209825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,1536,0.015321777926550971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,16384,0.22787290149264863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,5120,0.045590221881866455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,512,0.015549333559142219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,2048,0.01163733336660597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,16384,0.3604871167076959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,2048,0.014103111293580798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,64,0.02056711084312863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,32,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,128,0.03017510970433553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,6144,1.9551618364122179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,32,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,8192,0.033258666594823204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,2560,0.03103111187616984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,768,0.02368444369898902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,512,0.008610666626029545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,10240,0.03944889042112563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,12288,0.060544888178507485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,51200,0.9934462441338433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,64,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,768,0.009399111072222391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,6144,0.5387066735161675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,1536,0.03920977645450168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,51200,0.028819554381900366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,2048,0.12753600544399685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,51200,86.37150743272569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,12288,0.3709422217475043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,10240,0.07052444749408297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,1024,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,768,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,6144,0.14362933900621203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,256,0.03820266657405429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,12288,0.1281368864907159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,512,0.019476445184813607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,256,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,64,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,768,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,256,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,2048,0.005365333209435145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,51200,0.33116266462537974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,12288,0.12575999895731607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,512,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,3584,0.16680088308122423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,51200,0.05898044506708781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,2048,0.05435555511050754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,512,0.01292799992693795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,12288,0.2843671109941271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,65536,0.03233155608177185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,51200,5.440310160319011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,5120,0.19918399386935762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,10240,0.02838488750987583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,3584,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,64,0.008779555559158325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,65536,0.07996266418033175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,2560,0.007376888559924231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,16384,1.0017111036512587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,256,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,8192,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,16384,0.744999991522895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,6144,0.015593777100245157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,32,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,6144,0.03477333320511712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,10240,0.057124442524380155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,64,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,2048,0.28413955370585126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,1024,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,512,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,65536,1.5057866838243272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,128,0.008544888761308458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,8192,0.07817777660157946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,1536,0.012136889000733694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,4096,0.009568000005351173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,3072,0.006033777776691649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,256,0.013389333254761167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,5120,0.009725333087974125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,1536,0.014501333236694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,6144,0.05953688753975762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,10240,0.02658222284581926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,16384,0.10929777887132432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,5120,0.009975999593734741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,16384,0.07485244671503703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,5120,0.1148151159286499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,12288,0.036446223656336464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,6144,0.10440888669755723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,768,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,64,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,10240,0.015787555111779105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,7168,0.012133333418104382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,8192,0.021670222282409668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,4096,0.010115555591053432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,3584,0.17479199833340117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,32,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,12288,0.32009156545003253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,6144,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,3584,0.043987555636299976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,4096,0.11610489421420628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,2560,0.006879111131032308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,3584,0.025831111603313025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,7168,0.010069333016872406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,1536,0.03189599845144484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,51200,0.6668711238437228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,16384,0.02825511164135403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,10240,0.034889777501424156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,4096,0.039354665411843195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,128,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,2048,0.030817776918411255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,5120,0.027552889453040227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,256,0.0032026666320032547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,512,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,12288,0.020784889658292133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,3584,0.03568444318241543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,3584,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,10240,0.017038222816255357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,5120,0.014752000570297241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,512,0.03604088889227973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,16384,0.011307555768224927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,8192,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,768,0.13276622030470106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,2048,0.02572444412443373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,1024,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,3584,0.5464311175876194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,768,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,12288,0.3067591190338135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,512,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,6144,0.07775644461313884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,16384,0.016653334101041157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,256,0.031530665026770696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,3584,0.020999110407299463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,768,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,16384,0.02165866725974613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,32,0.0021137777302000257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,6144,0.02591466738118066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,512,0.12817955017089844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,8192,0.030418667528364394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,64,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,512,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,1536,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,256,0.01554844445652432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,51200,48.16888427734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,3072,0.016561778055297006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,1536,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,128,0.05679555733998617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,128,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,3072,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,5120,0.09011822276645237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,65536,0.7236506674024793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,1536,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,16384,0.025369778275489807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,65536,0.049477332168155246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,12288,0.24873243437872994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,2560,0.01200177768866221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,4096,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,3072,0.029577778445349798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,3072,0.060976889398362905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,512,0.012624888784355588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,2560,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,16384,0.029507554239696924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,256,0.013895110951529609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,256,0.019928889142142404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,10240,0.03730310996373495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,7168,0.9193795522054037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,12288,0.05773511197831896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,3584,0.048785779211256236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,4096,0.08739022413889568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,65536,0.36700089772542316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,2560,0.08383466800053914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,2560,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,64,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,65536,0.4823919932047526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,256,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,5120,0.023473777704768713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,6144,0.057090666559007436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,6144,0.030016889174779255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,64,0.013901333014170328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,128,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,12288,0.01724444495307075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,4096,0.031581332286198936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,6144,0.011651555697123209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,8192,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,64,0.017857778403494094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,32,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,2048,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,65536,0.41115644243028426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,64,0.005104889058404499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,16384,1.5275057686699762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,1536,0.1777182287640042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,128,0.009150221943855286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,6144,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,256,0.007173333730962541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,256,0.017218665944205392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,8192,0.01791111131509145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,3072,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,128,0.0023022221608294383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,8192,0.023412444525294836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,4096,0.04051288962364197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,1024,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,4096,0.38841867446899414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,1024,0.02788089050187005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,256,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,768,0.27270489268832737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,4096,0.038331554995642766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,12288,0.1378435558742947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,65536,0.22957332928975424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,2048,0.03290133343802558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,7168,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,7168,0.015135111080275642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,5120,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,1024,0.0028373334142896864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,12288,0.045520000987582736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,3072,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,3072,0.007898666792445714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,12288,0.5179840193854438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,65536,0.06730844577153523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,768,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,6144,0.04138311081462436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,8192,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,8192,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,1536,0.04385422335730659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,12288,0.01751733322938283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,51200,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,768,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,512,0.0026266665922270883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,3584,0.10217599736319648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,3584,0.00776711106300354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,6144,0.04029511080847846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,12288,0.015625778171751235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,8192,0.011427555647161273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,10240,0.044601778189341225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,2560,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,256,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,128,0.014949333336618213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,256,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,3584,0.035848001639048256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,768,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,7168,0.014693333042992486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,4096,0.010743111371994019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,2560,0.01995377739270528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,768,0.02387911081314087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,768,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,128,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,2560,0.08780533075332642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,128,0.00351555562681622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,5120,0.013886222408877479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,6144,0.01983644399378035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,12288,0.0524160001013014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,51200,1.1427493625217013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,1536,0.03783377673890855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,5120,0.02501511077086131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,65536,0.2269359959496392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,2560,0.03525066706869338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,8192,0.025758221745491028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,8192,0.37739377551608616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,2048,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,1536,0.015634665886561077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,128,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,2048,0.00925422211488088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,7168,0.04266755448447334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,65536,3.414338641696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,51200,0.052630222505993314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,3584,0.019647111495335896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,7168,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,1536,0.02981333269013299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,12288,0.05682222048441569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,32,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,6144,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,5120,0.014747555057207743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,5120,0.01681155628628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,65536,0.05211199985610115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,51200,0.053174220853381686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,256,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,128,0.01586933268441094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,32,0.02165066699186961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,512,0.05648710992601183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,2560,0.03909511036343045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,16384,2.0719786749945746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,2560,0.21222933133443198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,3584,0.03289777702755398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,51200,0.09679200251897176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,7168,0.09947822491327922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,32,0.12369955910576715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,8192,0.33861690097384983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,16384,0.03195110956827799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,1536,0.09581955273946126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,4096,0.010867555936177572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,128,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,128,0.019485332899623446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,768,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,1536,0.03293955657217238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,256,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,6144,0.010495111346244812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,64,0.008945778012275696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,768,0.002549333290921317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,256,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,12288,0.01716977854569753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,2560,0.07314488622877333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,2560,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,128,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,12288,0.02509333358870612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,65536,1.5308133231268988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,6144,0.06568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,16384,0.016195555528004963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,2048,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,3584,0.03571288784344991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,512,0.027095110880004034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,256,0.1036648882759942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,128,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,16384,0.07409333520465426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,16384,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,256,0.09199288818571304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,3072,0.03152533372243246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,65536,0.15324178006913927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,64,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,256,0.002784888984428512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,4096,0.01661422186427646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,4096,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,32,0.015493333339691162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,128,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,128,0.0018959999498393799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,1536,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,51200,0.11291555563608806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,4096,0.05523288912243313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,1536,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,5120,0.14810666773054335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,5120,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,512,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,128,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,8192,0.022486221459176805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,6144,0.2272204558054606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,7168,0.028904000918070476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,768,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,512,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,16384,0.02222577730814616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,16384,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,2560,0.058957331710391574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,5120,0.03355022271474203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,1536,0.008178667061858708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,3072,0.008912000391218398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,768,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,5120,0.02254044512907664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,1536,0.028727110889222886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,5120,0.013334222137928009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,5120,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,2048,0.3011351161532932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,5120,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,65536,1.08733336130778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,64,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,3072,0.09921866655349731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,3072,0.038613332642449275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,1536,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,3584,0.04308088951640659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,65536,0.06511200136608548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,7168,0.009879111415810054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,512,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,7168,0.03215822246339586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,7168,0.004925333377387789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,128,0.019655999210145738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,256,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,5120,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,6144,0.09655555751588608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,1536,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,32,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,3072,0.006670222100284364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,10240,0.014418666561444601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,3072,0.05667022201750013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,65536,0.13096622625986734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,12288,0.10593866639667088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,12288,0.059861335489484996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,2048,0.00453599997692638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,64,0.012784889174832238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,1536,0.01965777741538154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,16384,0.08367733160654704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,16384,0.4229297902848985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,512,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,64,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,2560,0.021624000536070928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,256,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,10240,0.05810666746563382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,5120,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,1536,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,1024,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,6144,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,256,0.031195554468366835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,16384,0.0733226670159234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,6144,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,2560,0.060024003187815346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,3072,0.01276622215906779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,4096,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,2048,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,65536,0.5696141984727647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,64,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,4096,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,8192,0.012979555461141797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,16384,0.034035556846194796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,2560,0.0642755561404758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,128,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,256,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,7168,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,6144,0.06107644240061442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,10240,0.053454223606321544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,64,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,2048,0.03087110983000861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,16384,0.13320355945163304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,1536,0.01348088930050532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,4096,0.03589422172970242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,128,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,51200,0.02420088814364539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,10240,0.006888888776302338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,10240,0.008656000097592672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,8192,0.15429866313934326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,768,0.006102222121424145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,4096,0.016371554798550077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,1024,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,6144,0.010668444136778513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,51200,0.5509875615437826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,8192,0.014720888601409065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,1536,0.017610667480362784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,4096,0.011760888828171624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,7168,0.12843999597761366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,3584,0.03578311204910278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,512,0.0566213329633077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,16384,0.042429334587521024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,768,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,1024,0.06535733408398099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,128,0.02029600077205234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,10240,0.041914665036731295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,1024,0.0329022208849589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,51200,0.05022133390108744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,768,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,256,0.015280889140235053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,65536,0.0935751133494907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,3072,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,3584,3.5649325052897134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,1024,0.015211555692884656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,65536,0.06456977791256376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,2048,0.004551110996140374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,3584,0.05127733283572727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,2560,0.2125422159830729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,51200,0.06723555591371325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,2048,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,128,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,64,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,10240,0.01497333414024777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,3584,0.03318666749530368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,768,0.01460800071557363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,2048,0.2212373283174303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,6144,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,51200,0.31082222196790904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,7168,0.011599111060301462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,128,0.02160888910293579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,512,0.015204444527626038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,51200,0.36522934171888566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,51200,0.13999556170569524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,3072,0.02018044392267863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,7168,0.02776355544726054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,3584,0.052842666705449425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,32,0.002208000048995018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,256,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,12288,1.5898862414889867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,51200,0.20885954962836373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,10240,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,12288,0.23262490166558158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,512,0.033837334977255926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,8192,2.2802132500542536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,2560,0.024866667058732774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,3584,0.06431821982065837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,128,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,7168,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,2560,0.04843644301096598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,256,0.03307911091380649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,2560,0.0503386656443278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,3072,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,6144,0.03136444422933791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,4096,0.012320000264379712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,2048,0.06578666634029813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,16384,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,2048,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,51200,0.06791466474533081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,65536,0.10622221893734402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,10240,0.3621795442369249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,8192,0.14917066362169054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,6144,0.05264444483651055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,8192,0.023638221952650283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,6144,0.04404711061053806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,256,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,2048,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,1536,0.056417776478661426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,5120,0.024107555548350017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,4096,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,3072,0.020643555455737643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,2048,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,32,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,12288,0.028093334701326158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,512,0.027431999643643696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,256,0.020774222082561918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,768,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,5120,0.024511999554104272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,768,0.04401333464516533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,6144,0.020941333638297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,768,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,16384,0.06839022371504042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,4096,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,6144,0.5491520033942329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,4096,0.008598222500748104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,12288,0.20044444666968453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,256,0.01129155523247189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,128,0.0993022190199958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,6144,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,8192,2.442047119140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,12288,0.38598132133483887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,8192,0.05786755349900988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,3072,0.017298666967286002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,2560,0.07746133539411756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,51200,0.05789155430263943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,51200,16.716808742947048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,6144,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,2048,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,51200,0.477143128712972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,3072,0.09979377852545844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,6144,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,3072,0.12837421894073486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,2048,0.036728888750076294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,3584,0.039671109782324895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,768,0.02017244365480211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,10240,0.043397333886888295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,5120,0.037816888756222196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,8192,0.014923555983437432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,10240,0.018429333964983623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,6144,0.03110044532352024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,65536,0.06263822317123413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,5120,0.024313777685165405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,2048,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,16384,0.03919644488228692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,51200,0.07319821914037068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,4096,0.10182666778564453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,1536,0.021235555410385132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,1536,0.0036142224238978494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,3584,0.009593778186374241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,5120,0.008806222014957005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,512,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,2048,0.03157155712445577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,256,0.017869333426157635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,512,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,8192,0.33752889103359646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,65536,1.3398426903618708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,16384,0.038485334979163274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,4096,0.028280887338850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,65536,0.04925510949558682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,65536,0.08899911244710286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,256,0.02095555596881443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,8192,0.01475911173555586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,4096,0.0236453331179089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,16384,0.09634311331642999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,1536,0.03671911027696397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,6144,0.01792622274822659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,512,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,51200,0.04423200090726217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,12288,0.04263910982343885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,1024,0.021469333105617102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,768,0.018896889355447557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,128,0.014157333307796054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,7168,0.162191112836202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,1024,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,6144,0.12106755044725205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,768,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,4096,0.10130133231480916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,10240,0.19918044408162436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,1536,0.18833511405520967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,256,0.014911999305089315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,1024,0.05328088998794556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,5120,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,3072,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,8192,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,65536,0.05943200323316786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,768,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,8192,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,6144,0.121869338883294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,12288,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,3072,0.007673777639865875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,1024,0.02092355489730835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,51200,0.05758488840527005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,10240,0.28038843472798664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,5120,0.40752267837524414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,1024,0.06068355507320828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,8192,0.059912886884477407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,65536,0.2535075611538357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,16384,0.07865066660775079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,3584,0.025196444657113817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,64,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,32,0.04010222355524699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,10240,0.03920088873969184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,512,0.030209776428010728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,12288,0.28903200891282826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,5120,0.043848001294665866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,512,0.046075556013319224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,8192,0.047664887375301786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,6144,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,51200,0.8084524472554525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,2560,0.02360533343421088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,3072,0.012972444295883179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,3072,0.01218844453493754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,12288,0.0365937782658471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,12288,0.016375111209021676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,3072,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,65536,1.6626773410373266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,5120,0.011718221836619906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,4096,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,51200,0.13280177116394043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,6144,0.6641973389519585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,7168,0.018279999494552612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,65536,0.18799732791052925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,5120,0.02234933276971181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,512,0.011343999869293638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,7168,0.16583377785152859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,16384,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,51200,0.0954444474644131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,3072,0.05921600262324015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,512,0.0063697778516345555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,6144,0.0143857780430052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,7168,0.12479289372762044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,16384,0.06683822472890218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,2048,0.011564444336626263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,2048,0.019959110352728102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,128,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,512,0.015328889091809591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,512,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,51200,0.055712891949547656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,10240,0.03637689020898607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,8192,0.03207822309599982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,3584,0.031552000178231135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,4096,0.008261333737108443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,12288,0.01757244434621599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,768,0.017539555827776592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,16384,0.06895822286605835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,1536,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,65536,0.2276079919603136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,4096,0.033951110310024686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,768,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,3584,0.03637422124544779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,32,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,51200,0.05995021926032173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,6144,0.027366222606764898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,1536,0.03222222129503886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,768,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,3584,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,3072,0.005522666705979242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,6144,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,51200,0.10290933317608303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,1024,0.020503110355801053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,3072,0.005422222117582957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,7168,0.01459822224246131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,2560,0.038443555434544883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,6144,0.9464471605088977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,2560,0.03808088766203986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,512,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,2560,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,768,0.014526221487257214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,64,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,16384,0.3477768898010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,256,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,4096,0.12452267275916205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,16384,0.24148178100585938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,2048,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,6144,0.018416888184017606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,7168,0.010656000011497073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,10240,0.045385777950286865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,128,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,3584,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,128,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,6144,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,12288,0.05082755618625217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,12288,0.01344622257683012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,8192,1.9244435628255208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,128,0.02714311083157857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,16384,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,10240,0.12551466623942056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,16384,0.08000355296664767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,4096,0.03887910975350274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,1536,0.004661333229806689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,2560,0.12240533034006755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,4096,0.02243288854757945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,6144,0.06003733476003011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,64,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,51200,0.15516177813212076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,8192,0.06587644418080647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,768,0.002567111204067866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,3072,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,16384,0.05318577753172981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,512,0.036944889360004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,51200,0.048918220731947154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,10240,0.025117332736651104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,2048,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,16384,0.05468177795410156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,7168,0.046797331836488515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,51200,0.05615555577807956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,2560,0.07362044519848294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,3072,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,2560,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,768,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,1536,0.009877333210574256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,10240,0.029575109481811523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,3072,0.08357777860429551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,51200,0.3776071071624756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,1536,0.05993688768810696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,10240,0.1046924458609687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,32,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,8192,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,32,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,7168,0.026086222794320848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,128,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,2048,0.011680000358157687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,2048,0.028099556763966877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,51200,0.15656533506181505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,8192,0.011361777782440186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,7168,0.023012444376945496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,2048,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,4096,0.050272001160515674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,6144,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,7168,0.046909332275390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,4096,0.013129777378506131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,6144,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,32,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,10240,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,512,0.033209777540630765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,128,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,128,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,1536,0.006833777659469181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,5120,0.21474756134880912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,6144,0.010276444256305695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,12288,2.430146747165256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,6144,0.011337777806652917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,6144,1.0172292921278212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,768,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,128,0.03310222095913357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,1024,0.012465777496496836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,6144,0.3256186644236247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,16384,0.2807377709282769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,65536,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,3584,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,16384,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,12288,0.031994667318132185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,1024,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,6144,0.019623999794324238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,10240,0.014950222439236112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,5120,0.04284444451332092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,7168,0.30259821150037974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,8192,0.051961776283052236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,1536,0.010308444499969482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,1536,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,8192,0.02866666515668233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,8192,0.033688889609442815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,128,0.007821333077218797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,12288,2.3660852644178605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,10240,0.9103253682454427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,256,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,1536,0.008089777496125963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,3072,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,7168,0.15928267108069524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,3584,0.029990222718980577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,6144,0.01981511049800449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,2560,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,1536,0.026885333988401625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,8192,0.07860177755355835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,768,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,8192,1.7857564290364583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,2560,0.06293866369459364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,65536,1.1469075944688585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,7168,0.012468444804350534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,4096,0.027434666951497395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,768,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,6144,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,128,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,2560,0.016521778371598985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,1024,0.04095910986264547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,16384,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,2560,0.025050666597154405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,3072,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,512,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,2560,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,64,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,768,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,6144,0.25194043583340114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,65536,6.798116472032334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,1024,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,51200,0.12022666136423747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,1536,0.017143110434214275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,4096,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,2048,0.013454222016864352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,6144,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,16384,0.023578666978412207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,3584,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,10240,0.016009777784347534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,65536,0.07243021991517809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,6144,0.018544889158672757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,5120,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,65536,1.1230612860785592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,256,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,512,0.00775822252035141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,7168,0.014105778601434497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,3072,0.015597333510716757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,5120,0.012258666257063547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,7168,0.027320888307359483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,51200,0.13952799638112387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,8192,0.33396445380316836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,5120,0.03333244389957852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,16384,0.027289777994155884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,10240,0.015656888484954834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,10240,0.22328623135884604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,2560,0.011284444067213269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,12288,0.09938489066229926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,51200,0.11067910989125569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,512,0.019549333386951022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,2560,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,3072,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,6144,0.009257777697510189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,1024,0.208168003294203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,32,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,128,0.013100444442696042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,768,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,768,0.00903555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,1536,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,16384,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,51200,0.0530373321639167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,2048,0.006079111248254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,32,0.005730666634109285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,5120,0.12336266040802002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,5120,0.10359644227557713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,12288,0.013678222066826291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,4096,0.05439111259248522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,12288,0.030525333351559106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,768,0.02932000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,10240,0.182532442940606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,3072,0.06662044260236952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,128,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,1024,0.013047110703256396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,512,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,5120,0.08509422010845608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,2560,0.031023999055226643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,1024,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,3072,0.026449778013759192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,7168,0.01774222155412038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,256,0.01273155543539259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,512,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,6144,0.22073066234588623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,12288,0.004582222137186262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,6144,0.028802666399214003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,51200,0.2806648943159315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,32,0.008226667013433244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,2560,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,16384,0.024078221784697637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,65536,0.31364888615078396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,16384,0.11311999956766765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,2048,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,4096,0.03749777873357137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,3584,0.03641688823699951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,51200,1.3608417510986328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,3072,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,1536,0.21351288424597847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,1536,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,2048,0.004534222185611725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,51200,0.07151822249094646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,51200,0.05974488788180881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,256,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,16384,0.02472266720400916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,2560,0.030290666553709242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,3072,0.004192000048028098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,12288,0.09594044420454238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,2048,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,10240,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,6144,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,768,0.021362667282422382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,32,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,128,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,65536,0.08421866761313544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,32,0.01277599980433782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,51200,0.06081333425309923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,512,0.18777067131466338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,5120,0.009126221968068017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,512,0.016626665989557903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,2560,0.07910577456156413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,1024,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,6144,0.05522488885455661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,32,0.1642284393310547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,51200,0.05063822203212314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,7168,0.2521795431772868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,3072,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,8192,0.016376000311639573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,4096,0.030908445517222088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,256,0.015003555350833468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,128,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,7168,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,4096,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,8192,0.24644621213277182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,16384,0.44327733251783585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,10240,0.17742933167351616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,2560,0.07141600052515666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,256,0.03307822346687317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,4096,0.1095199982325236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,16384,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,8192,0.2525804572635227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,64,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,768,0.02070755594306522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,512,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,10240,0.052377776967154614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,6144,0.08573689063390096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,4096,0.020294222566816542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,4096,0.025359110699759588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,512,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,65536,0.49049064848158097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,51200,0.06424533658557467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,512,0.03154933452606201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,2048,0.06222933530807495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,768,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,256,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,4096,0.12567378415001765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,12288,0.019334221879641216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,1536,0.03967377874586318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,51200,0.01955200069480472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,256,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,10240,0.03183644347720676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,1536,0.010584889186753167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,1024,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,65536,0.07044711377885607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,10240,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,65536,7.93009270562066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,128,0.026124444272783067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,1024,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,7168,1.4758941862318251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,10240,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,256,0.17840445041656494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,128,0.015616888801256815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,768,0.006045333213276333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,4096,0.007749333149856991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,16384,0.13660177919599745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,768,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,32,0.0019066666977273093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,12288,0.02961155441072252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,128,0.006476444502671559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,1024,0.009816889133718278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,8192,0.02693066663212246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,16384,0.0700862209002177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,64,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,10240,0.10298399792777167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,32,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,1536,0.09355022509892781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,12288,0.1818551090028551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,6144,0.041999111572901406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,2048,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,7168,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,128,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,5120,0.024674667252434626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,65536,1.5750648710462782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,512,0.059226665231916636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,6144,0.053119109736548535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,1536,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,256,0.017532444662517972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,64,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,5120,0.007363555332024892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,10240,0.07999289035797119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,3072,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,512,0.007695111135641734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,2560,0.01571822166442871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,8192,0.046642667717403836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,2048,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,2560,0.034689777427249484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,2560,0.034986668162875705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,128,0.015036443869272867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,51200,0.23063288794623482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,6144,0.018065777089860704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,7168,0.9323946634928385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,5120,0.05044088761011759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,2048,0.17292443911234537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,8192,0.02272355556488037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,5120,0.016906667086813185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,512,0.015331554743978711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,10240,0.3747822178734674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,1024,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,16384,0.09518133269415961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,768,0.018458666072951425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,3584,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,16384,0.6660106447007921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,2560,0.012321777641773224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,5120,0.08968177768919204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,65536,0.07489422294828627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,4096,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,768,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,65536,0.06463111109203763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,2048,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,512,0.0038968887594011095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,7168,0.07542666461732653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,128,0.014705777168273926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,512,0.16190489133199057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,256,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,2560,0.037132445308897234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,3584,0.015098666151364645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,16384,0.02562933332390255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,10240,0.09351555506388347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,10240,0.07477421893013848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,1024,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,6144,0.013669333524174161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,3072,0.08500711123148601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,3072,0.07000355588065253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,768,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,12288,0.07489689191182454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,16384,0.04830933279461331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,2560,0.07866488562689887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,1536,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,768,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,51200,0.13464889261457655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,65536,0.09029333459006415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,3584,0.4214462174309625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,1536,0.01648977730009291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,8192,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,128,0.08308000034756131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,512,0.02149066660139296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,512,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,7168,0.04130666785769992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,512,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,3072,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,1536,0.017881777551439073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,51200,0.06424799892637464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,3584,0.016068443655967712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,4096,0.3523288832770453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,128,0.025482666161325242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,2048,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,3584,0.010102222363154093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,6144,0.03152533372243246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,6144,0.03695644603835212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,6144,0.12366488244798447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,2560,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,7168,0.06915199756622314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,4096,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,5120,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,256,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,128,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,12288,0.012967111335860359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,1536,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,2560,0.02239111065864563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,6144,0.025672889418072168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,1024,0.01277511070171992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,2560,0.07438933187060885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,5120,0.01793599956565433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,2560,0.010830221904648675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,2048,0.011704000333944956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,16384,0.13932711548275417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,512,0.01351111133893331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,5120,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,512,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,3072,0.43569066789415145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,7168,0.0810497800509135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,12288,0.013768889009952545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,5120,0.026496888862715826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,1024,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,768,0.06921511226230197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,5120,0.01956533392270406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,2048,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,2560,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,7168,0.05999288956324259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,65536,8.602739969889322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,768,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,3072,0.020809777908855014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,4096,0.02788533435927497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,3072,0.1083937750922309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,64,0.024862221545643274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,65536,0.7185146543714734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,16384,0.1646275520324707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,16384,0.020336889558368258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,51200,0.048060443666246205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,6144,0.038600001070234514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,4096,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,768,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,8192,0.12287822034623887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,768,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,5120,0.009038222332795462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,2048,0.05029155479537117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,128,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,16384,0.053247110711203686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,512,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,128,0.013804444836245643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,2048,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,512,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,16384,0.13988533284929064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,4096,0.016514667206340365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,512,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,10240,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,768,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,1024,0.01942399972014957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,6144,0.027859555350409612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,6144,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,256,0.014941333068741692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,65536,0.05748889181349012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,7168,0.04218488931655884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,51200,0.0540924436516232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,1536,0.020809777908855014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,12288,0.020484444167878892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,7168,0.04527466826968723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,3072,0.00792711145348019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,3072,0.6904862191942004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,4096,0.027452444036801655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,4096,0.09722310966915554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,2560,0.01424711114830441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,65536,0.06928177674611409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,4096,0.02247911029391819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,2048,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,4096,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,32,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,2048,0.3925200038486057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,5120,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,65536,0.0713697804345025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,3072,0.021167111065652635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,1024,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,8192,0.0912640028529697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,2048,0.10096177789900039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,8192,0.1067768865161472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,128,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,2048,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,7168,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,32,0.03809955716133118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,128,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,8192,0.03353777858946059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,1024,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,512,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,51200,0.06243466668658786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,64,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,12288,0.031304889255099826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,1024,0.006124444305896759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,10240,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,51200,0.05690222316318088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,1536,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,256,0.035952889257007174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,2048,0.030101332399580214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,2560,0.051627556482950844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,512,0.01923111081123352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,1024,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,3584,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,256,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,3072,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,2560,0.6540657679239908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,2048,0.09508622354931301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,7168,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,16384,0.3653137683868408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,7168,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,51200,0.04186311033036974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,768,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,512,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,5120,0.02568799919552273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,32,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,51200,0.0511084430747562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,128,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,1536,0.20560533470577666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,12288,0.39940889676411945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,2048,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,5120,0.028216001060273912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,51200,0.418287992477417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,8192,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,10240,0.032221333848105535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,256,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,4096,0.06999644305970934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,256,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,12288,0.07302755779690213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,1536,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,7168,0.039224889543321394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,64,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,16384,0.14344444539811876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,8192,0.04195911023351881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,2560,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,16384,0.009498666558000777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,512,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,7168,0.6431786749098036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,256,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,6144,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,5120,0.41137154897054035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,12288,0.02054933375782437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,128,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,12288,0.13417689005533853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,8192,0.07605066564348009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,16384,0.16083289517296687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,51200,0.30246488253275555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,512,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,3584,0.02110222147570716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,2560,0.009684444301658207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,51200,1.066284391615126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,8192,0.06033777528338962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,65536,0.07220177517996894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,12288,0.04617422156863742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,16384,0.19602578216128877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,3072,0.07480888896518283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,12288,0.014723555909262763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,2048,0.15473600228627524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,2560,0.022857778602176245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,10240,0.04421866602367825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,65536,0.08513244655397202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,128,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,512,0.008443555898136562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,6144,0.032567110326555036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,2048,0.006783111227883234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,7168,0.06362222300635444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,7168,0.150309337510003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,16384,0.08991822269227769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,3072,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,51200,0.3850062158372667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,512,0.035764445861180626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,16384,0.07414755556318495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,3072,0.012479999827014076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,3584,0.021623111433453027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,768,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,768,0.017836444907718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,3584,0.021060445242457922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,1024,0.25416000684102374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,64,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,65536,0.07857955826653375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,128,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,768,0.003552000141806073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,256,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,768,0.020024889045291476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,768,0.02036888897418976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,1536,0.0038746665749284956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,8192,0.4098657766977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,2048,0.028862221373452082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,16384,0.07941777838601007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,51200,0.179439107577006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,8192,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,6144,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,16384,0.42342578040228945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,3584,0.03360177742110358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,7168,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,3072,0.10843555794821845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,2560,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,16384,0.06760355499055651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,10240,0.11374844445122613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,1536,0.057922667927212186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,768,0.01995822290579478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,256,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,512,0.009411555197503831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,1536,0.018258665998776753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,1024,0.03264710969395108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,256,0.015551111764378019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,768,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,10240,0.011901333100265928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,12288,0.047833777136272855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,512,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,3584,0.017696888910399545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,3072,0.01738666660255856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,256,0.005642666584915585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,12288,0.014697778556081982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,2048,0.010261333651012844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,8192,0.02490844494766659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,5120,0.08875733613967896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,3072,0.24007911152309844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,768,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,1536,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,2560,0.03425688876046075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,768,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,256,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,3072,0.13992622163560656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,1536,0.4720960193210178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,2560,0.09573600027296279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,128,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,1024,0.022264889544910852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,256,0.002692444456948174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,3584,0.012847999731699625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,65536,0.2811342345343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,6144,0.11078044441011216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,10240,0.0121742222044203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,12288,0.08227111233605279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,512,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,1024,0.00983199973901113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,128,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,1536,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,5120,0.02158044444190131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,51200,0.06028088596132067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,768,0.01888177792231242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,3072,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,1536,0.017870222528775532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,64,0.003650666524966558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,512,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,1024,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,2560,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,512,0.0199608885579639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,2560,0.017064000169436138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,8192,0.09694044457541572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,5120,0.02477600011560652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,12288,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,256,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,10240,0.035495999786588885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,3072,0.03143644332885742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,128,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,65536,0.08215555879804823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,2048,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,2048,0.03327466713057624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,4096,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,256,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,512,0.006519111080302133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,65536,0.1467777755525377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,5120,0.021735999319288466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,3584,0.006367110957702001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,1536,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,1536,0.035707556539111666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,65536,0.3282133208380805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,768,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,6144,0.011215999722480774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,7168,0.021755556265513103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,256,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,768,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,5120,0.009345778160625035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,4096,0.019864888654814828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,3584,0.020561777883105807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,6144,0.019721777902709115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,64,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,1024,0.016120000018013846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,1024,0.008879111044936711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,8192,0.08866666422949897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,2560,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,2560,0.022610666023360357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,51200,2.235033882988824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,7168,0.2565964327918159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,3072,0.011506666739781698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,1536,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,1536,0.017575111654069688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,1024,0.02666933337847392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,7168,1.9292666117350261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,12288,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,3072,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,1536,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,512,0.0032764443506797156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,1024,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,5120,0.049610667758517794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,8192,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,2048,0.269561767578125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,512,0.09836355845133464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,256,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,4096,0.12470755312177871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,8192,0.03261333372857835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,6144,0.010672888822025724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,65536,0.1385600037044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,6144,0.12211555904812282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,7168,0.10161422358618842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,256,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,3584,0.02234222160445319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,2560,0.023797333240509033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,512,0.006747555401590135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,512,0.005686222265164058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,1536,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,768,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,16384,1.5218675401475694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,128,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,1536,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,8192,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,10240,0.02552088929547204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,7168,0.4474124378628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,8192,0.12628889083862305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,5120,0.010107555323176915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,65536,0.04788711004787021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,16384,0.4827440049913194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,1536,0.019834667444229126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,7168,0.139101333088345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,128,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,65536,0.01575999955336253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,4096,0.030637333790461223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,2560,0.047651555803087026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,2560,0.3338337739308675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,1536,0.28002844916449654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,1024,0.006847111301289664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,1024,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,16384,0.027998222245110407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,128,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,1024,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,6144,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,4096,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,1024,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,8192,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,7168,0.10772533549202813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,12288,0.3064631091223823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,6144,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,1536,0.07011377811431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,3072,0.06063199705547757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,256,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,7168,0.3014142248365614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,4096,0.039353777964909874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,256,0.007556444240940942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,3584,0.013674666484196981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,8192,0.07089689042833117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,768,0.017852443787786696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,256,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,65536,0.24656356705559623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,2048,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,5120,0.036751998795403376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,128,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,3072,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,51200,0.03346933258904351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,8192,0.04099022348721822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,3584,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,6144,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,32,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,32,0.019273777802785236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,7168,0.07681866486867268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,1536,0.003900444342030419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,768,0.061679999033610024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,4096,0.02296888828277588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,32,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,3584,0.061956445376078285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,12288,0.05338044299019707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,1536,0.009173333644866943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,3072,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,5120,0.10787022113800049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,16384,0.358969767888387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,5120,0.049850665860705905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,6144,0.007430222299363878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,7168,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,2560,0.0210479994614919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,8192,0.023337778117921617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,7168,0.020207999481095206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,4096,0.01794666714138455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,65536,3.8962622748480906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,5120,0.010953777366214328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,16384,0.020264888803164165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,2560,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,10240,0.02765688962406582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,5120,0.02750933335887061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,51200,7.335266960991754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,1024,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,6144,0.004858666823969947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,12288,0.24815466668870714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,256,0.07622133360968696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,16384,0.08572977781295776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,256,0.012169777519173093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,7168,0.2094506687588162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,2048,0.004530666602982415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,7168,0.043756445248921715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,51200,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,5120,0.02222844461599986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,51200,0.17708444595336914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,3072,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,10240,0.027962668074501887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,3584,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,3072,0.19495556089613172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,2048,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,51200,0.10221333636177911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,6144,0.07259377506044176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,256,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,12288,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,3072,0.22341865963406035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,1024,0.0035982223020659555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,2560,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,768,0.005623111294375525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,128,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,5120,0.0762595534324646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,7168,0.018256889449225534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,2560,0.029116445117526587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,65536,0.13111555576324463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,8192,0.014915555715560913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,1536,0.04582311047448052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,1536,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,4096,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,2048,0.7817360030280219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,128,0.029504888587527808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,4096,0.02439555525779724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,1024,0.006516444600290722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,1536,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,8192,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,6144,0.039094222916497126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,256,0.019783111082182992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,6144,0.12667911582522923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,12288,0.05474044548140632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,8192,0.13309244314829508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,3072,0.009975999593734741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,8192,0.03244977858331468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,1024,0.026030222574869793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,1536,0.01260088880856832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,1024,0.02160266704029507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,10240,0.0499582224422031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,2560,0.19035732746124268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,4096,0.25815733273824054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,10240,0.028875556257035997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,12288,0.042986667818493314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,1536,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,768,0.014175110393100314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,32,0.012108444339699216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,256,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,5120,0.02383199996418423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,8192,0.04938933253288269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,65536,1.0946853425767686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,5120,0.038378665844599404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,768,0.004359111189842224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,65536,0.05708444118499756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,65536,0.3316737810770671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,12288,0.03628266519970364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,8192,0.02973066767056783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,6144,0.04422755704985725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,10240,0.05323555403285556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,2560,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,6144,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,8192,0.13163555992974177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,512,0.0042071110672420925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,2048,0.011580444044537015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,5120,0.08849066495895386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,3072,0.013053333593739403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,8192,0.05962666538026598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,1536,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,256,0.010445333189434476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,1024,0.018183110488785636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,3584,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,3072,0.031496001614464655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,512,0.03046755658255683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,5120,0.009331555830107795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,2560,0.030410667260487873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,64,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,768,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,16384,0.13000888294643825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,16384,0.02100533412562476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,6144,0.019522666931152344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,2560,0.00921688891119427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,1024,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,51200,16.58105638292101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,1536,0.019021333919631112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,8192,0.11582310994466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,2560,0.050717333952585854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,8192,0.016024000114864774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,12288,0.013720889058378009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,256,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,1536,0.034449779325061373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,51200,0.076910224225786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,16384,0.0772764417860243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,1536,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,5120,0.1709137757619222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,65536,0.048230220874150596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,32,0.0324906673696306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,8192,0.031176888280444678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,8192,0.01112711098459032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,1536,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,65536,1.1945466995239258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,8192,0.02437155610985226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,16384,0.027148443791601393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,10240,0.31619289186265737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,1536,0.05187466740608215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,128,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,2048,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,10240,0.18464799722035727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,3072,0.016224000189039443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,1536,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,128,0.012124444047609964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,4096,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,1536,0.03035466538535224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,8192,0.026162667406929865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,1024,0.19946577813890246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,8192,0.5579742325676812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,51200,5.31240971883138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,7168,0.06343644195132785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,2560,0.05324000120162964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,4096,0.04855288730727302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,2560,0.03303466571701898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,6144,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,5120,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,8192,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,1024,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,128,0.015283554792404175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,3584,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,2048,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,2048,0.5719448725382487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,2048,0.006217777729034424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,2560,0.04235733217663235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,16384,0.04559466573927137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,16384,0.5324844254387749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,2048,0.141047994295756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,16384,0.0869599978129069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,768,0.21982133388519287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,3072,0.021895110607147217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,16384,0.05081955591837565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,65536,1.882016075981988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,8192,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,128,0.003253333270549774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,3072,0.007237333390447829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,51200,0.12819555070665148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,128,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,8192,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,8192,0.011360888679822287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,64,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,51200,2.542848798963759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,1024,0.13221689065297446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,16384,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,2560,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,768,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,1024,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,512,0.016552888684802584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,5120,0.035183111826578774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,64,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,10240,0.01736177835199568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,256,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,2560,0.024446222517225478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,32,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,7168,0.01297688898113039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,3584,0.16408888498942056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,16384,0.09898133410347833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,512,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,65536,0.17948710918426514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,2048,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,3072,0.01677600045998891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,1536,0.0038311113086011675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,10240,0.025191111697090998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,4096,0.0184560004207823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,7168,0.02756533357832167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,3584,0.015556444724400839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,1024,0.06283022297753228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,6144,0.02209866709179348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,8192,0.04845155609978569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,4096,0.007884444461928474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,12288,0.06390844451056586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,2048,0.011927111281288994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,8192,0.011257777611414591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,3072,0.016023111012246873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,8192,0.019487111104859244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,4096,0.030768000417285498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,128,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,51200,0.07662755250930786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,6144,0.028103109863069322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,3072,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,5120,0.054209778706232704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,256,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,128,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,8192,0.009736888938479954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,12288,0.10555022292666966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,65536,0.0832808878686693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,10240,0.022545778089099463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,32,0.0019217777169413036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,256,0.026130666335423786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,8192,0.24325243631998697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,32,0.0024480000138282776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,1536,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,12288,0.23482044537862143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,10240,0.479856014251709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,10240,0.014947555131382413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,128,0.0058648888435628675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,12288,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,3584,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,1024,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,16384,0.022153778208626643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,128,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,12288,0.034266667233573064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,8192,0.04542311032613119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,1024,0.020918221937285528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,3584,0.017165333032608032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,2048,0.004318222403526306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,512,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,512,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,65536,1.6981289121839735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,6144,0.0424231125248803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,4096,0.010871111518806882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,128,0.018253333038753934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,2048,0.031214220656288996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,3072,0.1466346714231703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,2560,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,51200,0.08541333013110691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,10240,0.010861333045694562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,8192,0.0320995549360911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,7168,0.2202062209447225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,1536,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,12288,0.0704995526207818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,65536,0.16027377711402044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,2048,0.004252444538805219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,32,0.01814844376511044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,16384,0.058303998576270215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,12288,0.011239111423492432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,6144,0.8417546484205457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,3072,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,1024,0.02052888936466641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,2560,0.012897777888509961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,4096,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,512,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,64,0.013403555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,2560,0.06512710783216688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,1024,0.00815288888083564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,256,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,64,0.0037413334680928122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,768,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,51200,0.5475200017293295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,4096,0.03360533383157518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,10240,0.012815111213260226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,256,0.012757333616415659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,64,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,32,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,7168,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,256,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,1024,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,4096,0.22952267858717176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,51200,0.09624355369144016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,7168,0.0728728903664483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,2560,0.1023626658651564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,8192,0.03357777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,256,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,12288,0.11340444617801243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,7168,0.052765332990222506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,3584,0.8011760181850858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,3584,0.023408000667889912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,2048,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,51200,0.05316888954904345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,6144,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,6144,0.022965333527988855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,64,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,256,0.03718488746219211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,256,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,5120,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,6144,0.013433777623706393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,12288,0.021527111530303955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,5120,0.03472088774045309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,2560,0.006504888749784893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,5120,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,768,0.056259552637736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,3584,0.024454222785101995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,12288,0.3853982289632161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,4096,0.27919289800855845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,65536,0.4571146435207791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,3072,0.07818666431638929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,2048,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,768,0.011789333489206104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,12288,0.11884355545043945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,2048,0.08180444770389132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,3072,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,6144,0.27910489506191677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,256,0.026332444614834253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,1536,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,5120,0.03862222366862827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,512,0.012244444754388599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,1024,0.14042754968007407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,5120,0.04085777865515815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,768,0.11525689231024848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,128,0.0021173333128293357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,768,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,256,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,1536,0.03484800126817491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,7168,0.12554489241706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,128,0.009388444324334463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,64,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,12288,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,2560,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,8192,0.01716977854569753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,51200,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,65536,0.07306577761967976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,7168,0.15313600169287786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,16384,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,32,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,512,0.006858666737874349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,12288,0.01475288967291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,2560,0.06941777467727661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,3584,0.03991644581158956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,5120,0.058895111083984375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,32,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,5120,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,5120,0.03477955526775784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,3072,0.014096000128322177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,64,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,7168,0.2302791012658013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,3584,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,10240,0.05814488728841146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,3584,0.011392888923486074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,3584,0.04282222191492716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,16384,0.04109955496258206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,7168,0.08153244521882799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,1024,0.019612444771660697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,512,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,5120,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,10240,0.04176177912288242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,1024,0.10302666823069255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,5120,0.0396088891559177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,2560,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,512,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,768,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,1024,0.01886133352915446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,16384,1.9372969733344183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,8192,0.19785243935055205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,65536,1.8455742730034723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,256,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,3072,0.04298133320278592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,3584,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,1024,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,6144,0.010202666951550378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,128,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,12288,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,2048,0.221160888671875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,256,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,16384,0.07885689205593534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,256,0.030059556166330974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,256,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,3072,0.08028711213005914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,3584,0.03642933236228095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,8192,0.03820000092188517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,2048,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,3072,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,6144,0.12152800295088027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,16384,0.2626417742835151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,7168,0.03770666652255588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,7168,0.013391111459996967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,5120,0.08710844649208917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,64,0.00226488895714283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,10240,0.028770665327707928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,12288,0.08076088958316378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,128,0.08107288678487141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,1024,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,768,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,65536,0.47927199469672305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,768,0.009299555586444007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,4096,0.019910222954220243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,10240,0.03698844379848904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,3072,0.01702666613790724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,2560,0.020023110840055678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,5120,0.00925511121749878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,768,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,64,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,2560,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,128,0.0018053334206342697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,6144,0.019345778557989333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,8192,0.14447911580403647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,3584,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,32,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,12288,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,3584,0.1522622240914239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,768,0.03418311145570543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,12288,0.5742017957899306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,51200,0.25912533866034615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,32,0.002565333412753211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,3072,0.018070222602950204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,7168,0.021818665994538203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,1536,0.16894133885701498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,3584,0.04391910963588291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,3072,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,51200,0.0766577786869473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,256,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,10240,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,1536,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,65536,0.1480204396777683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,1536,0.028012444575627644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,3072,0.07214577992757161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,512,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,51200,0.5250710911220974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,51200,0.05407822132110596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,1024,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,6144,1.1249440511067708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,768,0.03399111164940728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,6144,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,64,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,1024,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,32,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,32,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,65536,0.1298604408899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,5120,0.03148266673088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,8192,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,128,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,65536,2.6113404168023004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,3072,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,2048,0.02092622220516205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,32,0.0070159998204973005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,16384,0.08107378085454305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,8192,0.051608890295028687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,12288,0.1535617775387234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,6144,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,7168,0.015560888581805758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,4096,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,3072,0.010672888822025724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,65536,0.3873324394226074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,12288,0.02017244365480211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,2048,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,3584,0.00959911114639706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,1536,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,4096,0.009713778065310584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,2560,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,8192,0.07526577843560113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,64,0.07775466971927218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,2560,0.027430221438407898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,64,0.011563555234008364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,51200,1.9237564934624567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,512,0.02568711174858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,7168,0.044529779089821704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,768,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,16384,0.0458204448223114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,1024,0.008819555242856344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,1024,0.19736799928877088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,16384,0.10103022389941746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,512,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,4096,0.008056888977686564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,768,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,51200,0.04912177721659342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,65536,0.2646026611328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,64,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,10240,0.03229688935809665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,256,0.0255022214518653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,128,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,3584,0.21452089150746664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,2048,0.010788444843557147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,4096,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,65536,0.41394665506150985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,8192,0.02397777802414364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,32,0.06080444653828939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,768,0.018529777725537617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,16384,0.03052622079849243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,1536,0.019623999794324238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,3072,0.013757333159446716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,5120,0.2197457816865709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,5120,0.01797422269980113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,10240,0.10195022159152561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,12288,0.015225778023401896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,3584,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,51200,17.23576185438368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,4096,0.031052443716261122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,6144,0.13247466087341309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,12288,0.16136622428894043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,2560,0.07694222529729207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,1024,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,5120,0.0042453333735466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,2048,0.009583999713261923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,65536,0.17824978298611113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,3584,0.007680888805124495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,2048,0.006335111127959357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,64,0.003149333306484752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,64,0.006464889066086874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,4096,0.04133955637613932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,512,0.011355555719799466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,32,0.0042168887125121225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,4096,0.035959998766581215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,768,0.016125332978036668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,8192,0.011680000358157687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,12288,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,1536,0.02016533414522807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,2048,0.021670222282409668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,6144,0.03156000044610765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,12288,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,256,0.015127999915017022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,1536,0.006342222293217977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,1024,0.016788444585270353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,1024,0.04475377665625679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,1536,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,2560,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,1536,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,5120,0.04530844423505995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,2048,0.0510773327615526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,256,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,512,0.02773511078622606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,16384,0.15385866165161133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,64,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,65536,0.1574524508582221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,3072,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,16384,0.03267644511328803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,6144,0.019018666611777413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,256,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,32,0.008062221937709385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,12288,0.1921671099132962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,7168,0.3993848959604899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,128,0.00534400012758043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,128,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,1024,0.01829688913292355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,6144,0.25114666091071236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,1536,0.006680000159475539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,256,0.012107555237081317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,16384,0.21107910739050972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,5120,0.0956968863805135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,12288,0.4415857791900635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,1536,0.016840888394249808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,512,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,2048,0.009466666314336989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,6144,0.021363554729355708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,256,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,65536,0.23800267113579643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,16384,0.02195466723706987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,64,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,2560,0.06401866674423218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,4096,0.10447733269797431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,2048,0.5177262094285753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,7168,0.01923466722170512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,3072,0.04849866694874234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,8192,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,768,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,2560,0.011076444553004371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,3584,0.050844444168938525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,8192,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,6144,1.1150195863511827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,51200,0.7482044431898328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,2048,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,3584,0.032272001107533775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,2048,0.020251555575264823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,4096,0.01772444446881612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,32,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,3072,0.11028177870644464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,1024,0.0029271110478374693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,5120,0.017758222089873422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,512,0.015095111396577625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,768,0.003759110967318217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,2048,0.011791110866599612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,8192,0.10874933004379272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,1536,0.014347554908858405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,256,0.008437333007653555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,3584,0.010584889186753167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,12288,0.049646221929126315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,6144,0.027264888087908428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,768,0.01936177743805779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,256,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,2048,0.022431999444961548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,768,0.014948444234000312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,10240,0.0359208881855011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,16384,0.052652445104387074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,5120,0.033810665210088096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,4096,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,2560,0.023382221659024555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,128,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,1536,0.028229332632488672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,1024,0.006360888895061281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,8192,0.10087022516462539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,1536,0.03740266627735562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,8192,0.036995556619432234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,5120,0.08751466539171006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,3072,0.019235556324323017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,2048,0.02032088902261522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,256,0.005427555491526921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,64,0.002587555597225825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,5120,0.18001777595943877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,128,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,51200,0.1435991128285726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,1536,0.014563555518786112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,16384,0.24249778853522408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,2048,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,8192,0.03877244393030802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,2560,0.013694222602579327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,2048,0.017863111363516915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,6144,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,768,0.003608889049953885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,2048,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,256,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,1536,0.017493334081437852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,16384,0.22261155976189506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,128,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,12288,0.19212978416019014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,256,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,12288,0.013872000078360239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,3072,0.03514044483502706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,51200,0.29222223493787974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,51200,0.15773244698842367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,64,0.010503110786279043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,16384,0.01863911085658603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,4096,0.07181155681610107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,51200,4.704555511474609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,51200,0.3928079870012071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,256,0.021966222259733412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,10240,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,256,0.02848177817132738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,10240,0.13800977336035833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,7168,0.04953866534762912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,3584,0.02853155467245314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,12288,0.11364889144897461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,768,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,768,1.6655048794216578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,512,0.013463999662134381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,7168,0.116939557923211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,512,0.029448890023761328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,128,0.0034017778105205963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,6144,0.07312266694174872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,10240,0.025881777207056682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,256,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,1024,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,16384,0.06570311387379964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,6144,0.013184000220563678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,768,0.011108444796668159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,7168,1.9656230078803167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,512,0.01110666659143236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,51200,3.6435894436306424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,51200,0.22879555490281847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,128,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,256,0.022621333599090576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,10240,0.03104089034928216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,2560,0.028512001037597656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,3584,0.08476622237099542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,2560,0.1182746622297499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,512,0.02196888956758711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,10240,0.01200977795653873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,16384,9.934849209255642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,2560,0.03496355480617947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,1024,0.018963555494944256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,4096,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,512,0.045257776975631714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,64,0.00812000036239624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,7168,0.030717333157857258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,3584,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,5120,0.13603199852837458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,8192,0.02053600052992503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,65536,0.13550578223334417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,4096,0.12491910987430149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,3072,0.014548444085650973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,256,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,1536,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,2048,0.00758133331934611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,2048,0.23255909813774955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,2048,0.007045333584149678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,16384,0.3841706646813287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,8192,0.025052444802390203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,16384,0.02216622233390808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,16384,0.13672178321414524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,256,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,1536,0.02124533388349745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,12288,0.40508000055948895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,12288,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,1536,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,3072,0.020577778418858845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,2048,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,5120,0.029891557163662378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,6144,0.03375022278891669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,768,0.0540924436516232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,4096,0.07169689072502984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,128,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,4096,0.04755111204253303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,32,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,10240,0.02667377723587884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,512,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,4096,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,7168,0.011398221883508893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,4096,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,6144,1.92939207288954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,51200,0.10907822185092503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,3584,0.044884443283081055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,1536,0.013131555583741931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,65536,0.13403377268049452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,8192,0.013228444589508904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,2560,0.02017155620786879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,768,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,32,0.017831999394628737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,16384,0.1928995582792494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,1024,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,512,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,6144,0.07549244165420532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,7168,0.20772266387939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,256,0.010393777655230628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,4096,0.17200177245669893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,3584,0.046353777249654136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,12288,0.07157511181301542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,2048,0.008914666871229807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,12288,0.06065422296524048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,1536,0.047269334395726524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,512,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,2560,0.021485333641370136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,7168,0.012551999754375882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,2048,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,2048,0.006745777610275481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,5120,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,512,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,65536,0.7578186459011502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,8192,0.019719999697473314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,4096,0.03363200028737386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,5120,0.02568622264597151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,1024,0.014878221684032016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,256,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,256,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,128,0.014946666028764514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,3072,0.025992888543340895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,8192,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,256,0.020598222812016804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,2048,0.004867555366622077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,16384,0.06750488943523832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,4096,0.1063413355085585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,2560,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,2560,0.010426667001512317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,65536,1.5417822731865778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,7168,0.016795555750528973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,3584,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,1024,0.022973333795865376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,1024,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,51200,9.465108235677084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,51200,0.18120533890194365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,3584,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,2048,0.012710221939616732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,16384,0.1385457780626085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,1024,0.02643555568324195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,2560,0.15680444240570068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,10240,0.4934524430169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,64,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,3584,0.020652444826232064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,12288,0.020948444803555805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,768,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,2048,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,512,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,10240,0.19407733281453451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,10240,0.030823998981051978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,8192,0.01975022256374359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,512,0.11838044060601129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,3072,0.14698933230506048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,1536,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,1024,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,256,0.021736888421906367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,7168,0.018975999620225694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,4096,0.049888001547919385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,6144,0.03985333442687988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,5120,0.06507555643717448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,16384,0.05752711163626777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,3584,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,4096,0.03315644462903341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,3584,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,3072,0.00978488889005449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,3072,0.01796355512407091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,51200,1.6066950692070856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,12288,0.18995111518436006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,512,0.0185217774576611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,256,0.08820444345474243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,6144,0.054025779167811074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,12288,0.02121066715982225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,7168,0.07219377491209242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,16384,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,10240,0.07772888739903767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,1536,0.023424888650576275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,4096,0.008107555409272512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,5120,0.010583110981517367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,256,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,3072,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,65536,0.41892533832126194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,16384,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,2048,0.01756177842617035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,1536,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,7168,0.03575199842453003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,4096,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,4096,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,5120,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,12288,0.013755555782053204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,12288,0.18589156203799775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,128,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,128,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,5120,0.01973777843846215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,512,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,512,0.012719111310111152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,6144,0.1296373340818617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,768,0.018135999639829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,8192,0.05775377485487196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,32,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,3584,0.04967022273275587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,51200,0.04986310998598734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,6144,0.049534221490224205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,512,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,4096,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,2560,0.026363554928037856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,1024,0.02238311039076911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,256,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,6144,0.3893582291073269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,7168,0.010263111028406356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,5120,0.10674311055077447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,51200,0.07190755340788099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,2560,0.018803555104467604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,51200,1.6221031612820094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,12288,0.37578667534722227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,7168,0.10546133253309463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,64,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,12288,0.06083110968271891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,1536,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,51200,1.2487209108140733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,1536,0.01757333344883389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,16384,0.07101777527067396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,3072,0.13536800278557673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,8192,0.29549598693847656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,6144,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,1024,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,64,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,16384,0.010779555473062726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,2560,0.03432888785998026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,10240,0.98544798956977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,5120,0.01740533279048072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,64,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,1024,0.012836444709036084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,1536,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,3584,0.020543111695183646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,10240,0.028896888097127277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,3584,0.06122577852673001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,512,0.006863111009200414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,4096,0.019727110862731934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,3584,0.019623999794324238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,1024,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,3072,0.019916445016860962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,32,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,10240,0.015592000550693937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,51200,0.10401778088675605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,8192,0.031298667192459106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,5120,0.005985777825117111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,3072,0.0816026661131117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,4096,0.03341866532961527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,16384,0.23451821009318033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,10240,0.19501333766513398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,1536,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,10240,0.026437333888477747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,4096,0.02331377731429206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,4096,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,512,0.16027911504109701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,7168,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,65536,0.0663742224375407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,3584,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,65536,0.13675467173258463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,6144,0.17804711394839814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,3584,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,512,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,768,0.0514959990978241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,256,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,4096,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,1536,0.023735110958417256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,256,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,2560,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,6144,0.034952001439200506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,12288,0.1026186678144667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,4096,0.014761777387724983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,64,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,128,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,10240,0.2769102255503337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,4096,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,65536,2.0286933051215277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,3584,0.02495644489924113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,3584,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,8192,0.025607110725508794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,5120,1.1138106452094185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,5120,0.011143111520343356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,2560,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,2560,0.027449778384632532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,64,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,4096,0.03046311272515191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,2048,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,8192,0.08195822106467353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,51200,2.0763492584228516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,4096,0.017749332719379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,256,0.017252445220947266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,768,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,51200,1.452219539218479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,3072,0.16755466991000703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,3072,0.03066666589842902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,1536,0.01715288890732659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,65536,0.12333510981665717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,3072,0.028343111276626587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,3072,0.025655110677083332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,51200,0.06499733527501424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,3584,0.0637306637234158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,16384,0.03755555550257365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,51200,0.4440595573849148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,6144,0.0461422238085005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,16384,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,768,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,4096,0.04044177797105577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,16384,0.03445155421892802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,7168,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,1536,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,256,0.011792889071835412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,2560,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,12288,0.02850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,256,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,65536,0.08175910843743218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,4096,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,128,0.005436444448100195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,2048,0.026135999295446608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,12288,0.5189555486043295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,768,0.023964444796244305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,3584,0.2442622184753418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,2048,0.025937777426507738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,64,0.00350577798154619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,1536,0.038174221913019814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,65536,0.0950017770131429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,51200,0.050548444191614784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,16384,0.03564000129699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,65536,0.5290177663167318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,10240,0.1140693293677436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,2048,0.02674666709370083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,256,0.005526222288608551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,5120,0.05520800087187025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,4096,0.04559200008710226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,8192,0.006034666465388403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,1024,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,16384,0.13221600320604113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,51200,0.06865511337916057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,1024,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,12288,0.1885084443622165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,256,0.0074942219588491656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,8192,0.01941777765750885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,12288,0.12593955463833278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,128,0.15175999535454646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,128,0.016275554895401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,512,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,6144,0.009684444301658207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,2560,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,4096,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,10240,0.08038222127490573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,3072,0.18318400118086073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,1536,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,3584,0.014828445182906257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,32,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,2560,0.03274755676587423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,3072,0.010102222363154093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,51200,0.1406382189856635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,512,0.015249777171346875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,5120,0.07036711110009088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,7168,0.028785778416527644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,8192,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,10240,0.1622977786593967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,12288,0.7068453364902072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,5120,0.11362400319841172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,51200,0.06015377574496799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,128,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,12288,0.14794666237301296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,1536,0.016377778516875375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,6144,0.030187557140986126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,256,0.016715556383132935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,7168,0.019122666782803006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,2560,0.0299288895395067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,5120,0.011579555769761404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,8192,0.20410399966769746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,7168,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,3584,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,1536,0.16472089290618896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,1536,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,51200,0.769038200378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,512,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,16384,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,16384,0.09844355450736152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,7168,0.06100355254279243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,32,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,6144,0.03923822111553616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,6144,0.06933866606818305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,128,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,3584,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,6144,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,6144,0.037681778271993004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,3584,0.034867554903030396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,16384,0.17007377412584093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,2048,0.07700088951322767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,10240,0.027443556321991816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,512,0.0067528887755341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,12288,0.09012977944480048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,128,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,1024,0.034449779325061373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,4096,0.015127999915017022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,1024,0.02309688925743103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,51200,9.134221394856771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,5120,0.021664889322386846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,64,0.048787554105122886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,3072,0.012108444339699216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,128,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,2048,0.04549244377348158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,3072,0.019537778364287484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,4096,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,32,0.00981777740849389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,7168,0.009157333109113906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,256,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,65536,0.032469332218170166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,10240,0.0123831108212471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,1024,0.015554666519165039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,65536,0.05874400006400215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,512,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,7168,0.020555555820465088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,6144,0.01993600030740102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,8192,0.02055911057525211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,64,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,7168,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,256,0.018591110905011494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,128,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,6144,0.06664178106519911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,6144,0.03057866626315647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,64,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,6144,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,8192,0.019542222221692402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,51200,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,6144,0.8098559909396701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,4096,0.021134222547213238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,7168,0.020918221937285528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,1024,0.0041422223051389056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,64,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,51200,0.6172364552815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,1536,0.023947555157873366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,128,0.03762755460209317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,2048,0.010575110713640848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,1024,0.1950684388478597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,7168,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,3584,0.011076444553004371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,7168,0.02199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,10240,0.3709733221266005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,32,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,51200,3.176653332180447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,10240,0.23232356707255045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,12288,0.014849777022997538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,128,0.04164266586303711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,6144,0.09076355563269721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,768,0.005392888767851724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,2048,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,7168,0.025172443853484258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,8192,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,12288,0.018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,65536,0.27526312404208714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,3584,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,8192,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,5120,0.02090577781200409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,4096,0.017415111263593037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,3072,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,5120,0.09319111373689438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,6144,0.5365031030442979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,12288,0.029554665088653564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,65536,0.009328888522254096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,3584,0.036834667126337685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,16384,0.08197955290476482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,4096,0.3395306534237332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,12288,0.07238311237759061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,2560,0.021708443760871887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,128,0.024337778488794964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,4096,0.029137776957617864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,32,0.0063324446479479475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,8192,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,7168,0.020308444897333782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,12288,0.08996355533599854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,65536,2.043445375230577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,5120,0.01575555569595761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,7168,0.0873146653175354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,51200,0.07957688967386882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,65536,0.0858008861541748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,65536,0.06816533539030287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,2560,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,16384,0.08202933602862887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,512,0.04658400019009908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,4096,0.01735911104414198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,2048,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,65536,0.07258222500483195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,3072,0.0075191110372543335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,3072,0.015224888920783997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,5120,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,3584,0.017836444907718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,3072,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,1024,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,256,0.007834666305118138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,32,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,10240,0.020954666866196528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,65536,0.12670932875739202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,3072,0.0321342216597663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,32,0.009825777676370408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,3584,0.01442311041884952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,3584,0.023997333314683702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,1024,0.013737777868906656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,7168,0.02642044425010681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,7168,0.034860445393456355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,5120,0.020344000723626878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,3072,0.029078221983379785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,8192,0.011027555498811932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,128,0.020759110649426777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,65536,0.31855911678738064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,2560,0.013734222286277346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,3584,0.008691555923885768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,2048,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,12288,0.017518222332000732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,8192,0.9226826561821831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,1024,0.014940443966123792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,1024,0.03269244564904107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,10240,0.031529777579837374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,2048,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,768,0.01312355531586541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,2560,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,65536,3.300199932522244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,10240,0.09999910990397136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,6144,0.03852444556024339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,32,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,12288,0.016846223009957206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,65536,0.054274668296178184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,128,0.014101333088344999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,8192,0.12924888398912218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,3072,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,16384,0.026064889298544988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,16384,0.020058666666348774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,7168,0.17892089154985216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,6144,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,6144,0.010953777366214328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,4096,0.11451466878255208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,2048,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,12288,0.22308709886338976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,768,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,64,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,65536,0.12507555219862196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,12288,0.01830844415558709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,16384,0.046987556748920016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,3584,0.03545866741074456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,512,0.012608889076444836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,1536,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,7168,0.024685333172480266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,128,0.0041244443919923575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,12288,0.059099555015563965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,16384,3.0526631673177085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,128,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,16384,0.0247697780529658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,128,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,512,0.01687466601530711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,8192,0.0753706693649292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,1024,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,128,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,512,0.013016000390052795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,7168,0.048882666561338634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,10240,0.019522666931152344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,51200,0.24840799967447916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,768,0.15726489490932888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,3584,0.035585777627097234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,10240,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,65536,0.07181600067350599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,256,0.012148444851239523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,51200,0.15530044502682155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,3584,0.01848266687658098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,32,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,1536,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,3584,0.010999999940395355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,5120,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,4096,0.11584711074829102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,12288,0.028699556986490887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,768,0.12935022513071695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,2560,0.017038222816255357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,7168,0.028766221470303003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,128,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,51200,0.27310310469733345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,16384,0.043326222234302096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,7168,0.022064889470736187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,512,0.07589511076609294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,3072,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,5120,0.014502222339312235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,12288,0.06727021932601929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,4096,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,3072,0.00609688874748018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,3584,0.0236791107389662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,16384,0.02012977831893497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,12288,0.03395377927356296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,16384,0.5534213383992513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,256,0.011555555793974133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,128,0.02253955602645874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,6144,0.01923288901646932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,32,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,2048,0.05448266532686022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,512,0.006799110935793982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,51200,0.05093600021468269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,65536,0.05791733662287394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,1536,0.015991111596425373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,1536,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,1536,0.1817422178056505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,2560,0.047622223695119224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,1536,0.004294222427739037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,2560,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,10240,0.042466666963365346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,256,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,3584,0.01586933268441094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,10240,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,1536,0.022089777721299067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,5120,0.10607733329137166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,5120,0.016454221473799813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,3072,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,10240,0.11901599831051296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,128,0.011658666862381829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,12288,0.015391111373901367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,16384,3.5615857442220054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,4096,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,768,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,2048,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,5120,0.05036977926890055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,12288,0.024262222978803847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,8192,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,128,0.0028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,2048,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,1024,0.009970666633711921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,1024,0.07723999685711332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,512,0.010589333044158088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,256,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,4096,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,12288,0.05916355715857612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,16384,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,512,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,6144,0.43070043457878965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,768,0.01666844387849172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,128,0.029504888587527808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,128,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,2048,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,51200,0.05788977940877279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,6144,0.022222222553359136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,12288,0.026365333133273657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,6144,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,5120,0.019187556372748483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,2048,0.026949332820044622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,16384,0.04240711198912727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,3072,0.05981510877609253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,3072,0.07170844740337796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,3584,0.4951670964558919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,51200,0.3538515567779541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,2048,0.005693333430422678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,65536,4.960784064398871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,3072,0.020985777179400127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,8192,0.08376177814271714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,10240,0.0988053348329332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,2048,0.04048266675737169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,512,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,128,0.02313511073589325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,3584,0.011016888750924004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,512,0.017948443690935772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,256,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,4096,0.015223110715548197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,3072,0.032255109813478254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,256,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,1024,0.02881422307756212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,512,0.022841778066423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,12288,0.10070844491322835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,12288,0.08761777480443318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,3584,0.08067111174265544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,12288,0.034508443540996976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,4096,0.01628800067636702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,2048,0.05651199817657471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,10240,0.028734223710166082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,2048,0.0046728890803125174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,3072,0.14015555381774902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,768,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,51200,0.6022737820943197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,2560,0.05833066834343804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,768,0.003920888735188378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,256,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,4096,0.03989955451753404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,12288,0.024130667249361675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,7168,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,768,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,3072,0.029412445094850328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,2048,0.03808800048298306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,768,0.09124266439014012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,3584,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,512,0.00816533300611708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,768,0.05001422100596958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,10240,0.1869964467154609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,512,0.05718400080998739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,5120,0.023181334137916565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,8192,0.05639377567503187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,6144,0.02032355633046892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,3072,0.01603555513752831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,65536,0.7816586494445801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,128,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,2048,0.03680266605483161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,3584,0.025983999172846477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,65536,0.13140000237358943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,12288,0.11793600188361274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,64,0.0034826666944556763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,12288,0.013564444250530668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,7168,1.9941804673936632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,5120,0.0875040027830336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,7168,0.013632888595263163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,256,0.01586577792962392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,1536,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,3072,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,16384,0.02240088913175795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,16384,0.026314665873845417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,768,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,2560,0.059261335266960986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,1024,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,3584,0.015810666812790763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,6144,0.02667288813326094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,2560,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,8192,0.009429333110650381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,8192,0.01681777834892273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,12288,0.034853332572513156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,128,0.02663822140958574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,1024,0.022745778163274128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,12288,0.03905511233541701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,512,0.014202665951516895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,1024,0.006106666806671355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,8192,0.03137066629197862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,2048,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,3072,0.015448888142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,7168,0.07016533613204956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,51200,3.2431155310736766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,4096,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,512,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,8192,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,1536,0.01090133355723487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,8192,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,6144,0.026397332549095154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,3072,0.03335377905103896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,4096,0.007752888732486301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,3072,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,1536,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,64,0.0018151110659042995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,256,0.003463111196955045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,5120,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,65536,0.129713773727417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,256,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,51200,0.6276177830166286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,768,0.1429191165500217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,65536,0.045591112640168935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,1024,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,64,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,2560,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,3584,0.06416889031728108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,128,0.09611199961768256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,6144,0.10386488835016887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,12288,0.058486223220825195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,512,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,6144,0.09785155455271403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,128,0.047372443808449634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,2560,0.1080720027287801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,3584,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,1536,0.0350213348865509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,8192,0.0524160001013014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,1024,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,7168,0.031119111511442397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,65536,0.09691199991438125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,10240,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,51200,0.5554951032002767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,2048,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,4096,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,3072,0.013431111143694984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,3072,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,7168,0.020232889387342665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,5120,1.1773306528727214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,7168,0.019888000355826486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,10240,0.02881244487232632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,51200,0.11232711209191217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,1024,0.01276622215906779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,512,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,16384,0.14358933766682944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,7168,0.016131555040677387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,51200,0.048134220971001514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,1536,0.06639822324117024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,4096,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,65536,0.12541245089636907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,1024,0.013473778135246701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,2560,0.033139556646347046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,256,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,2560,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,1536,0.016898666818936665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,5120,0.01737422247727712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,768,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,65536,0.19571822219424775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,6144,0.030160887373818293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,3584,0.03950577643182542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,8192,0.03863644599914551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,256,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,6144,0.13758488496144614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,2560,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,768,0.0042462220622433555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,1024,0.027702222267786663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,1536,0.0032711111836963226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,6144,0.009310222334331935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,12288,0.043246222866906055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,2560,0.02913244565327962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,128,0.0053582220441765254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,1024,0.07081333133909437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,7168,0.11548444959852432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,128,0.00564533347884814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,12288,0.030586666531032983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,5120,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,65536,0.07898577716615465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,3584,0.024898666474554274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,3072,0.0074933336840735535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,512,0.028628445333904688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,12288,0.019520888725916546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,128,0.0021342221233579847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,8192,0.014027555783589682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,1024,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,51200,0.1185342205895318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,10240,0.01575911045074463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,12288,0.031919111808141075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,1024,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,2560,0.004929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,1024,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,256,0.6403368843926324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,16384,0.14787288506825766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,3584,0.2275235652923584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,8192,0.011381333072980246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,4096,0.47116888893975156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,128,0.009846222069528367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,51200,0.0537120004494985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,51200,0.12000532944997151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,256,0.035564445787005954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,1536,0.023558222585254248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,1536,0.008408889174461365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,10240,0.04333244429694282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,7168,0.010910222099887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,7168,0.12446399529774983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,1024,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,1024,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,512,0.025291555457644995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,8192,0.07256711191601224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,1536,0.01606399979856279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,8192,0.022652443912294176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,4096,0.01369244439734353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,51200,0.31979287995232475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,5120,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,128,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,2048,0.02535199953450097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,65536,0.13751733303070068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,2048,0.0587022238307529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,64,0.013621333572599622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,768,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,2048,1.6609937879774306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,4096,1.3684719933403864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,3072,0.012988444831636218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,65536,0.0659244457880656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,65536,0.8061813248528374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,2560,0.025213332639800176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,2048,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,3072,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,768,0.0679875546031528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,51200,8.157208760579428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,16384,0.022395556171735127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,6144,0.753177801767985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,256,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,16384,0.04361600014898512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,1536,0.02156177825397915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,128,0.012585777375433179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,64,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,8192,0.013734222286277346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,2560,0.04110844598876105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,1024,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,1536,0.07215111123190986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,2048,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,10240,0.04661422305636936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,4096,0.021390222840838965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,3072,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,1024,0.007085333267847697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,4096,0.014516444669829475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,3584,0.12818755043877497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,65536,0.10803111394246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,51200,0.2642595503065321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,51200,0.06309600008858575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,2048,0.005414222263627582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,3072,0.9485724767049154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,12288,0.03870044483078851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,51200,0.17211288876003691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,1024,0.0038702223036024305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,3584,0.05625066492292616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,51200,0.32039109865824383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,51200,0.07636621925565931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,8192,0.01776711146036784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,3072,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,6144,0.026084444589085046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,1024,0.031601776679356895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,64,0.15002311600579157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,5120,0.045851555135515004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,16384,0.0321999986966451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,3584,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,12288,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,12288,0.01868088874551985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,4096,0.008084444536103142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,2048,0.1386186679204305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,512,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,12288,0.07900889052285089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,2048,0.6684693230523003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,128,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,8192,0.09298488828870986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,512,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,32,0.002542222125662698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,12288,0.03601422243648105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,3584,0.007798222204049428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,3072,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,768,0.042141334878073804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,3072,0.005535999933878581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,3072,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,12288,0.23035732905069986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,32,0.0064080001579390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,2048,0.00443111111720403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,512,0.019023999571800232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,128,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,7168,0.03164622187614441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,3584,0.059121780925326876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,256,0.0387644436624315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,4096,0.03560622202025519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,51200,0.06936799817615084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,51200,0.050889776812659375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,3584,0.05140799946255154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,3584,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,64,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,3584,0.02199199961291419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,6144,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,2048,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,5120,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,7168,0.011072888970375061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,1536,0.01920622256067064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,2560,0.0409786668088701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,2048,0.040525333748923406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,6144,0.01742044422361586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,10240,0.06518399715423584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,3072,0.24924090173509386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,7168,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,7168,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,64,0.03092355529467265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,10240,0.04248799880345663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,51200,0.12346222665574814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,128,0.07247289021809895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,2560,0.03464711043569777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,10240,0.18238755067189535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,12288,0.033997333712048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,51200,0.6381591161092123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,65536,0.06390044424268934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,768,0.21544978353712294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,5120,0.1660968859990438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,3584,0.02995022137959798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,1024,0.024795555406146582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,3072,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,1536,0.01331911070479287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,65536,0.06361955404281616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,6144,0.030251555972629126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,1024,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,16384,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,6144,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,64,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,12288,0.028747555282380845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,3584,0.08111111323038737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,64,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,3584,0.03521600034501817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,10240,0.05141777793566386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,1024,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,8192,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,51200,0.1396684381696913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,7168,0.06259555286831327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,4096,1.9599325391981337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,2560,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,7168,0.01425333321094513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,8192,0.21288622750176323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,3584,0.026939556002616882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,2048,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,5120,0.018380444910791185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,65536,0.14829599857330322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,2048,0.020586666133668687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,6144,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,10240,0.014173333843549093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,64,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,2560,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,4096,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,10240,0.011446221835083432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,12288,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,65536,0.1755688852734036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,2560,0.08157333400514391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,1536,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,16384,0.03643733263015747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,768,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,7168,0.1858337720235189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,256,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,1024,0.013912889692518445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,64,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,10240,0.012230222423871359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,65536,0.18224799633026123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,6144,0.01686844395266639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,12288,1.6011226442125108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,2560,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,128,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,8192,0.011207111179828644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,2048,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,32,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,768,0.009418666362762451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,16384,0.35698665512932676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,768,0.0026933333526055017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,1024,0.006713777780532837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,2560,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,7168,0.0206977774699529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,4096,0.7552515665690104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,256,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,5120,0.017501332693629794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,768,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,3072,0.05663555860519409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,1536,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,51200,0.907959090338813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,10240,0.03984888725810581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,12288,0.5575324694315592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,64,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,51200,0.2519822120666504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,51200,0.10375733508004083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,6144,0.02686044408215417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,1024,0.007151111132568783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,3584,0.017487999465730455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,10240,0.04731200138727824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,32,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,1024,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,1024,0.016403555870056152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,6144,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,1024,0.02165066699186961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,2560,0.006691555596060223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,2048,0.012030222349696688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,64,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,1024,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,7168,0.012540444731712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,10240,0.051700443029403687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,256,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,5120,0.08912177880605061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,6144,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,5120,0.10490933391782974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,4096,0.03316355413860745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,3072,0.13734222782982722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,7168,0.13620355394151476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,128,0.005145777844720417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,16384,0.4000977675120036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,4096,0.03156444430351257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,12288,0.004931555440028508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,8192,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,7168,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,256,0.014988443917698331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,1024,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,10240,0.17411555184258354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,5120,0.010676444404655032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,5120,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,51200,0.04060711132155524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,6144,0.343556457095676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,1024,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,256,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,512,0.007059555914666917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,3584,0.12170933352576362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,51200,0.13003644678327772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,16384,0.035820444424947105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,5120,0.024919999970330134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,10240,0.05037422312630547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,16384,0.02457511093881395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,6144,0.009948444035318162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,6144,0.009388444324334463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,1536,0.02233955595228407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,256,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,1536,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,768,0.01739555597305298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,2048,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,3584,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,2560,0.01311288856797748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,256,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,51200,13.733092414008247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,6144,0.14744799666934544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,4096,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,2560,0.3312968942854139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,12288,0.28605511453416615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,7168,0.045873777733908765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,2048,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,12288,0.0790319972568088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,6144,0.08562222454282974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,4096,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,3072,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,10240,0.017771555317772757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,6144,0.014255111416180929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,512,0.006718222051858902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,51200,0.9263839721679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,12288,0.20873333348168266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,10240,0.023718222975730896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,2560,0.031453334622912936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,7168,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,8192,0.10649510886934067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,3584,0.026613333159022864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,2560,0.02131555477778117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,12288,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,5120,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,7168,0.08186311191982694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,1024,0.003625777860482534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,65536,0.06328622500101726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,32,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,2048,0.0041857775714662345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,12288,0.5572835604349772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,12288,0.04074755642149184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,1536,0.005976888868543837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,1024,0.018613333503405254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,12288,0.004942222187916438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,1024,0.005761777775155173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,7168,1.467500474717882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,512,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,10240,0.029319110843870375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,51200,0.4219964345296224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,3584,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,8192,0.0350542234049903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,5120,0.018710222509172227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,4096,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,51200,0.9919662475585938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,5120,0.060453335444132485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,3072,0.01705511079894172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,12288,0.04345155424541897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,10240,0.190156446562873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,256,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,64,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,1024,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,3072,0.8040409088134766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,3584,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,7168,0.02567377852069007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,1024,0.005375111268626318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,6144,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,1536,0.0331928895579444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,512,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,8192,0.1143875519434611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,2560,0.04407733347680834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,3584,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,256,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,65536,0.15541422367095947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,1536,0.003839111162556542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,51200,68.01655409071181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,6144,0.05376177695062426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,10240,0.0547235541873508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,65536,0.33009688059488934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,7168,0.012336889074908363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,5120,0.012200888660218982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,768,0.021167111065652635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,128,0.030914667579862807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,256,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,5120,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,65536,0.9080702463785807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,7168,0.41239378187391496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,128,0.0169813334941864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,256,0.0024480000138282776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,4096,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,6144,0.1362933317820231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,7168,0.028162668148676556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,256,0.01960177719593048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,512,0.019282667173279654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,3072,0.006773333168692059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,256,0.023581332630581323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,1024,0.01719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,1024,0.0053342220683892565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,7168,0.7013431125217013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,2560,0.029038220643997192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,10240,0.012090666426552666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,10240,0.030418667528364394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,64,0.05877066983116997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,1024,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,12288,0.025314667158656653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,1536,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,5120,0.03674755493799845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,768,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,2560,0.017494221528371174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,2048,0.0520977775255839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,10240,0.06463111109203763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,2048,0.38488178782992893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,2048,0.013754666679435305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,2560,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,10240,0.07429066631529066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,8192,0.022496889034907024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,768,0.014160888062583076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,3584,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,768,0.015079999963442484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,768,0.03763377666473389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,12288,0.006561777657932705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,1024,0.021298666795094807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,65536,0.1223413281970554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,1536,0.02506133251720005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,3072,0.02977688776122199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,65536,0.15457867251502141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,2560,0.014148443937301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,512,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,65536,0.33864445156521267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,256,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,51200,0.11057155662112766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,768,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,512,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,7168,0.0517137779129876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,3584,0.10044711165957981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,3584,0.04753155509630839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,5120,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,512,0.02176977859603034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,65536,0.1799253357781304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,4096,0.008708444734414419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,128,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,768,0.0038613333470291565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,64,0.006125333408514659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,51200,0.03773066732618544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,10240,0.03161955542034573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,7168,0.029789331886503432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,8192,0.030010667112138536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,2048,0.04410488737954033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,768,0.03363822235001458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,8192,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,16384,0.010822222464614444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,128,0.07154933611551921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,6144,0.39256710476345485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,65536,0.728149308098687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,5120,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,6144,0.16442222065395778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,512,0.009425777528021071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,256,0.026371555195914373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,51200,0.11765866809421116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,1536,0.02512888941499922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,3072,0.019647111495335896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,6144,0.5586622026231554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,256,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,3072,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,12288,0.6009386910332574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,256,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,4096,0.05926489167743259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,768,0.018541332748201158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,8192,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,8192,0.038409776157803006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,16384,0.28553512361314565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,3584,0.03602577911482917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,512,0.07716977596282959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,1536,0.027313777142100867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,2560,0.010460444622569613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,2560,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,1024,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,51200,0.6582284503512913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,16384,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,3584,0.08835911088519627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,7168,0.027445332871543035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,2560,0.037187556425730385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,3584,0.16193600495656332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,6144,0.017736888594097562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,4096,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,8192,0.02402399977048238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,51200,0.05767822265625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,16384,0.042706665065553456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,64,0.010562666588359408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,16384,0.03754666778776381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,512,0.009939555492666032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,51200,1.2125635147094727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,1024,0.0307262208726671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,4096,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,768,0.015984889533784654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,2048,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,64,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,16384,0.1598115497165256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,32,0.004238222208287981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,51200,0.05293066634072197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,4096,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,768,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,12288,0.045534223318099976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,3072,0.016576889488432143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,5120,0.0247022228108512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,6144,0.019164444671736825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,5120,0.04220266805754768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,10240,0.15941066212124294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,128,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,6144,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,16384,0.061612447102864586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,8192,0.035976890060636736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,1024,0.06519466638565063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,5120,0.01978755493958791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,8192,0.19059556060367158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,1024,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,256,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,256,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,16384,0.046485334634780884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,8192,0.024332443873087566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,32,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,768,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,10240,0.012320000264379712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,2048,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,1024,0.12015733453962539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,64,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,10240,0.01833777791923947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,128,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,4096,0.13422221607632107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,10240,0.36960887908935547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,512,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,6144,0.21452000406053331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,2048,0.01990133358372582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,6144,0.011976888610257043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,5120,0.033760001262029014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,1024,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,65536,0.07692444324493408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,16384,0.18332178062862822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,3072,0.0075795559419526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,8192,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,3072,0.018567111757066514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,3072,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,2048,0.017469333277808297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,5120,1.0595226287841797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,1536,0.018522666560279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,768,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,3584,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,2048,0.010905777414639791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,5120,0.12885510921478271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,51200,0.059304886394076876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,16384,0.5107955402798122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,12288,0.063664886686537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,768,0.017488888568348356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,3072,0.050253334972593516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,512,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,3072,0.007182222273614671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,2048,0.004547555413511065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,1024,0.03560355636808608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,2560,0.018616000811258953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,5120,0.06260355313618977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,3072,0.013822221921549903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,3584,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,3072,0.008561777571837107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,8192,0.17342844274308947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,16384,0.03352355625894334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,768,0.005720888988839255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,12288,0.12054577138688828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,768,0.02781244450145298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,6144,0.07720000214046903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,6144,0.1110924416118198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,128,0.01566844516330295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,4096,0.025388444463411968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,7168,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,256,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,8192,0.014398222168286642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,3072,0.02872177627351549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,256,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,1536,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,256,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,16384,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,1536,0.01811288959450192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,12288,0.03561955690383911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,128,0.009973333113723332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,6144,0.02125955621401469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,51200,0.07201244433720906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,3584,0.06913866599400838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,5120,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,12288,0.059418665038214795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,512,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,10240,0.3702053229014079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,3584,0.0322053333123525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,51200,0.15060977141062418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,64,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,16384,0.2508808771769206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,768,0.00563733321097162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,1536,0.014242667290899487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,3072,0.029930667744742498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,3584,0.009878222313192155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,32,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,512,0.007104888558387756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,5120,0.024517332514127094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,256,0.002327999927931362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,65536,0.4303297731611464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,10240,0.04825599988301595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,2560,0.036660446061028376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,32,0.13594488302866617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,5120,0.3018302122751872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,10240,0.04286400145954556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,512,0.0029484445436133277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,10240,0.0052933332820733385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,2048,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,128,0.01701244380739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,16384,0.032531556155946516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,128,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,10240,0.03758755657407973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,51200,0.1723324457804362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,12288,0.01347644461525811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,5120,0.04388711187574598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,3072,0.014127110441525778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,3072,0.013002666334311167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,65536,0.6758977572123209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,5120,0.049765331877602466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,3584,0.029153777493370905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,6144,0.10986133416493733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,3072,0.010915555887752108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,1024,0.0041680000722408295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,3072,0.023240889112154644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,1024,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,51200,0.08284088638093737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,10240,0.01532977819442749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,5120,0.5452817810906304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,4096,0.008077333370844523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,1536,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,512,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,1024,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,3584,0.0664079984029134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,8192,0.014323555760913424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,1024,0.02900800108909607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,2560,0.5080106523301866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,4096,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,3072,0.0586106644736396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,512,0.021309332715140447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,3584,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,5120,0.008896888958083259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,10240,0.016473778420024447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,5120,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,12288,0.05235822333229912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,65536,0.8483822080824109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,2560,0.040307554933759905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,65536,0.1644044452243381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,2560,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,5120,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,4096,0.009773333039548662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,6144,0.049591110812293164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,64,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,4096,0.03254844413863288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,1536,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,8192,0.05742044581307305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,7168,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,3584,0.01688799924320645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,16384,0.17134666442871094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,2560,0.033131556378470525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,768,0.019858666592174105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,16384,0.025427555044492085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,2560,0.005040888984998067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,10240,3.5935482449001737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,16384,0.12943111525641546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,8192,0.030951112508773804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,8192,0.030846221579445735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,10240,0.06844355662663777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,256,0.015787555111779105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,10240,0.05843377775616116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,4096,0.034344888395733304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,3584,0.013957333233621387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,768,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,8192,1.1319857703314888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,1536,0.021075555019908484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,128,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,1536,0.023057777020666335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,1536,0.04861866765552097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,7168,0.013241777817408243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,51200,0.11528621779547797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,12288,0.2952924569447835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,10240,0.023726221587922838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,4096,0.025756445195939805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,8192,0.02271822260485755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,256,0.11572089460160996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,8192,0.05003199974695841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,51200,0.09143466419643825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,1536,0.022841778066423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,512,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,1536,0.004508444418509801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,8192,0.043346666627460055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,6144,0.06905066967010498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,3584,0.03404355711407132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,2560,0.004270222038030624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,4096,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,5120,0.10525155729717678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,7168,0.02834222382969326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,6144,0.08897244267993504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,32,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,10240,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,4096,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,3584,0.03503911031617059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,128,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,4096,0.011250666446155973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,16384,0.0458568897512224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,512,0.014900444282425774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,4096,0.12359644307030572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,3072,0.18320355150434706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,128,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,10240,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,51200,0.021393777595625982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,2048,0.05782133340835571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,10240,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,16384,0.8765510982937283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,12288,0.03216266632080078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,3584,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,256,0.17243466112348768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,8192,0.03294666608174642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,512,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,7168,0.06608533196979098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,128,0.015802666544914246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,8192,0.023235556152131822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,3072,0.033215999603271484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,3072,0.027624888552559748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,64,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,65536,0.22907021310594347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,1536,0.033498668008380465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,2560,0.004932444542646408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,16384,0.16887822416093615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,512,0.0069919998447100324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,10240,0.01606044504377577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,65536,0.09018222490946452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,128,0.01756977703836229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,6144,0.11366132895151775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,5120,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,65536,0.12142933739556207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,2560,0.05157688922352261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,256,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,8192,0.012337778177526263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,3584,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,12288,0.028985778490702312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,32,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,7168,0.06946400139066908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,3072,0.19049244456821016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,16384,0.22886578241984049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,32,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,1536,0.0057813334796163775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,2048,0.01475822263293796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,8192,0.09356355667114258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,512,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,5120,0.020816889074113634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,16384,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,1536,0.009926222264766693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,5120,0.06037422021230062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,3584,0.007320000065697565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,7168,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,12288,0.101492444674174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,7168,0.01904533306757609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,768,0.022109333011839125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,6144,0.01814933286772834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,768,0.02296977738539378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,512,0.006337777607970768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,7168,0.28153954611884224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,64,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,512,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,512,0.015896888242827523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,4096,0.006561777657932705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,768,0.016570667425791424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,16384,0.037167112032572426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,3072,0.2923591136932373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,65536,1.4777741962009008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,65536,0.2502737840016683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,10240,0.1550568872027927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,128,0.02312622302108341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,10240,0.023286221755875483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,64,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,8192,0.039388444688585066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,5120,0.027446221974160936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,2560,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,768,0.01721600029203627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,10240,0.07678221993976168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,768,0.03358133302794562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,51200,0.8467280069986979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,65536,0.1474791102939182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,32,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,128,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,256,0.01828888886504703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,512,0.0351128876209259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,7168,0.025415110919210646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,1024,0.0031431110368834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,8192,0.18608889314863417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,256,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,768,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,1536,0.04868622289763557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,128,0.01460088955031501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,65536,0.8601253297593859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,3584,0.011317333413494958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,51200,0.10291733344395955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,16384,0.042528887589772545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,256,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,51200,1.1231066385904949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,12288,0.05255822340647379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,4096,0.01276000009642707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,1024,0.011023999916182624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,12288,0.09585244125790066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,5120,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,10240,0.4241537782880995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,3072,0.030947556098302204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,12288,0.06113688813315498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,768,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,6144,0.010132444401582083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,51200,0.09916533364189996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,128,0.009743111001120673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,10240,0.02569777766863505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,128,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,16384,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,2048,2.8369216918945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,1536,0.19883288277520073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,7168,0.13235644499460855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,64,0.009911110831631554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,10240,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,3584,0.1141200065612793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,1536,0.12007821930779351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,256,0.00618488879667388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,2560,0.01715555621518029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,256,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,8192,0.06461155414581299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,3584,9.108853658040365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,64,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,5120,0.05838311380810208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,2560,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,7168,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,768,0.03939822316169739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,1024,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,512,0.019880889190567862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,2048,0.004584889031118817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,1536,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,6144,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,12288,0.06633689006169637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,512,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,6144,0.13337688975863987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,1536,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,2048,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,2048,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,512,0.020408888657887776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,32,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,3584,0.014316444595654806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,2560,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,256,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,16384,0.03278222348954942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,5120,0.11611910661061604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,65536,0.13548621866438124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,768,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,51200,0.4098995526631673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,768,0.0028817777832349143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,65536,0.13157688246832952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,51200,3.6432452731662326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,10240,0.022835556003782485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,51200,0.2112222247653537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,51200,0.02316088808907403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,1024,0.06454400221506755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,1536,0.010210666391584609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,7168,0.018793778287039865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,256,0.026384888423813715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,128,0.011244444383515252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,2048,0.021181333396169875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,1024,0.005986666513813867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,2048,0.026058667235904272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,2560,0.02148444453875224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,512,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,8192,0.11061600181791519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,2048,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,12288,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,16384,0.06006044149398804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,1536,0.0605306691593594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,5120,0.07148089011510213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,4096,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,51200,0.050147556596332125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,2560,0.051698668135537036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,6144,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,64,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,5120,0.12158221668667263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,1024,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,4096,0.10103555520375569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,5120,0.009129777550697327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,5120,0.04375822345415751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,8192,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,512,0.01309866706530253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,51200,0.37570932176378036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,3584,0.04342844420009189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,1024,0.02089244458410475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,16384,0.07509155405892266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,3584,0.014126222994592456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,768,0.01998844411638048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,256,0.015186667442321777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,256,0.01554844445652432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,6144,0.8999457889133029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,256,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,512,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,512,0.0028942223224374983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,3072,0.7477626800537109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,768,0.007089777953094906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,32,0.0063368889192740125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,6144,0.029810667037963867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,768,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,8192,0.02128000060717265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,2560,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,1536,0.04610133171081543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,32,0.017495110630989075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,8192,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,6144,0.06709600157207914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,10240,0.15502933661142984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,10240,0.01564711166752709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,51200,0.36368266741434735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,16384,0.04680800106790331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,1024,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,6144,0.13686488734351263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,8192,0.04500888784726461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,3584,0.029608001311620075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,32,0.010104888843165504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,4096,0.01574488812022739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,128,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,7168,0.02330844435426924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,768,0.018031110366185505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,12288,0.32656534512837726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,10240,0.03417155477735732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,4096,0.04594577683342827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,2560,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,256,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,5120,0.008944888909657797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,8192,0.09060533179177178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,4096,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,10240,0.02912799848450555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,1536,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,12288,0.019849777221679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,1024,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,5120,0.0603004429075453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,512,0.03415377934773763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,12288,0.035813334915373064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,2560,0.032398223876953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,4096,0.03129955463939243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,7168,0.02365333338578542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,3584,0.06823466883765326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,512,0.01443911095460256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,6144,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,6144,0.09195466836293538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,7168,0.01922044489118788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,768,0.011736888852384357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,1536,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,768,0.004129777765936322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,8192,0.06735199689865112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,2048,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,512,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,768,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,12288,0.017494221528371174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,3072,0.016509332590632968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,8192,0.01108088923825158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,4096,0.02793955471780565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,3584,0.007678222325113084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,6144,0.04205866654713949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,65536,0.06702311171425714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,1024,0.00582755563987626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,3584,0.07666666640175714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,2048,0.1035519970787896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,8192,0.7336524327596029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,3584,0.021976000732845728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,5120,0.06670488913853963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,10240,0.00928177767329746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,10240,0.5089742342631022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,7168,0.0786693361070421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,128,0.02685600022474925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,4096,0.03955466548601786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,7168,0.012069332930776807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,512,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,64,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,6144,0.028911110427644517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,65536,0.9396488401624891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,5120,0.05144266618622673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,7168,0.03860355416933695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,6144,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,2048,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,128,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,7168,0.048222220606274076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,512,0.023139556248982746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,12288,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,12288,0.05539733171463013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,1536,0.05788355403476291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,2048,0.009401777552233802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,32,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,7168,0.01372977760103014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,5120,0.013996443814701505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,3072,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,32,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,2048,0.006383111079533895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,2560,0.02782755427890354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,3072,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,512,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,512,0.12067200077904595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,16384,3.336660385131836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,1024,0.008594666918118795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,4096,0.07406755288441975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,1024,0.008810666700204214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,1536,0.012732444538010491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,3584,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,2560,0.06230666902330187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,6144,0.05366666780577766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,3584,0.1183297766579522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,12288,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,1024,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,256,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,7168,0.21495555506812203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,128,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,256,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,51200,0.15904800097147623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,256,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,4096,0.008911111288600499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,4096,0.029616001579496596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,5120,0.043077333105934985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,3072,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,12288,0.03565866748491923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,5120,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,7168,0.8098346922132703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,65536,2.4597030215793185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,3584,0.11874043941497803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,128,0.02382133404413859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,3072,0.01824711097611321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,5120,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,3072,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,2048,0.3448568979899089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,12288,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,8192,0.2559537887573242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,6144,0.037032001548343234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,51200,0.07426044676038954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,7168,0.02369955513212416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,6144,0.025158221522967022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,10240,0.010672888822025724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,1536,0.00433688900536961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,7168,0.011286222272449069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,4096,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,51200,0.09319200118382771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,32,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,1536,0.013238222234778933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,10240,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,12288,0.02311644454797109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,128,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,256,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,5120,0.023800000548362732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,768,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,65536,0.06253866354624431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,3584,0.016753777861595154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,10240,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,7168,0.05258666806750827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,3072,0.0206977774699529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,768,0.005793777604897817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,64,0.011004444625642566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,3072,0.022279111875428095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,8192,0.36320087644788956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,10240,0.32104267014397514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,1024,0.009480000370078618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,65536,0.36865777439541286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,4096,0.015787555111779105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,12288,0.016935111747847665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,768,0.01763288842307197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,768,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,768,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,10240,0.01867288847764333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,1024,0.021168000168270532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,768,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,65536,0.2001404497358534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,4096,0.01681511104106903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,5120,0.020536889632542927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,10240,0.015944000747468736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,2048,0.03320088982582092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,512,0.013061333033773633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,256,0.0072266666425599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,16384,0.09028355280558269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,3584,0.02713511056370205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,512,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,1536,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,12288,0.022429333792792425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,6144,0.043192889955308705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,6144,0.03388977713055081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,3072,0.12087733215755886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,3072,0.03643111056751675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,51200,0.17916622426774767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,10240,0.017210667332013447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,16384,0.0634488860766093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,7168,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,65536,0.05312800076272753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,6144,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,3072,0.029615110821194116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,12288,0.017694221602545846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,64,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,256,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,3584,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,2048,0.08455644713507758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,65536,0.10702044434017605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,8192,0.024637333220905725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,64,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,1024,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,1536,0.09122221999698216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,4096,0.02595022155178918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,10240,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,768,0.002785777880085839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,32,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,2048,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,10240,0.03153333399030898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,2048,0.04000622365209792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,3584,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,3072,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,51200,0.06859199868308173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,16384,0.40818132294548887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,4096,0.4281493292914496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,12288,0.0425164434644911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,4096,0.02036177780893114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,5120,0.05924533473120796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,1024,0.011439111497667102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,512,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,512,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,768,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,768,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,32,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,51200,0.9510320027669271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,65536,0.23290756013658312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,3072,0.01861422260602315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,2560,0.03319911162058512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,1536,0.022330666581789654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,512,0.025947555899620056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,10240,0.030007110701666936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,12288,0.05563022030724419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,6144,1.453203519185384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,2048,0.0337084432442983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,6144,0.054415110084745616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,12288,0.017758222089873422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,5120,0.010100444157918295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,65536,0.25961867968241376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,5120,0.1264639960394965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,3072,0.022875555687480505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,512,0.09191644191741943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,12288,0.0936711099412706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,5120,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,768,0.026766222384240892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,64,0.007703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,5120,0.9926719665527344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,4096,0.01982666717635261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,2560,0.018175111876593698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,51200,0.17711022165086535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,2048,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,2048,0.6991022427876791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,1536,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,2560,0.025736888249715168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,4096,0.032454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,7168,0.028204444381925795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,6144,0.021312889125612047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,65536,1.6764613257514105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,10240,0.1336488856209649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,4096,0.028025776147842407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,2560,0.489532470703125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,64,0.00328533330725299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,64,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,128,0.016196444630622864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,3072,0.013274666335847644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,128,0.00552977787123786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,2048,0.007171555525726742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,1024,0.022643556197484333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,16384,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,6144,0.016142222616407607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,7168,0.17319733566708037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,12288,0.021583999196688335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,128,0.03896355628967285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,16384,0.047028445535235934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,64,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,2560,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,6144,0.019588443968031142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,8192,0.07095644209120008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,5120,0.04741511080000135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,3072,0.004618666652176115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,8192,0.03190755512979295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,2560,0.0075777777367168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,10240,0.05445333321889242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,51200,0.11483466625213623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,2560,0.04260444309976366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,512,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,2560,0.01700088878472646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,16384,0.29793866475423175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,3072,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,32,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,51200,0.05044800043106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,512,0.01643200053109063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,10240,0.1876088910632663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,1536,0.006872889068391588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,256,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,10240,0.008807111117574904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,12288,0.019927110936906602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,51200,0.05168799890412224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,1024,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,16384,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,4096,0.00906399968597624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,128,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,3584,0.11171200540330674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,8192,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,1536,0.003654222107595868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,65536,0.5368960168626573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,16384,0.5085155698988173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,6144,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,65536,0.6372159851921929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,6144,0.47498491075303817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,10240,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,512,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,7168,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,8192,0.02627022233274248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,16384,0.1795813375049167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,2048,0.0599493318133884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,64,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,51200,0.1104782223701477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,128,0.08047111166848077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,6144,0.013268444273206921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,3072,0.014939554863505893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,6144,0.025996444953812495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,768,0.003142222141226133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,12288,0.014120888378885059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,3584,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,8192,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,1536,0.024029332730505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,2560,0.03803022371398078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,512,0.04677866564856636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,16384,0.052645332283443876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,51200,0.32516177495320636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,32,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,4096,0.031735112269719444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,65536,0.06546133094363742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,768,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,128,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,10240,0.014731556177139282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,1536,0.029504001140594482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,128,0.065138664510515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,2048,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,65536,0.07154844204584758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,1536,0.008900444540712569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,3072,0.007420444654093848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,7168,0.033639109796947904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,256,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,768,0.016524444023768108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,6144,0.02232355541653103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,5120,0.030026667647891577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,51200,2.7804427676730685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,2048,0.059847997294531934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,3072,0.02884977724817064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,3072,0.051046222448349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,3072,0.11537155840131973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,8192,0.013656888571050433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,12288,0.05040622088644239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,3072,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,16384,0.06505955590142144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,12288,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,3584,0.011427555647161273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,3584,0.027439999911520217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,768,0.04279466801219516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,4096,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,32,0.0018782222436534034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,1024,0.010576888918876648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,51200,0.9116053051418729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,1536,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,51200,0.11892444557613796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,3072,0.01921599937809838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,65536,0.5037084685431586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,1536,0.00738933351304796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,1536,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,10240,0.01297155519326528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,256,0.007700444095664554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,4096,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,1536,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,2048,0.008363555702898238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,2560,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,512,0.017045333981513977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,12288,0.10620888736512925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,3072,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,8192,0.011574221981896294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,256,0.014361777239375643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,256,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,12288,0.05494488941298591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,12288,0.019709333777427673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,128,0.006768888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,51200,1.988684548272027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,3584,0.04626577761438158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,128,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,7168,0.06478399700588651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,256,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,16384,0.16247022151947021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,12288,0.014743111199802823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,2560,0.014876445134480795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,3584,0.029620445436901514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,128,0.020631111330456205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,1024,0.21797777546776664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,1536,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,51200,0.05152177810668945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,1536,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,256,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,2560,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,128,0.012048000262843238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,2048,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,6144,0.04930488930808174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,32,0.0018622221218215094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,768,0.013904000322024027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,5120,0.0765155553817749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,8192,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,4096,0.6722346411810981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,7168,0.015431111057599386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,65536,0.22746756341722277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,768,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,4096,0.01035288886891471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,16384,0.03758222195837233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,1024,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,3584,0.03875733415285746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,1536,0.03797511259714762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,512,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,10240,0.016203555795881484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,7168,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,256,0.02772533396879832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,32,0.0022133334229389825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,1024,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,1024,0.09722310966915554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,5120,0.026730666557947796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,7168,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,51200,1.8052711486816406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,64,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,10240,0.021426666114065383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,768,0.050586667325761586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,768,0.03774844606717428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,1536,0.019223110543357003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,16384,0.18844889269934761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,12288,0.046670221620135836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,10240,0.014352000421947904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,3072,0.01553422212600708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,3584,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,16384,0.0854675571123759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,256,0.013223999904261695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,3584,0.017691555950376723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,512,0.006827555596828461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,512,0.025962667332755193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,5120,0.008936888641781276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,5120,0.008553778131802877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,3072,0.7071084446377225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,1024,0.008366222182909647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,12288,0.019575110740131803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,12288,0.032857778999540545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,10240,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,10240,0.22403377956814238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,2560,0.057110223505232066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,2560,0.024259555670950148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,65536,0.05863910913467407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,1024,0.014757333530320061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,65536,0.07544710901048449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,2560,0.008525333470768398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,768,0.020601777566803824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,1024,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,4096,0.12003200583987766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,1536,0.006406222366624408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,4096,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,6144,0.056962668895721436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,65536,0.7112488746643066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,768,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,7168,0.13034489419725206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,65536,0.46451732847425675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,5120,0.047934220896826855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,1024,0.0035217776894569397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,4096,0.10740444395277236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,1024,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,12288,0.6904684702555338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,51200,0.017656889226701524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,3584,0.03465866711404588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,12288,0.03086755673090617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,1024,0.011243555280897351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,3584,0.16270932886335585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,6144,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,768,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,1024,0.0032195556494924757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,256,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,128,0.015388444066047668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,4096,0.020745777421527438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,64,0.0016231111561258633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,768,0.010016889207892949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,16384,0.014679110712475248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,768,0.03245066603024801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,6144,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,5120,0.017255110873116385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,6144,0.015744000673294067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,12288,0.017140444782045152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,3584,0.0053004444473319584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,2560,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,128,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,3072,0.3413280116187201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,7168,0.29842400550842285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,512,0.013342222405804528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,2048,0.01570666664176517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,51200,0.1760986646016439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,2048,0.024360888534122046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,6144,0.013016000390052795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,16384,0.2990017731984456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,65536,0.06567111280229357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,3072,0.017108443710539076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,2560,0.007156444092591603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,3072,0.08418844143549602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,16384,0.024759110477235582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,3584,0.41015288564893937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,4096,0.032131556007597185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,8192,0.018704889549149405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,1536,0.0069982219073507525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,65536,16.09933132595486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,5120,0.12025066216786702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,128,0.02612000041537815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,64,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,32,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,6144,0.02578488820128971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,7168,0.030572444200515747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,8192,0.09141511387295193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,2560,0.030385779009924993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,1024,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,5120,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,512,0.014898666077189975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,3584,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,7168,0.044616887966791786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,12288,0.011548444628715515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,3584,0.013712888790501488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,512,0.007335111498832703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,512,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,6144,0.04759555392795139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,16384,0.03405422303411696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,1536,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,4096,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,12288,0.10716088612874348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,51200,0.12219288614061143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,2560,0.020405333903100755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,1536,0.037865777810414634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,16384,0.04537422127193875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,8192,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,1024,0.008386666576067606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,128,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,5120,0.0586106644736396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,1024,0.016007999579111736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,6144,0.020435555113686454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,12288,0.03393777873780992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,768,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,6144,0.08874933587180243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,4096,0.011965333587593503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,4096,0.042527112695905894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,8192,0.004585777719815572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,3584,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,64,0.0039057777159743835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,3072,0.01959111127588484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,1536,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,65536,26.456190321180557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,12288,0.013422222601042854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,5120,0.18269066015879312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,51200,0.07721066474914551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,512,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,128,0.023346667488416035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,128,0.0440959996647305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,64,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,65536,0.6463982264200846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,3584,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,3584,0.0750639968448215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,1024,0.034274667501449585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,2048,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,12288,0.08973422315385605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,3072,0.22815023528205025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,2560,0.010410666465759277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,2560,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,2048,0.06548888815773858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,7168,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,1024,0.003195555466744635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,4096,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,8192,0.15855644808875188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,65536,0.10211822059419419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,2048,0.03124089042345683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,768,0.01885688967174954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,3072,0.07147822115156385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,7168,0.33468977610270184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,128,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,2048,0.33922044436136883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,65536,0.18355733818478057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,10240,0.16321244504716662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,7168,0.017465778523021273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,6144,0.03494488861825731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,1024,0.019914666811625164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,3072,0.018968888454967074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,16384,0.025161777933438618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,1536,0.010696888797812991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,512,0.04984800020853678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,8192,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,3072,0.03673955467012193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,128,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,10240,0.012198222180207571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,65536,0.11942399872673883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,12288,0.017352888981501263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,7168,0.019355555375417072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,256,0.014628445108731588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,2048,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,128,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,12288,0.03252355588807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,12288,1.1466692818535698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,5120,0.02083911167250739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,6144,0.05066666669315762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,2048,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,65536,0.90047730339898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,768,0.04596977763705783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,2048,0.05791022380193075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,51200,0.26263199912177193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,32,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,3584,0.015240889456537036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,1024,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,12288,0.004967999955018361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,2560,0.004879111217127906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,64,0.0041866666740841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,8192,0.313935120900472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,16384,1.8121395111083984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,6144,0.5222222010294596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,5120,0.026776888304286536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,4096,0.2301093207465278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,10240,0.08104799853430854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,1024,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,2048,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,32,0.026501332720120747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,4096,0.08696177932951187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,3072,0.014548444085650973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,10240,0.018794667389657762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,512,0.023352000448438857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,32,0.19732088512844512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,51200,0.8603270848592123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,7168,0.04787910977999369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,10240,0.034056001239352755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,256,0.09316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,6144,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,2560,0.026611554953787062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,51200,1.6974088880750868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,5120,0.04668000009324816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,10240,0.031141334109836157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,10240,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,1536,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,1536,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,65536,0.24863023228115508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,128,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,6144,0.016113777955373127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,64,0.002577777744995223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,5120,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,10240,0.05749599801169502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,256,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,3072,0.015222221612930298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,65536,0.787468433380127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,3072,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,3584,0.010660444696744284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,256,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,1024,0.16096533669365778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,3584,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,8192,0.024657777614063684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,3072,0.03598489032851325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,256,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,51200,1.3809351391262477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,1536,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,2560,0.025690666503376428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,10240,0.01221422188811832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,4096,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,3584,0.02769244379467434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,8192,0.03307911091380649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,8192,0.20424266656239828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,8192,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,6144,0.02381333377626207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,768,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,8192,0.011173333558771344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,10240,0.15221244759029812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,7168,0.022669333550665114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,768,0.0028773333049482773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,1024,0.022267555197079975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,768,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,16384,0.014055111342006259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,1536,0.04548711246914334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,2560,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,65536,0.45751020643446183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,3072,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,512,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,65536,0.17038044664594862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,64,0.05862666500939263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,1536,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,128,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,256,0.023952000670962866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,7168,0.01535555554760827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,7168,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,2560,0.017343999611006845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,64,0.018834667073355783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,5120,0.020703110429975722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,16384,0.05055377880732218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,128,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,6144,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,256,0.006468444648716185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,51200,0.1923795541127523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,3072,0.03158222304450141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,2048,0.006789333290523953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,6144,0.010653333531485664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,512,0.030392001072565716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,3584,0.01997599999109904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,8192,0.07284444570541382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,51200,0.12949689229329428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,7168,0.271692434946696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,32,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,512,0.012298666768603854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,12288,0.015204444527626038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,3072,0.0846995578871833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,768,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,12288,0.16260266304016113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,51200,4.082319047715929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,4096,0.05303288830651177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,2048,0.006535111202134027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,16384,0.07891466882493761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,65536,0.07902577850553724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,4096,0.3405342102050781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,2048,0.07866399817996554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,10240,0.028612444798151653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,51200,0.07935733265346952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,64,0.0018151110659042995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,4096,0.008536888493431939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,6144,0.08535110950469971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,2048,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,8192,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,8192,0.03645066751374139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,6144,0.03185688787036472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,128,0.009826666778988307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,2560,0.029522667328516643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,12288,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,1536,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,10240,0.08504622512393528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,4096,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,16384,0.0369742214679718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,1536,0.1074097752571106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,128,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,8192,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,4096,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,256,0.03394666645261977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,1536,0.13367288642459446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,65536,0.39577423201666934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,6144,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,2048,0.03467911150720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,768,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,51200,0.07237778107325236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,16384,0.16166222095489502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,16384,0.025066667132907446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,2560,0.009117333425415887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,256,0.015186667442321777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,2560,0.014928888943460254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,12288,0.9046693378024631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,6144,0.0355751117070516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,32,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,512,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,51200,0.14637777540418836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,65536,0.011327111058764987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,128,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,4096,0.022855111294322546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,1024,0.22498933474222818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,5120,0.01756800048881107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,768,0.007702222300900354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,6144,0.0041697778635554844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,6144,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,128,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,1024,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,4096,0.02571644385655721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,16384,0.2825048764546712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,1536,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,1024,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,16384,0.022052443689770166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,8192,0.014642667439248828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,128,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,256,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,1024,0.03240088952912225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,65536,0.6371253331502279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,8192,0.06724622514512804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,7168,0.12553866704305014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,3584,0.07053066624535455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,1536,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,1024,0.0034248888906505373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,3584,0.009218666288587783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,7168,0.06136622031529745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,512,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,51200,0.12148622671763103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,3584,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,12288,0.19069688849978975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,3584,0.061153776115841336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,2048,0.05091911223199633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,768,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,7168,0.05297155512703789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,7168,0.18233333693610299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,65536,5.772292243109809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,1024,0.010126222338941362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,6144,0.016789333687888253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,6144,0.011738667057620155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,3072,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,16384,0.0554639995098114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,6144,0.023687111006842718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,2560,0.02267644471592373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,1536,0.003548444559176763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,4096,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,512,0.018571555614471436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,1536,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,256,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,5120,0.011992000043392181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,3584,0.02643022272321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,32,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,768,0.018364444375038147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,1536,0.008919999831252629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,16384,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,4096,0.10192177693049113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,7168,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,16384,0.1504960060119629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,1536,0.02420622275935279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,2048,0.03938577903641595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,65536,1.7580071555243597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,65536,0.09987911250856187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,12288,0.10388000143898858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,8192,0.04660355713632372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,1536,0.0042160000238153665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,8192,0.03488444288571676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,256,0.01719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,10240,0.44301154878404403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,51200,0.07434044281641643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,12288,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,1024,0.02405333353413476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,512,0.019517333971129525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,4096,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,7168,0.012305777933862476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,51200,0.36441956626044375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,1024,0.003195555466744635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,1536,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,5120,0.029765334394243028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,3584,0.01777866648303138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,512,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,3072,0.08259289132224189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,4096,0.021792888641357422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,10240,0.02674755619631873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,16384,0.09921155373255412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,51200,0.5828435685899522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,4096,0.1751013331943088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,65536,0.08004266685909696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,1024,0.02055644492308299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,12288,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,5120,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,65536,0.5408479902479384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,5120,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,1024,0.02311022248533037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,2560,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,512,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,8192,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,768,0.05813066826926338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,128,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,2560,0.03702222307523092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,128,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,4096,0.07973955737219916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,128,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,7168,0.0407315558857388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,4096,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,32,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,3584,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,128,0.021073778470357258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,128,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,3072,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,3072,0.007361777954631382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,768,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,2560,0.03526044554180569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,10240,0.016014221641752455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,51200,0.05063377817471822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,3584,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,2560,0.02978044417169359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,512,0.006114666660626729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,256,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,2048,0.18974933359358045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,65536,0.06712977753745185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,3584,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,8192,0.4913662274678548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,51200,0.07713955640792847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,2560,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,2048,0.019492444064882066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,8192,0.022031999296612207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,8192,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,64,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,2560,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,1536,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,51200,0.06519555383258395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,2048,0.19167021910349527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,12288,0.025394666526052687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,4096,0.031098667118284438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,65536,0.06487822532653809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,65536,0.0638257794910007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,2560,0.0273617770936754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,4096,0.020434666011068556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,6144,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,8192,0.024199111594094172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,16384,0.03938666648334927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,8192,0.03345244460635715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,16384,0.0751511123445299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,4096,0.15027199851142034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,512,0.008374222450786168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,12288,0.060439997249179415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,8192,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,5120,0.03223289052645365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,1536,0.027980443504121568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,512,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,2048,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,8192,0.01605955594115787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,12288,0.019884443945354886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,10240,0.031848000155554876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,2560,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,8192,0.019949333535300363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,7168,0.013770666387346057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,128,0.014583110809326172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,512,0.01924444403913286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,5120,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,6144,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,128,0.02961777647336324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,768,0.02307644486427307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,64,0.0025680000997251934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,5120,0.012522666818565793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,1024,0.04074133435885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,3072,0.0148062225845125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,2048,0.020277332928445604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,2560,0.0148062225845125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,8192,0.19544977611965605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,16384,0.2654577891031901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,256,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,2048,0.053174220853381686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,65536,0.5382115576002333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,10240,0.3654293219248454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,128,0.00776800016562144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,1024,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,10240,0.00775733341773351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,256,0.017553778158293832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,3584,0.04344088832537333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,3584,0.0210844443904029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,3072,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,64,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,16384,0.03702844513787164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,768,0.05052533414628771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,64,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,128,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,10240,0.02880711025661892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,65536,0.1547493272357517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,4096,0.0874933335516188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,16384,0.013960889644092984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,16384,0.07577955722808838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,65536,0.1348044474919637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,128,0.0019226666125986313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,5120,0.02382133404413859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,12288,0.04147911071777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,2560,0.03509599963823954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,7168,0.24963021278381348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,3072,0.011975111232863532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,5120,0.03566044569015503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,16384,0.050364444653193154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,8192,0.015778667396969266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,7168,0.411077340443929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,1536,0.03241777751180861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,1024,0.007800000409285228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,1536,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,512,0.0199937770764033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,7168,0.07624889082378812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,3072,0.03644711110326979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,51200,0.06514222092098661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,2560,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,512,0.021485333641370136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,12288,0.0509111119641198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,128,0.04502844479348925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,64,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,1536,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,6144,0.04370489054256015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,8192,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,2048,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,2560,0.01459377838505639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,3584,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,10240,0.012299555871221753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,512,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,2560,0.022109333011839125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,1536,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,16384,0.025234666135576036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,10240,0.12165689468383789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,512,0.04314577910635206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,7168,0.01995022263791826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,12288,0.03821422325240241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,2560,0.03200799889034695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,256,0.020657777786254883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,512,0.011794666449228922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,256,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,1536,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,3584,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,3584,0.015376889043384127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,1024,0.026328000757429335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,2560,0.026692445079485577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,5120,0.46592802471584743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,16384,0.07533155547247992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,7168,0.018223111828168232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,2048,0.08671555254194473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,12288,0.021960000197092693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,8192,0.042111999458736844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,8192,0.03537599907981025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,6144,0.055296887954076133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,128,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,2560,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,7168,0.01352355546421475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,65536,0.12191111511654323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,7168,0.02830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,4096,0.027782221635182697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,4096,0.03417866759830051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,65536,1.6747333738538954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,51200,0.06642755534913805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,512,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,3584,0.010432889064153036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,32,0.06667911344104342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,128,0.005417777846256892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,2048,0.02390844457679325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,4096,0.044158223602506846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,512,0.04573333263397217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,1536,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,51200,0.1140506664911906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,2560,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,8192,0.025626666016048853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,1536,0.10111466381284927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,1024,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,32,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,5120,0.011431111229790581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,51200,0.19323910607231987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,1024,0.0031386667655573953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,16384,0.26458311080932617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,128,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,4096,0.09764622317420112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,2560,0.026421333352724712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,7168,0.014305777019924588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,4096,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,4096,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,2048,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,3584,0.024862221545643274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,51200,0.11992177698347305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,4096,0.010382222632567087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,4096,0.010031999813185798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,12288,0.28423110644022626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,1536,0.014717333846622042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,1024,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,12288,0.10193955236011082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,6144,0.01351466692156262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,512,0.015106666419241162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,1024,0.056645333766937256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,512,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,1536,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,3584,0.016574222180578444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,2048,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,6144,0.024335111180941265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,256,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,65536,0.24890311559041342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,5120,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,4096,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,768,0.02137600051032172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,4096,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,256,0.09828977452384101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,5120,0.02885955572128296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,16384,0.02479733361138238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,51200,0.04552622305022346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,2560,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,32,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,65536,0.28682488865322536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,6144,0.06096533272001478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,3072,0.01607111096382141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,768,0.015061333775520325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,4096,0.6596817970275879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,65536,0.07530222336451213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,2048,0.08642311228646173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,3072,0.08993955453236897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,8192,0.02961777647336324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,32,0.013232000172138214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,1024,0.02908710969818963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,3072,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,5120,0.1550017727745904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,512,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,128,0.00328533330725299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,8192,0.010951110886202918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,6144,0.02141155633661482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,16384,0.01146933353609509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,1024,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,4096,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,512,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,32,0.04367555512322319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,512,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,1024,0.06184177928500705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,4096,0.07642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,7168,0.037774221764670476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,10240,0.016693333784739178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,2048,0.4658631218804254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,2048,0.023149333066410486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,16384,0.2365510993533664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,65536,0.0745484431584676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,4096,0.010645333263609143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,1024,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,512,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,12288,0.034862223598692156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,2048,0.04225511021084256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,3584,0.01236088905069563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,2048,0.01614755557643043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,32,0.005872000008821487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,16384,0.08185422420501709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,12288,0.032283554474512734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,2560,0.011592888997660743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,7168,0.05689422289530436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,128,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,51200,1.3920710881551106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,8192,0.018595554762416415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,65536,0.14829688602023655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,64,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,12288,0.3215840127733019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,7168,0.3921991189320882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,3072,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,1024,0.09171644184324478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,32,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,10240,0.018562666243977018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,32,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,512,0.009863999982674917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,1024,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,1536,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,2048,0.040553778409957886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,10240,0.030086222622129653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,16384,0.04899466700024075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,2048,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,5120,0.02364088926050398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,3584,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,1024,0.062275555398729115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,64,0.008903111020723978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,4096,0.009571555587980483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,1024,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,512,0.034669333034091525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,5120,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,3072,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,256,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,256,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,65536,0.06348622507519193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,7168,0.012116444607575735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,8192,0.18399377663930258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,1024,0.006320888797442119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,2048,0.044157332844204374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,1536,0.03166933192147149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,16384,1.2872977786593967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,4096,0.270576900906033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,3584,0.012522666818565793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,768,0.025775111383861963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,4096,0.04051555527581109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,51200,0.027406222290462915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,8192,0.014864888456132678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,768,0.02166844407717387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,128,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,64,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,2560,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,8192,0.010748444332016839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,5120,0.009135111338562435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,16384,1.3550675710042317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,256,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,3584,0.1066266695658366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,8192,0.047162665261162647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,51200,0.049145778020222984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,3072,0.3454879919687907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,4096,0.022084444761276245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,3072,0.033279998434914485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,6144,0.15793244043986002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,7168,0.037992000579833984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,32,0.019768888751665752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,1536,0.07100711266199748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,2048,0.0413680010371738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,10240,0.01698755555682712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,768,0.01943555474281311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,2560,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,51200,0.040018667777379356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,512,0.003354666754603386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,65536,1.5842435624864366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,512,0.012798222402731577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,3584,1.3515253067016602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,65536,0.09118666913774277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,16384,0.04801155461205376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,16384,0.28050133917066783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,1536,0.07707377937104967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,6144,0.0787573324309455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,3072,0.01629066632853614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,3072,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,2560,0.14751111136542425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,512,0.008402666283978356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,10240,0.02658133374320136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,1536,0.33055467075771755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,6144,0.038520889149771795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,4096,0.014071111877759298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,65536,0.108551111486223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,5120,0.10991733604007298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,1536,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,256,0.012127999630239276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,5120,0.0630906687842475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,64,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,1536,0.011342222491900126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,4096,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,7168,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,12288,0.04274133178922865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,1024,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,2560,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,10240,0.37900532616509336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,256,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,16384,0.052592890130149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,4096,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,128,0.01443911095460256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,3584,0.012917333179050021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,1536,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,2048,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,32,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,2560,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,128,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,51200,0.053039110369152494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,5120,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,2048,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,65536,0.07340088817808363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,128,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,7168,0.010308444499969482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,16384,0.0746871100531684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,1536,0.039220445685916476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,32,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,5120,0.02754844393995073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,1536,0.010088000032636855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,8192,0.011420444481902652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,64,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,6144,0.012727999852763282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,5120,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,8192,0.025399999486075506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,5120,0.016140444411171805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,65536,0.37766043345133465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,12288,0.056517336103651256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,1536,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,3584,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,10240,0.11585599846310085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,6144,0.0476471119456821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,4096,0.01670222149954902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,6144,0.5374177826775445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,8192,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,2048,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,3072,0.004587555511130227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,256,0.03200888964864943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,2560,0.10124888685014512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,6144,0.10318044159147476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,12288,0.013192888763215808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,5120,0.012150222228633033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,1024,0.10740000009536743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,1536,0.030534221066368952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,2560,0.019541333119074505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,3072,0.030031111505296495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,3584,0.06833777825037639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,32,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,8192,0.018601778480741713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,65536,1.0898959901597765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,65536,0.13991110854678684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,7168,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,2560,0.04774311184883118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,4096,0.024865777956114873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,8192,0.06724977493286133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,10240,0.04099555479155646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,16384,0.2623342143164741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,256,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,4096,0.047930667797724404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,8192,0.01408622165520986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,8192,0.02537244392765893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,51200,1.8211599985758464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,7168,0.013730666703648038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,3072,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,8192,0.010790222220950656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,128,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,256,0.006066666709052191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,32,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,768,0.09968800014919704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,8192,0.016173332929611206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,3072,0.05340266558859083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,12288,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,51200,0.06910222106509738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,1536,0.029130667448043823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,65536,0.06292266978157891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,3072,0.021980444590250652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,256,0.01182755579551061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,6144,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,16384,0.047260443369547524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,64,0.010813333094120026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,16384,0.18898933463626436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,128,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,6144,0.018027555611398485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,5120,0.11250310473971897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,768,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,3584,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,51200,0.5075182384914821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,10240,0.06068888637754652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,2560,0.04288977715704176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,16384,0.05946222278806898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,768,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,2560,8.176268683539496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,3072,0.011520889070298938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,16384,0.02718933257791731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,64,0.04972355564435323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,2048,0.019172444939613342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,4096,0.22294312053256563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,7168,0.18885421752929688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,7168,0.013938667045699226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,4096,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,1024,0.02036355601416694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,7168,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,2560,0.04426577687263489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,3584,0.02019999921321869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,65536,0.049257778459125094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,65536,0.061032891273498535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,2560,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,65536,0.3841555648379856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,8192,0.01609777741962009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,512,0.018626666731304593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,12288,2.0159901512993708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,10240,0.06361689170201619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,6144,0.040431999497943454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,16384,0.021987555755509272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,2560,0.01740977830357022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,4096,0.011607111328177981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,8192,0.016768889294730294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,8192,0.059644447432623975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,51200,0.548725340101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,3072,0.018236445056067575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,512,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,6144,0.08806577656004165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,8192,0.040031999349594116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,1536,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,10240,0.048143111997180514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,32,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,6144,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,1024,0.032163557079103254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,65536,0.29057510693868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,7168,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,8192,0.022981334063741896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,12288,0.3650008837381999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,256,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,65536,0.24806043836805555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,65536,0.17267378171284994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,768,0.027858667903476294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,6144,0.025715554753939312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,2560,0.025412445267041523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,16384,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,65536,0.25397777557373047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,32,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,2560,0.40754400359259707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,3072,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,256,0.04204622242185804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,16384,0.03329244587156508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,16384,0.08092355728149414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,64,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,128,0.021712889273961384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,3584,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,32,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,10240,0.18811021910773384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,10240,0.014504889647165934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,128,0.005097777893145879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,2048,0.028412444723976985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,8192,0.02072888943884108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,768,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,2048,0.028920888900756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,1536,0.020264888803164165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,7168,0.021329777108298406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,2560,0.016623111234770883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,3072,0.0553653339544932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,1024,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,16384,0.845793777041965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,64,0.001773333383931054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,512,0.0032142222755485107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,16384,0.036828445063696966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,128,0.010580444501505958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,10240,0.023690667417314317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,1024,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,6144,0.17030044396718344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,4096,0.05115200082461039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,51200,0.06265599860085382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,32,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,65536,0.07413066758049859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,64,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,512,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,512,0.020096888144810993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,7168,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,7168,0.05542666713396708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,768,0.00869599978129069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,7168,0.07101866934034559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,4096,0.027666666441493567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,1536,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,4096,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,6144,0.07331289185418023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,51200,0.42587465710110134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,8192,0.10741155677371556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,768,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,768,0.020797333783573575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,256,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,2048,0.03143289022975498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,1024,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,7168,0.07073066631952922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,2560,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,65536,0.45510043038262266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,128,0.002442666639884313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,8192,0.035283555587132774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,2048,0.014168888330459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,32,0.018671999375025432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,3584,0.015429332852363586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,256,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,128,0.010568888651000129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,3072,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,6144,0.03128177920977274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,512,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,768,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,3584,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,6144,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,128,0.020033778415785897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,3584,0.007582222421964009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,16384,0.022343110707071092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,512,0.022650667362742957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,64,0.007899555895063613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,768,0.02808977829085456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,2048,0.03882666760020786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,512,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,10240,0.07586133480072021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,51200,0.09560177723566692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,2560,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,2560,0.020397333635224234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,256,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,4096,0.018014222383499146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,10240,0.10145066844092475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,6144,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,7168,0.03914133376545376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,65536,0.24018398920694986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,8192,1.8162665896945531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,512,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,2048,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,256,0.12111022737291123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,1536,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,10240,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,128,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,3584,0.07775821950700548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,16384,0.015050666199790107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,4096,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,3072,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,8192,0.10072266393237644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,16384,0.016534222496880423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,3584,0.012970666918489667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,10240,0.32048622767130536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,8192,0.037408000893063016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,512,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,256,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,3072,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,1024,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,3584,0.02824977702564663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,16384,0.031694223483403526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,768,0.05711555480957031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,6144,0.011769777370823754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,1024,0.023286221755875483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,256,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,12288,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,7168,0.08199822240405612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,12288,0.21334132883283827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,2048,0.30452177259657115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,2560,0.02204888893498315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,65536,0.33321332931518555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,7168,0.024706666668256123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,8192,0.022207111120224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,8192,0.029911110798517864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,3072,0.008931555681758458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,3584,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,128,0.028581334484948054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,7168,0.014673777752452426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,3072,0.01700888905260298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,768,0.020114666885799833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,8192,0.05495822098520067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,2048,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,2048,0.012490666574902005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,16384,0.023698666029506262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,768,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,5120,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,1024,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,4096,0.04257955484920078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,8192,0.013822221921549903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,5120,0.019887111253208585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,4096,0.020394666327370536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,65536,0.06356710857815213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,4096,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,65536,0.07587733533647326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,64,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,3072,0.09139021899965073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,5120,0.1473875575595432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,768,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,128,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,12288,2.135332531399197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,1024,0.03607644306288825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,10240,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,128,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,5120,0.013120888835854001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,3584,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,51200,0.06447111235724555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,3584,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,16384,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,5120,0.0612551106346978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,3584,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,3584,0.03872266742918227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,51200,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,3584,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,2560,0.008599999878141616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,12288,0.07294044229719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,1536,0.012545777691735161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,512,0.0037813331517908308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,128,0.09174044264687432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,768,0.012857777376969656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,128,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,64,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,8192,0.024513777759340074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,256,0.015446222490734525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,51200,0.08610666460461086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,2560,0.031203554736243352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,16384,0.03222489025857713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,3584,0.013949332965744866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,1536,0.009846222069528367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,1024,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,1536,0.026127111580636766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,5120,0.023704888092146978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,10240,0.3331742286682129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,256,0.047713776429494224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,128,0.01330311099688212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,64,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,4096,0.011396444506115384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,65536,0.48735464943779844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,3072,0.009481777747472128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,2560,0.007043555378913879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,6144,0.19346400101979574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,3072,0.010280888941552904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,1024,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,5120,0.009378666679064432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,5120,0.039255999856524997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,65536,0.12367110782199436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,8192,0.1554737753338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,64,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,1536,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,16384,0.015785778562227886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,768,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,65536,0.21255466673109266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,3072,0.07328710953394572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,6144,0.015691555208630033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,7168,0.06574489010704888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,6144,0.24962933858235678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,256,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,2560,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,32,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,128,0.009962666365835402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,128,0.08255555232365926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,1536,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,16384,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,10240,0.10908355315526326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,2560,0.03264355659484863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,3072,0.026895999908447266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,2560,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,128,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,10240,0.05162222186724345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,16384,0.021212443709373474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,7168,0.025732444392310247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,6144,0.015955555770132277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,128,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,2560,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,10240,0.027211555176311072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,6144,0.010413332945770687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,3072,0.02777333392037286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,2560,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,4096,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,512,0.005719999886221356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,7168,0.08260889185799493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,2560,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,51200,0.07667111025916205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,65536,0.4061440096961127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,7168,0.02478755513827006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,1536,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,3584,0.0052915554907586836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,512,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,768,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,256,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,7168,0.023787554767396715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,1536,0.03683111071586609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,1536,0.013542222479979197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,1024,0.013951111170980664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,256,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,65536,0.24878221087985566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,6144,2.078555636935764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,64,0.009191999832789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,6144,0.042295998997158475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,65536,1.6667839686075847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,12288,0.10473866595162286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,768,0.02475111186504364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,8192,0.23695643742879233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,51200,0.7718044386969672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,2048,0.03582133187188043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,65536,0.07191555367575751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,256,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,768,0.31114400757683647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,12288,0.02088711162408193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,8192,0.0335822237862481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,3072,0.057068447271982826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,1536,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,4096,0.009116444322797988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,128,0.0038684445122877755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,512,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,256,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,12288,0.03241511185963949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,6144,0.03120888935195075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,16384,1.5616221957736547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,256,0.022742221752802532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,2048,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,5120,0.012878222597969903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,7168,0.09137244357003106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,512,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,3072,0.030144890149434406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,1024,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,1536,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,7168,0.18716533978780112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,768,0.026741334133678015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,32,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,3584,0.010495111346244812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,512,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,51200,0.12794221772087946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,10240,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,32,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,3584,0.5130222108629015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,51200,0.5210515658060709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,512,0.03175910976197984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,4096,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,65536,0.06583822435802884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,51200,0.9188755883110894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,2560,0.011353777514563667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,512,0.007144889069928064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,2560,0.02164177762137519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,2048,0.2950346734788683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,3072,0.008355555435021719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,10240,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,512,0.015073777900801765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,512,0.02927466564708286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,768,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,2560,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,65536,0.25595911343892414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,8192,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,2048,0.035028444396124944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,16384,0.017157332764731515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,4096,0.018361777067184448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,3072,0.029214223225911457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,65536,0.18096800645192465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,256,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,3584,0.036601778533723615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,8192,0.391887108484904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,3584,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,2048,0.026174222429593403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,65536,3.235367033216688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,16384,0.03506933318244086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,12288,0.026347556047969397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,4096,0.09583910968568589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,16384,0.047410666942596436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,2048,0.029437333345413208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,2560,0.023031999667485554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,7168,0.1182755496766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,64,0.005079111291302575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,51200,1.5133315192328558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,2048,0.030291554000642564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,16384,0.038374221987194486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,4096,0.04468977782461378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,768,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,4096,0.01622222198380364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,7168,0.027902222341961328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,5120,0.0650213360786438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,7168,0.07670488622453478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,6144,0.051242666112052075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,2560,0.10924444595972697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,512,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,1024,0.09016622437371148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,8192,0.015388444066047668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,128,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,10240,4.006223890516493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,51200,0.5989564259847006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,10240,0.18131644195980498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,128,0.007697777615653143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,3072,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,7168,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,5120,0.028552889823913574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,7168,0.021593777669800654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,1024,0.013600000076823764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,7168,0.03218844532966614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,2560,0.42835201157463926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,1024,0.05739911397298177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,5120,0.3662159972720676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,8192,0.025802666942278545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,512,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,512,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,4096,0.03973955578274197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,3072,0.026525333523750305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,65536,2.66556888156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,7168,0.0481697784529792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,3072,0.022255111071798537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,512,0.02644177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,1536,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,8192,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,3584,0.026798221800062392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,4096,0.016511110795868766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,6144,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,2048,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,256,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,32,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,3072,0.02130844526820713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,6144,0.03253333436118232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,64,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,128,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,51200,0.12242310576968723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,1536,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,2048,0.021998221675554912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,65536,0.05919910801781548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,10240,0.04511733187569512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,256,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,2560,0.0518124434683058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,8192,0.21343910694122314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,32,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,512,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,12288,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,2048,0.014775110615624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,512,0.0023111111174027124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,6144,0.06823466883765326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,3584,0.00684533350997501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,2560,0.02574577762020959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,1536,0.08449600140253703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,8192,1.499383078681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,5120,0.08943111366695827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,10240,0.05577066871854994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,1024,0.014523555835088095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,3072,0.015966221690177917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,1024,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,5120,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,3072,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,1024,0.012440889245933957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,7168,0.0366213321685791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,1536,0.024881778491867915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,65536,0.06176800198025174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,7168,0.024529778295093115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,2048,0.01145600030819575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,1024,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,768,0.016626665989557903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,256,0.02537422213289473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,51200,0.2656533453199598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,8192,0.015544889701737298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,16384,0.14240177472432455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,16384,0.2546044455634223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,6144,1.2525680330064561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,12288,0.040208889378441706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,65536,4.5354737175835504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,128,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,128,0.08840888738632202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,768,0.025383111503389146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,4096,0.008789333204428354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,256,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,1536,0.007465778125656976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,8192,0.01865688959757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,3584,0.05633866786956787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,768,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,6144,0.015541333291265698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,64,0.00544177782204416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,768,0.01737244427204132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,4096,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,2048,0.1659439934624566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,8192,0.024187554915746052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,5120,0.012033777932325998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,51200,0.6910542382134331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,8192,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,16384,0.08309866984685262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,16384,0.054472890165117055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,256,0.02091111077202691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,2560,0.008186666501892937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,10240,0.06023999717500475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,2560,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,8192,0.04386666748258802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,4096,0.016935111747847665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,768,0.17884443865882027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,256,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,8192,0.10204088687896729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,2048,0.05526133378346761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,12288,0.018350222044520907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,7168,0.026174222429593403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,10240,0.013873777455753751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,51200,9.489445156521267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,512,0.024692444337738886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,6144,0.24426311916775176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,64,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,768,0.0070879997478591064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,3072,0.0589617755677965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,128,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,12288,0.03934489025010003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,16384,4.43522220187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,16384,0.029343999094433252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,12288,0.02197244432237413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,7168,0.6132888793945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,10240,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,51200,0.017135999268955655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,2560,0.026005332668622334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,3584,0.012606222596433429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,256,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,256,0.007096889118353526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,7168,0.046576887369155884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,5120,0.008363555702898238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,2048,0.013843555417325763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,2048,0.0746506651242574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,3072,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,7168,2.0830764770507812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,12288,0.030004445049497817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,128,0.023680888944202002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,12288,0.2588062286376953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,12288,0.12160800562964545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,5120,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,1536,0.0032231110251612137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,4096,0.051683555046717324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,512,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,2048,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,3584,0.01887822151184082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,2560,0.03388266762097677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,2560,0.37357332971360946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,1536,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,65536,0.10557511117723252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,128,0.05910044246249729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,2048,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,12288,0.05453333258628845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,12288,0.07254488600624932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,12288,0.03287555442916022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,1024,0.020568889048364427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,12288,0.06254666381412081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,128,0.0020702222569121257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,128,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,8192,0.038274665673573814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,51200,0.24517777231004503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,7168,0.11724978023105198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,10240,0.018175111876593698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,8192,0.04618488748868307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,7168,0.15256444613138834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,768,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,6144,0.011062222222487131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,10240,0.17686222659216988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,64,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,6144,0.7396346728006998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,6144,0.05409511261516147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,4096,0.08419377936257257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,16384,0.09278844462500678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,7168,0.04655999938646952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,5120,0.05642933315700955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,3072,0.025117332736651104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,64,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,51200,0.1944755580690172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,4096,0.009053332938088311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,10240,0.03613600134849548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,3072,0.34787734349568683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,128,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,2048,0.038690666357676186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,512,0.016895999511082966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,1536,0.011384888655609555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,3072,0.024501333634058636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,2560,0.012292444705963135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,768,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,16384,0.04985510971811083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,51200,0.12038311693403457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,51200,0.14537689420912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,16384,0.06001422140333387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,32,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,3584,0.0514613323741489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,16384,0.6158293088277181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,16384,0.05272355675697327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,5120,0.09848799970414902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,4096,0.05645155244403415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,3072,0.012960000170601739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,3584,0.013425778183672162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,6144,0.019620445039537217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,1024,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,32,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,10240,0.388791110780504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,65536,0.07566844754748873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,6144,0.01668444441424476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,3072,0.032136887311935425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,3584,0.03156355685657925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,4096,0.0083155557513237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,5120,0.03467288944456313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,12288,1.4326595730251734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,1024,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,2048,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,16384,0.5130595631069607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,10240,1.9815288119845922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,768,0.01070755554570092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,4096,0.013542222479979197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,1536,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,4096,0.033991999096340604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,1024,0.3750604523552789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,10240,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,2048,0.028965334097544353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,12288,0.07258933120303683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,512,0.011416888899273343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,3072,0.019500444332758587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,51200,1.2814782460530598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,1024,0.0328764451874627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,2560,0.009951999617947472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,256,0.009379555781682333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,4096,0.021440888444582622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,2560,0.007072000039948358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,4096,0.017247110605239868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,128,0.005391110976537068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,1536,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,5120,0.03943555553754171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,3584,0.014577777849303352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,16384,0.04339111182424757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,256,0.002642666714058982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,2560,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,7168,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,10240,0.8358364635043674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,10240,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,51200,0.09207110934787327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,3072,0.011374221907721626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,1536,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,32,0.003084444544381566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,65536,0.03868533174196879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,2560,0.004921777794758479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,5120,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,5120,0.014711999230914645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,256,0.016534222496880423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,256,0.013984888792037964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,6144,0.050550222396850586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,2048,0.014697778556081982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,4096,0.5794684622022841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,4096,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,12288,0.12334844801161025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,3584,0.036694222026401095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,256,0.0025537777692079544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,2560,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,8192,0.04147466686036851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,12288,0.02271733350223965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,6144,0.017416889468828838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,8192,0.037470221519470215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,64,0.08305155568652682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,3584,0.05305333269966973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,256,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,51200,0.07893155680762397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,1024,0.0034346667428811393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,512,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,8192,0.07348889112472534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,65536,0.9469253751966688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,1024,0.009335111412737105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,16384,0.019691555036438834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,2560,0.044960889551374644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,128,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,4096,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,768,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,3072,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,3072,0.016707556115256417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,7168,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,3072,0.032764444748560585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,12288,0.5828720198737251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,2048,0.02827289038234287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,3072,0.03788177834616767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,6144,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,8192,0.02066044509410858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,32,0.0019288888821999233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,512,0.17713867293463814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,6144,0.04294755392604404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,256,0.022279111875428095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,65536,0.19961955812242296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,256,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,10240,0.14977422025468615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,65536,0.06007822354634603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,5120,0.03128711051411099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,16384,0.022900443938043382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,3072,0.017833777599864535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,256,0.0021342221233579847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,2560,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,1024,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,2048,0.03727733426623874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,65536,0.0701706674363878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,6144,0.013811555173661975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,256,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,2560,0.03487911158137851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,4096,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,256,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,512,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,16384,0.04097155398792691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,51200,0.8434711032443576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,4096,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,12288,0.01937599976857503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,64,0.014191110928853353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,8192,0.021299555897712708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,51200,0.09722399711608887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,32,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,7168,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,51200,0.04866933491494921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,3584,3.6823857625325522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,7168,0.03962044583426581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,8192,3.2027431064181857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,3072,0.006335111127959357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,256,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,65536,0.09787022405200535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,8192,0.011219555305110084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,64,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,256,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,12288,0.24833599726359049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,12288,0.029586666160159644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,51200,1.2870266172620985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,65536,0.7981102201673719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,3072,0.026983110441101923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,128,0.022008889251285132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,512,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,128,0.00665155549844106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,128,0.01741333305835724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,5120,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,1536,0.06425422430038452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,256,0.023764444722069636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,51200,0.3835386700100369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,3584,0.016653334101041157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,12288,1.5772711435953777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,8192,0.04051377707057529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,2048,0.04773333337571886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,5120,1.0495582156711154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,512,0.00943999985853831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,8192,0.4689111179775662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,64,0.024647111694018047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,12288,0.026705778307384912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,4096,0.04606044292449951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,1536,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,1024,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,3584,0.09431644280751546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,12288,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,6144,0.0203128887547387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,51200,1.8261608547634547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,3584,0.09198577536476983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,65536,0.1747839980655246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,3072,0.014860444598727755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,16384,0.05565244621700711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,512,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,12288,0.031217777066760596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,12288,0.025358221597141687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,16384,0.06901599963506062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,7168,0.010096000300513374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,12288,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,16384,0.034435556994544134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,512,0.016509332590632968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,8192,0.012223111258612739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,51200,3.4759396447075734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,8192,0.07671378056208293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,4096,0.013440888788965015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,3072,0.039121776819229126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,5120,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,32,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,3584,0.05368444323539734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,3072,0.02892000145382351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,3584,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,64,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,256,0.0053084443012873335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,1024,0.03427555494838291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,256,0.003485333174467087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,65536,0.48120445675320095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,64,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,256,0.0074720001882976955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,12288,0.033831109603246055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,51200,0.19952532980177137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,1024,0.06732800271775988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,512,0.32441422674391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,5120,0.026212443908055622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,3072,0.012198222180207571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,4096,0.07089777787526448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,10240,0.1881528827879164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,4096,0.23253332244025338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,8192,0.03201422095298767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,51200,0.9106889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,12288,0.043122665749655835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,128,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,2560,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,8192,0.5158008999294704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,12288,0.15764889452192518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,6144,0.2445333268907335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,4096,0.015328889091809591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,8192,0.43961066669887966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,12288,0.04220088985231188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,1536,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,2048,0.11104266511069404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,2560,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,3584,0.05439910954899258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,12288,0.03372888763745626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,8192,0.0866862204339769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,6144,0.05603733327653673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,10240,0.06230577495363024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,16384,0.28925244013468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,12288,0.033053331904941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,256,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,65536,0.08964977661768596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,5120,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,6144,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,7168,0.06894577874077691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,16384,0.2614195611741808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,5120,0.015547555353906421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,768,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,6144,0.5994773440890843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,64,0.18653066953023276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,65536,0.22253065639071992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,2560,0.0694248874982198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,12288,0.06057066387600369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,32,0.1559600035349528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,6144,0.014840000205569796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,64,0.010409778190983666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,768,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,1536,0.014147554834683737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,32,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,10240,0.039120889372295804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,128,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,4096,0.03832533293300205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,12288,0.04692622356944614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,7168,1.05968443552653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,1536,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,2048,0.032307555278142296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,7168,0.12169867091708714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,12288,0.0627431141005622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,2560,0.03303822212749057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,256,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,2048,0.23159554269578722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,128,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,4096,0.09274044301774766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,3584,0.3046169016096327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,6144,0.057645334137810596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,6144,0.03086400032043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,512,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,256,0.02569155560599433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,51200,0.08806844552357991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,4096,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,5120,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,2048,0.02125688890616099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,2048,0.05995733208126492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,8192,0.14463999536302355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,1024,0.027487110760476854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,3584,0.008264000217119852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,12288,0.15801778104570177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,10240,0.24571466445922852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,1024,0.09716177648968166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,16384,0.01610844499535031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,768,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,7168,0.012071111136012606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,512,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,7168,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,6144,0.25448976622687447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,32,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,16384,0.03052799900372823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,256,0.014087110757827759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,512,0.012874667015340594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,4096,0.01835555500454373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,65536,1.2990923987494576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,7168,0.11961421701643203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,10240,0.31773421499464244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,3584,0.005971555494599872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,64,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,2560,0.6117822329203287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,64,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,10240,4.5758310953776045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,5120,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,32,0.06542666753133138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,32,0.010284444524182213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,128,0.024413333998786077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,32,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,4096,0.034133334954579674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,768,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,256,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,2560,0.007357333269384172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,768,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,256,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,1024,0.008480889101823172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,10240,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,256,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,51200,0.6202693515353733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,2560,0.017234666479958426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,128,0.01906399925549825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,32,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,4096,0.1609599987665812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,12288,0.012457778056462606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,4096,0.0090515555606948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,2560,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,5120,0.0453324450386895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,6144,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,51200,0.6558426751030816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,4096,0.07605599694781832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,6144,0.030140442980660334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,1536,0.02743911080890232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,512,0.03353510962592231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,1024,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,4096,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,768,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,8192,0.8663191265530057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,12288,0.17805333932240805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,3584,0.007755555212497711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,8192,0.04869955446985033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,10240,0.205749339527554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,3072,0.027841776609420776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,6144,0.07024977604548137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,8192,0.07310755385292901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,1024,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,65536,0.7323724428812662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,256,0.030786666605207656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,51200,0.06211466921700371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,65536,0.16580177678002253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,256,0.0042071110672420925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,2560,0.31085067325168186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,256,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,3072,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,7168,0.11888799402448867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,768,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,2560,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,16384,0.2608275678422716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,8192,0.02755733331044515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,3072,0.04013955593109131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,32,0.02887822190920512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,1024,0.012113778127564324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,65536,0.07891288730833265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,6144,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,6144,0.05714133050706652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,16384,0.07381066348817614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,5120,0.08966310818990071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,8192,0.2957822216881646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,768,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,65536,0.3467697832319472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,2560,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,12288,0.018428444862365723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,1024,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,5120,0.02441066669093238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,512,0.013532444834709167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,4096,0.008092444803979661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,10240,0.022298667165968154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,2048,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,3584,0.020603555772039626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,2048,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,512,0.016883555385801528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,5120,0.0177182224061754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,32,0.015691555208630033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,51200,0.5517066849602593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,32,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,10240,0.044488890303505786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,8192,0.021515554851955835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,1536,0.008767999708652496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,65536,0.678984906938341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,3584,0.009588444398509132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,10240,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,10240,0.10286400053236221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,7168,0.04033599959479438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,12288,0.008642666869693333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,51200,0.06945244471232097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,128,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,5120,0.12177777290344238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,65536,0.06622044245402019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,65536,0.10868799686431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,10240,0.08774844143125747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,6144,0.02014044423898061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,5120,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,16384,0.24392443233066133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,10240,0.08639555507236057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,3072,0.03090311090151469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,5120,0.24960798687405059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,768,0.01824977828396691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,10240,0.010676444404655032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,65536,0.2219057745403714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,6144,0.0576488905482822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,128,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,3584,0.016544888416926067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,256,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,256,0.010268444816271463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,128,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,5120,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,51200,0.1649155616760254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,1024,0.030581331915325586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,3072,0.038133333126703896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,32,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,512,0.15468977557288274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,10240,0.03978666663169861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,10240,1.0358390808105469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,64,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,256,0.020608888732062448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,1536,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,4096,0.14308088355594212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,256,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,1536,0.020420443680551317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,10240,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,64,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,256,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,1536,0.043841779232025146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,3072,0.0159288893143336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,10240,0.014024888475735983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,6144,0.08350221978293525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,7168,0.02937155630853441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,4096,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,1024,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,6144,0.019367999500698514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,3072,0.3083537684546577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,64,0.006815111057625876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,4096,0.029791110091739233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,16384,0.6188631057739258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,4096,0.046486222081714206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,64,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,1024,0.006124444305896759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,5120,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,512,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,5120,0.035014222065607704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,10240,0.03234400020705329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,65536,0.11590311262342666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,2048,0.058873779243893094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,8192,0.09316177500618829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,5120,0.09477155738406712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,1024,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,5120,0.022796443767017786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,51200,0.06655289067162408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,1536,0.01201866649919086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,6144,0.11405866675906712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,7168,0.20174222522311738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,16384,0.16176088651021323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,32,0.02370133333735996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,256,0.00254488881263468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,1024,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,512,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,3584,0.03614044520590041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,7168,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,128,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,4096,0.0503111117415958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,1536,0.003576888806290097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,6144,0.04074133435885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,8192,0.1446266704135471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,12288,0.01919644408755832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,65536,0.07322400146060519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,2048,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,512,0.016371554798550077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,128,0.020637333393096924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,4096,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,3584,0.027496889233589172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,1536,0.025819554924964905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,1024,0.006305777778228124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,768,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,3584,0.022962666220135156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,64,0.006108444597986009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,768,0.01755199995305803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,3072,0.027467555469936792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,5120,0.021180444293551978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,12288,0.01724799970785777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,6144,0.026817777090602454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,51200,0.10971111059188843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,64,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,6144,0.12403021918402778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,2560,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,1536,0.019375110665957134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,51200,0.4579342206319173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,10240,0.03167377909024557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,65536,0.6371413336859809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,512,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,12288,0.021010667085647583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,16384,0.5142640007866753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,64,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,2560,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,8192,0.07691999938752916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,10240,0.04599555664592319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,4096,0.015577778220176697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,768,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,8192,0.03202666507826911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,1024,0.012552888856993781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,10240,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,768,0.1484453280766805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,64,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,4096,1.3394124772813585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,7168,0.39403287569681805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,2048,0.0278479986720615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,4096,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,1536,0.015618667006492615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,65536,0.10535999801423813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,1024,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,4096,0.03300711181428697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,768,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,512,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,2560,0.010763555765151978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,12288,0.12796089384290907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,3584,0.18892621994018555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,1024,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,2048,0.04017688830693563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,1024,0.016908443636364408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,256,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,1024,0.025075554847717285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,8192,1.6482924355400934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,128,0.020580444071027968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,3072,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,7168,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,6144,0.029046220911873713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,65536,2.107020484076606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,512,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,2560,0.6879315376281738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,10240,0.0281333327293396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,4096,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,10240,0.12095110946231419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,5120,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,768,0.0362044440375434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,768,0.012792000340090858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,3584,0.014920888675583733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,4096,0.06564178069432576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,512,0.01724799970785777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,65536,0.21377600563897026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,768,0.002965333354141977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,1536,0.039432889885372586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,51200,1.507149378458659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,16384,0.2124133374955919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,16384,0.03832533293300205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,64,0.008851555486520132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,128,0.022781333989567224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,6144,0.04753422074847751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,4096,0.026033777329656813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,1024,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,128,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,1536,0.16169955995347765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,512,0.013216888739003075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,2048,0.03663555449909634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,1024,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,1536,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,2560,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,2560,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,768,0.041287110911475286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,65536,15.523829142252604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,128,0.0038542221817705366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,10240,0.6374293433295356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,8192,0.07840444644292195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,51200,0.2163511117299398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,4096,0.11671555042266846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,10240,2.7306435902913413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,2560,0.017503110898865592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,51200,0.05415022373199463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,3072,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,4096,0.036928001377317644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,1536,0.015265777707099915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,1024,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,128,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,16384,0.07675911320580377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,3072,0.6481155819363064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,16384,0.049792889091703624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,128,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,4096,0.04569777846336365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,2048,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,512,0.021904889080259535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,256,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,12288,0.12718844413757324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,768,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,3072,0.015405333704418607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,256,0.01202666676706738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,65536,0.12005421850416396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,4096,0.015280000037617154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,32,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,768,0.011404444773991903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,128,0.019543111324310303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,3072,0.054882665475209556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,6144,0.07168800301022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,512,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,5120,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,2560,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,1024,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,8192,0.17504355642530653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,8192,0.029643555482228596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,6144,0.02537422213289473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,256,0.017500445246696472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,32,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,3072,0.030518223841985066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,2048,0.027735999888843958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,8192,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,8192,0.03715377714898851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,16384,0.033891555335786604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,128,0.016075554821226332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,16384,0.23492354816860625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,2560,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,12288,0.11005244652430217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,3072,0.009146666361225976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,64,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,3584,0.004553777890072929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,128,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,1536,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,768,0.028196444114049275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,51200,0.1735911104414198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,12288,0.014056889547242058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,51200,0.04641333222389221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,7168,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,256,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,10240,0.040940443674723305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,1024,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,5120,0.03155111273129781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,10240,0.05089422067006429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,3584,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,7168,0.07178844345940484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,3584,0.005305777821275923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,5120,0.08532800277074178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,1536,0.010204444328943888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,8192,0.10141333606508042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,2560,0.015556444724400839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,1536,0.10733155409495036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,12288,0.20047288470798066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,8192,0.09777955876456366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,16384,0.2747724321153429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,5120,0.03804444604449802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,2560,0.044440888696246676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,1024,0.006162666612201267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,5120,0.034661332766215004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,4096,0.013336000343163809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,6144,0.05448089043299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,7168,0.015092444088723926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,3584,0.0047395556337303585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,65536,0.0770124461915758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,8192,0.09681510925292969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,1536,0.9780267079671224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,512,0.005643555687533484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,1024,0.038351999388800725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,256,0.021409778131379023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,3584,0.12098577287462021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,128,0.01144266708029641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,2048,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,1024,0.031798223654429116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,768,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,1536,0.18475377559661865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,12288,0.15943200058407253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,10240,0.04002666804525588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,6144,0.05800977680418226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,64,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,7168,0.01901511185699039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,1536,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,6144,0.03222666515244378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,12288,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,2048,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,32,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,1536,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,1024,0.004488888714048597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,65536,0.0890986654493544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,16384,0.13478133413526747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,128,0.03378488951259189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,1024,0.0032684444967243406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,16384,0.04364000095261467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,1536,0.1308008829752604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,3584,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,7168,0.012839111189047495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,5120,0.007479111353556315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,128,0.014074666632546319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,32,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,4096,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,256,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,2560,0.011367999845080905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,3072,0.05724177757898966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,2048,0.050866666767332286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,3072,0.032969776127073504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,3584,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,4096,0.059742218918270536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,512,0.002537777854336633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,256,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,128,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,512,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,7168,0.010514666636784872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,10240,0.023405333360036213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,64,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,1024,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,7168,0.16575645075903997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,32,0.013046222428480784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,12288,0.040752000278896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,1024,0.0034542222403817703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,3584,0.10540355576409234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,7168,0.021075555019908484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,65536,0.48367823494805234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,5120,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,51200,0.06738489203982882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,256,0.09779466523064508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,512,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,5120,0.013205333716339536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,6144,0.019840889506869845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,4096,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,6144,0.010048000348938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,5120,0.021385777327749465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,5120,0.06344978014628093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,5120,0.05670666694641113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,51200,2.9928578270806203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,51200,0.055050667789247304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,1536,0.02188088827662998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,1024,0.01328000012371275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,51200,0.9268515904744467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,2048,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,1536,0.0042462220622433555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,2560,0.29493244489034015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,1536,0.08717422352896796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,16384,0.027610666222042505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,2048,0.0719475547472636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,16384,0.01994311147265964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,5120,0.01688888834582435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,5120,0.04982044299443563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,65536,0.08923377593358357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,256,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,3072,0.03782222337192959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,3072,0.026313778426912095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,3584,0.01961955593691932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,512,0.018536888890796237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,65536,0.4479982058207194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,512,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,5120,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,51200,0.05812178055445353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,12288,0.2088008854124281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,6144,0.10740800036324395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,12288,0.0643084446589152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,4096,0.05298044284184774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,6144,0.03696711195839776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,8192,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,51200,0.3319671154022217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,16384,0.417846229341295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,768,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,10240,0.940424919128418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,65536,0.1453733311759101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,2560,0.011443555355072021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,256,0.011000889043013254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,512,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,128,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,51200,0.13166577286190456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,8192,0.020250666472646926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,5120,0.016388444436921012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,3072,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,12288,0.11651110649108887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,64,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,51200,0.10288888878292507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,32,0.006126222097211414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,3584,0.015611555841233997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,5120,0.01847022275129954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,256,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,5120,0.02273066673013899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,3072,0.03858310977617899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,2048,0.006457777900828256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,256,0.009706666900051964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,1536,0.07651466793484159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,256,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,6144,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,3072,0.034966223769717746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,1024,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,3072,0.013909333282046847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,512,0.002700444517864121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,3584,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,10240,0.028725332683987085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,65536,0.09706666734483506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,512,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,2048,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,512,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,2048,0.029233776860766943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,1536,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,3584,0.09077333079444037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,8192,0.09110133515463935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,2560,0.1736968888176812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,6144,0.020982222424613103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,6144,0.009453333086437648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,256,0.04387199878692627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,3072,0.007678222325113084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,2048,0.024039111203617517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,7168,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,4096,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,256,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,32,0.0019475554840432273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,5120,0.02186222208870782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,1536,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,6144,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,8192,0.3383715682559543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,2048,0.004883555488453971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,16384,0.03350844317012363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,6144,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,10240,0.0588231086730957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,10240,0.04919999837875366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,7168,0.010319111247857412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,10240,0.17717333634694418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,8192,0.031027555465698242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,12288,0.033490667740503945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,256,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,32,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,128,0.01860000027550591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,12288,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,16384,0.37294578552246094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,3072,0.08358577887217204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,65536,0.16335822476281062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,128,0.002198222196764416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,512,0.002986666642957263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,10240,0.005714666512277391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,7168,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,51200,0.04220622115665012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,3584,0.03192977772818671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,256,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,5120,0.00942755573325687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,16384,0.02646222213904063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,8192,0.05137688914934794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,3584,0.026518222358491685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,128,0.032440887557135686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,8192,0.4471591313680013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,256,0.030778666337331135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,6144,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,4096,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,12288,0.04490577843454149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,7168,0.010418666733635796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,128,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,5120,0.04387288954522875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,768,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,8192,0.04840266704559326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,3584,0.041871110598246254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,2560,0.01276000009642707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,4096,0.03417511118782891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,256,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,6144,0.024358222881952923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,65536,0.10336089134216309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,3584,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,128,0.01809244520134396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,10240,0.01791733337773217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,3072,0.01500266624821557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,16384,0.15089155568016901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,128,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,10240,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,2560,0.007350222104125553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,2560,0.020591111646758187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,4096,0.014912888407707214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,3072,0.0550177792708079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,128,0.025535111625989277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,2048,0.02108000053299798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,4096,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,512,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,3584,0.01587733295228746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,2048,0.015037332971890768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,2048,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,3584,0.01513155632548862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,256,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,2560,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,16384,0.028387556473414104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,32,0.15966667069329157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,6144,0.11251110500759548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,16384,0.04124266571468777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,768,0.011981333295504252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,7168,0.03738933470514085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,3072,0.059608002503712974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,2048,0.07551199860042997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,51200,0.3317182328965929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,768,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,6144,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,6144,0.06572533316082425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,256,0.014256889621416727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,8192,0.023613333702087402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,3584,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,16384,0.076200889216529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,12288,0.018114666144053142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,3072,0.007072000039948358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,512,0.01754044493039449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,2560,0.0458568897512224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,65536,0.06123822265201145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,768,0.03572888837920295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,1536,0.2153395546807183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,7168,0.023297778434223596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,5120,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,3072,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,64,0.020227554771635268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,16384,0.04310311211480034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,16384,0.0373599992858039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,3072,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,768,0.018787556224399142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,7168,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,16384,0.1712275611029731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,768,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,1536,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,5120,0.012512889173295764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,5120,0.02651111119323307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,16384,0.03980799847178989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,8192,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,8192,0.01075111081202825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,2560,0.08180177874035306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,512,0.009131555755933126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,768,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,512,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,3584,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,2048,0.015105777316623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,256,0.02886311213175456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,1024,0.011815999945004782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,1024,0.012048000262843238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,16384,0.04195288817087809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,3584,0.220688886112637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,32,0.1259333292643229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,4096,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,51200,0.6617573102315267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,4096,0.007692444655630324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,51200,0.17340355449252656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,12288,7.356602138943142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,2560,0.05762488974465264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,12288,0.3612017896440294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,3072,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,51200,0.05013066530227661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,5120,0.011807111402352652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,1024,0.06483022371927898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,4096,0.018759111563364666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,1024,0.0388008885913425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,32,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,128,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,16384,0.229784886042277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,6144,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,12288,0.016433777080641854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,51200,0.017888888716697693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,8192,0.014874666929244995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,12288,0.24888179037306044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,1536,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,64,0.0018275556051068837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,51200,7.6218151516384545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,128,0.005646222167544895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,10240,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,1536,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,2048,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,8192,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,32,0.00573333352804184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,768,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,1536,0.018221333622932434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,5120,0.04676977793375651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,2560,0.004945777770545748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,65536,0.07274222373962402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,3072,0.021130666136741638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,768,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,12288,0.02959111001756456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,16384,5.6068471272786455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,65536,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,12288,0.01787911189926995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,5120,0.010874666273593903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,7168,0.01202488856183158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,4096,0.03886399997605218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,2560,0.00489777781897121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,5120,0.01830844415558709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,6144,0.0801724460389879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,768,0.020299555526839364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,4096,0.02481511069668664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,5120,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,768,0.022487999664412603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,1024,0.014488000008794995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,2560,0.011128000087208219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,64,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,7168,0.05019911130269369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,16384,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,4096,0.009539555344316695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,2560,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,64,0.05757155683305529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,5120,0.025278222229745653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,1024,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,8192,2.098714616563585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,4096,0.21293511655595568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,3072,0.0673182209332784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,6144,0.05190844337145487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,8192,0.023876445160971746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,3072,0.056074665652381055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,1024,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,1536,0.016150222884284127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,8192,0.010515555739402771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,1024,0.05242400036917793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,2560,0.05802222092946371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,16384,0.03935555617014567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,1536,0.021672889590263367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,2048,0.025702221526039973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,512,0.010493333141009012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,1536,0.18066933419969347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,2560,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,256,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,1024,0.0038124442928367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,128,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,8192,0.03593066665861342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,3072,0.00849422232972251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,1024,0.061736000908745654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,768,0.019596444235907663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,128,0.014067555467287699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,2560,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,2048,0.031465778748194374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,3584,0.005304000029961268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,51200,0.11233688725365533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,6144,0.06928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,16384,0.049145778020222984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,4096,0.42338400416904026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,51200,0.02708622150950962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,6144,0.019719999697473314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,128,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,256,0.002236444503068924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,51200,0.050986667474110924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,6144,0.487960868411594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,10240,0.10669955280092026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,3584,0.0352497763103909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,2048,0.007098666495747036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,12288,0.07623910903930664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,512,0.008374222450786168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,1024,0.024446222517225478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,51200,0.11975822183820937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,16384,0.05579377545250786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,3584,0.0583911140759786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,5120,0.15744800037807888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,5120,0.007328000333574083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,6144,0.023733332753181458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,10240,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,3072,0.025969778498013813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,2048,0.01313066648112403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,128,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,1024,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,6144,0.023696889479955036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,6144,0.009983110758993361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,65536,0.13039467069837782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,4096,0.011342222491900126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,3072,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,128,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,768,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,1024,0.004510222209824456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,2048,0.008471999731328752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,6144,0.03714933329158359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,512,0.018716444571812946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,6144,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,128,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,7168,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,768,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,2048,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,10240,0.06526755624347262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,2560,0.013648000028398303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,128,0.008231110870838165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,16384,0.08398666646745469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,2048,0.016128000285890367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,768,0.06400799751281738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,2048,0.479489803314209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,51200,0.07941066556506686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,16384,0.09850044382943048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,65536,21.163560655381946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,5120,0.027952889601389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,12288,0.013471110827393003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,2560,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,2560,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,1024,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,256,0.015972443752818637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,2048,0.05304888884226481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,16384,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,10240,0.03778666588995192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,1536,0.024707555770874023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,5120,0.0041786668201287585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,65536,0.07227200269699097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,6144,0.010475555227862464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,8192,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,768,0.030481778913074072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,2560,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,65536,6.642174190945095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,16384,0.020115555988417733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,5120,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,512,0.020220445262061227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,8192,0.026717333330048457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,3584,0.012800888882742988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,1024,0.01962577799956004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,1536,0.006426666759782367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,7168,0.01348888874053955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,2048,0.018602665927675035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,3072,0.07297866874270968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,2048,0.006314666734801398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,768,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,10240,0.03666399916013082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,5120,0.015824000040690105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,768,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,2048,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,6144,0.07016088565190633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,512,0.023053333163261414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,12288,0.014848889576064216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,7168,0.03343200021319919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,12288,0.005984000033802456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,12288,0.013222222526868185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,4096,0.003914666672547658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,512,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,16384,0.01796533332930671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,8192,0.014326221413082547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,65536,0.327202664481269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,8192,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,10240,0.20265511671702066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,8192,0.020875554945733812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,51200,0.05529422230190701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,65536,4.273732503255208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,4096,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,3072,0.03554577628771464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,2560,0.020244444410006206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,768,0.027110222313139174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,512,0.018216888109842937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,64,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,5120,0.015976889265908133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,2048,0.1976355579164293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,16384,0.09053422345055474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,768,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,6144,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,3584,0.02165777815712823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,4096,0.028582221931881372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,16384,0.17313688331180146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,6144,0.012833777401182385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,8192,0.12144977516598171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,1536,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,8192,0.07094133562511869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,4096,0.020611556039916146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,1024,0.01740977830357022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,10240,0.025800888737042744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,3072,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,1536,0.0041555555330382455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,256,0.014882667197121514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,1536,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,3584,0.015663110547595553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,10240,0.146814227104187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,5120,0.010763555765151978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,3072,0.08308000034756131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,2048,0.045217778947618276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,3072,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,7168,0.036460445986853704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,2560,0.01590666671593984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,65536,2.03074582417806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,512,0.012768000364303589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,768,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,10240,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,7168,0.12633689244588217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,1536,0.00703111125363244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,1024,0.05192888776461283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,2048,0.007206222249401941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,512,0.021137777302000258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,64,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,512,0.21155556042989096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,8192,0.10945600271224976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,32,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,1024,0.02207999924818675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,4096,0.04952355557017856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,5120,0.017316444052590262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,65536,0.2507849004533556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,6144,0.0812879999478658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,10240,0.023372444841596816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,7168,0.02991644541422526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,16384,0.01644888851377699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,10240,0.03592355714903937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,512,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,3584,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,64,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,768,0.006735111276308696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,12288,0.04859644505712721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,64,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,3584,0.017310221989949543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,512,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,3584,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,8192,0.09753244453006321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,6144,0.0092257774538464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,3584,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,5120,0.049919111861122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,10240,0.035655998521380954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,10240,0.24199912283155653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,768,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,1024,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,512,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,512,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,51200,1.2531742519802518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,12288,0.036552889479531184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,2048,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,32,0.016639999217457242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,65536,0.34051020940144855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,51200,0.3284577793545193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,7168,0.029537777105967205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,2560,0.012295111185974546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,12288,5.901634640163845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,512,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,65536,0.8421813117133247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,51200,0.21044800016615126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,51200,8.021007961697048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,12288,0.019031110737058852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,10240,0.011061333119869232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,256,0.022030222747060988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,32,0.0020764445265134177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,1536,0.04936000042491489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,2560,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,768,0.09496088822682698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,3584,0.011234666738245221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,3584,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,5120,0.011186666786670685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,5120,0.11865422460767959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,1024,0.010253333383136326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,10240,0.013262222210566202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,32,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,8192,0.13771288924747044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,32,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,65536,0.37679023212856716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,768,0.009249777429633671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,1536,0.006664888726340399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,3072,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,128,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,51200,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,10240,2.3139254252115884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,3584,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,8192,0.02355466617478265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,512,0.02818044357829624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,128,0.03426400158140395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,2048,0.028862221373452082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,1024,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,768,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,2560,0.019194665882322524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,65536,0.0562115576532152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,10240,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,128,0.005514666852023866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,65536,0.09220088852776422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,10240,0.012057777908113269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,128,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,3584,0.020865778128306072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,10240,0.01828977796766493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,2560,0.06538755363888211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,1536,0.008361777497662438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,7168,0.022461333208613928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,12288,0.01349688900841607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,6144,0.050986667474110924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,2560,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,4096,0.08440533611509536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,1536,0.04467822114626566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,16384,0.05605866511662801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,3072,0.004247111164861255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,768,0.005161777966552311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,10240,0.27298132578531903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,2048,0.05250755614704556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,4096,0.02071466710832384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,128,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,3072,0.054831110768847995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,2560,0.10888177818722194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,2560,0.007070221834712558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,51200,0.06633866495556302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,768,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,10240,0.032941334777408175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,768,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,3584,0.01601777805222405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,4096,0.03314755691422357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,8192,0.2563813262515598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,3072,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,1024,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,5120,0.021500445074505273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,128,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,12288,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,1024,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,2048,0.007135999699433644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,6144,0.012096000214417776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,32,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,7168,0.019113777412308585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,256,0.007147555549939473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,128,0.021946666969193354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,65536,1.0116675694783528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,1536,0.012200888660218982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,3584,0.033731556600994535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,12288,0.02959644463327196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,16384,0.023913777536816064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,512,0.1401875548892551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,4096,0.016594666573736403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,16384,0.023400889502631292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,1536,0.41405510902404785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,3072,0.21104266908433703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,256,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,10240,0.3319484392801921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,128,0.013589333328935834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,10240,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,512,0.05275555451711019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,256,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,512,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,12288,0.07218133078681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,6144,0.018910222583346896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,12288,0.036935110886891685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,1536,0.034225778447257146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,3584,0.028032888968785603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,3584,0.012472000386979846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,2560,0.03820266657405429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,12288,0.010265777508417765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,10240,0.01678311162524753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,8192,0.16803733507792154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,1024,0.005346666607591841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,7168,0.1413911051220364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,6144,0.03869955407248603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,2048,0.004384888956944148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,10240,1.2935057746039498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,2048,0.023552000522613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,2560,0.059009777175055615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,3072,0.024199111594094172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,3584,0.3593706554836697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,2048,0.05240799983342489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,3584,0.05006400081846449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,32,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,128,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,10240,0.012432888978057437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,51200,10.370914883083767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,6144,0.017699556218253244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,6144,0.08579288588629828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,768,0.016953777935769822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,5120,0.03839644458558824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,3584,0.19870932896931967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,16384,0.01889600025282966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,12288,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,3584,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,512,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,8192,0.08563911252551609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,8192,0.009879111415810054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,256,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,5120,0.02753422160943349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,16384,0.025106666816605464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,512,0.0025822222232818604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,512,0.025811556312772963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,10240,0.048784001006020435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,6144,0.02286311156219906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,6144,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,6144,0.0273973329199685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,768,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,3584,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,2048,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,2048,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,12288,0.2267528904808892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,3072,0.05528177817662557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,4096,0.03590133455064561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,256,0.009381333159075843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,2560,0.02105955613984002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,4096,0.014515555567211576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,1024,0.042678223715888135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,1024,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,16384,0.05103288756476509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,5120,0.0506915549437205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,8192,0.021556443638271753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,256,0.013256889250543384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,10240,0.1996071073744032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,256,0.015509333875444202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,3072,0.034549332327312894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,768,2.1364258660210504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,2560,0.04165244433614943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,4096,0.02285066743691762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,65536,0.13134045071072048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,65536,0.0744568904240926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,512,0.0047386665311124595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,1024,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,7168,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,512,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,51200,0.10186222526762222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,256,0.01053422192732493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,12288,0.050741334756215416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,64,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,2048,0.017605332864655387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,2560,0.05519555674658882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,10240,0.04775822162628174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,512,0.011925333076053195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,768,0.06452800167931451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,1024,0.01662488944000668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,65536,0.09962399800618489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,768,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,12288,1.8445057339138453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,2560,0.00871022211180793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,3072,0.030604445272021826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,4096,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,8192,0.0224924451775021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,6144,0.0354968872335222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,16384,0.06962222523159452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,7168,0.11506311098734538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,6144,0.016763556334707472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,6144,0.14118844932980007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,51200,0.10698311196433173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,4096,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,256,0.02125599980354309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,5120,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,51200,0.35376887851291233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,32,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,16384,0.06783822509977552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,3072,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,10240,0.032122668292787336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,51200,0.02012622190846337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,256,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,3072,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,4096,0.008028444316652086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,51200,0.1345262262556288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,12288,0.02369155486424764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,6144,0.22278043958875868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,2048,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,7168,0.017504000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,3072,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,7168,0.03329688972897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,12288,0.18145333396063912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,51200,0.6493502193027073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,5120,0.009101333717505137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,3072,0.02306844459639655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,2048,0.006680888848172293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,51200,0.05052533414628771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,128,0.0022133334229389825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,7168,0.19873244232601592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,1536,0.004351110921965705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,1024,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,768,0.2619653277926975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,768,0.0654711127281189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,1024,0.005324444423119227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,16384,0.03909511036343045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,10240,0.043604443470637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,10240,0.11483821603986953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,1024,0.00871377769443724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,32,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,16384,0.02608355548646715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,7168,0.014859555496109856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,256,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,4096,0.306534210840861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,128,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,10240,0.024015110399987962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,1536,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,128,0.061179558436075844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,2048,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,768,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,512,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,2560,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,4096,0.006874666859706243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,51200,0.01832711034350925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,4096,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,8192,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,51200,0.04725333385997348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,12288,0.016113777955373127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,256,0.0064586665895250105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,5120,0.05498844385147095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,3072,0.004933333231343163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,12288,0.036664889918433294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,65536,0.19752266671922472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,2560,0.02591733303334978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,768,0.0038399998512532976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,65536,0.019160888261265226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,65536,0.11756977770063613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,512,0.20325866010453966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,3584,0.022210665875011023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,32,0.0018880000958840053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,16384,0.3146959940592448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,2560,0.02161600026819441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,768,0.011031111081441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,6144,0.07418488793902926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,3072,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,12288,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,3072,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,7168,0.06230400005976359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,12288,0.13719911045498318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,512,0.7534284591674805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,2560,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,5120,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,65536,2.210223091973199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,6144,0.01407199932469262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,2048,0.03394666645261977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,768,0.08281244171990289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,51200,1.7947396172417536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,256,0.003321777822242843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,8192,0.045775112178590566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,256,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,2560,0.04495733314090305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,51200,1.032849735683865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,65536,0.4852897855970595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,8192,0.13467110527886286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,5120,3.9792391459147134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,16384,0.024838222397698298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,7168,0.023619555764728125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,1536,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,1024,0.1906426615185208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,1536,0.012799111505349478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,6144,0.02472355630662706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,6144,0.12281244330936009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,6144,0.6991733445061578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,768,0.030728889836205378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,16384,0.054754667811923556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,10240,0.016248888439602323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,10240,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,2560,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,16384,0.38133955001831055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,1536,0.008805333740181392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,1536,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,1024,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,2048,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,64,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,12288,0.06898488601048787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,16384,0.10785422060224746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,10240,0.010407999985747868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,128,0.005388444496525659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,51200,0.05084088775846693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,3072,0.04016800059212579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,51200,0.5345519913567437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,51200,0.061391108565860324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,8192,0.040059556563695274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,5120,0.023709333605236475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,65536,0.20051111115349662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,65536,0.05678755707210965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,5120,0.004243555582231945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,2560,0.030822220775816176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,768,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,3072,0.028526223368114893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,3584,0.016934222645229764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,256,0.011594666375054253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,4096,0.00812977800766627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,2048,0.011418666276666852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,1024,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,51200,0.4597173266940647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,512,0.017457778255144756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,51200,0.04288711150487264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,12288,0.2436079978942871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,5120,0.021112889051437378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,7168,0.15857244862450495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,2048,0.27519112163119847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,16384,0.3092995484670003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,65536,0.11975377135806614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,3584,0.0090524446633127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,512,0.0028293333533737394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,128,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,768,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,10240,0.029209779368506536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,768,0.04131288992034064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,16384,8.493169996473524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,7168,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,16384,0.02658577760060628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,512,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,512,0.10159377919303046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,2048,0.019557333654827543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,1536,0.030311998393800523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,3072,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,3072,0.08132533232371013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,768,0.025060445070266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,1536,0.010054222411579555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,128,0.011883555187119378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,64,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,1536,0.03600622216860453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,6144,0.04479555620087517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,4096,0.09732533163494533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,3584,0.06463111109203763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,5120,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,512,0.00574311117331187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,256,0.013291555146376291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,7168,0.04265244470702278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,1024,0.06389244397481282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,6144,0.004936000125275718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,6144,0.05142755640877617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,256,0.037999110089408025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,7168,0.23096267382303873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,512,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,512,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,2048,0.09111555417378743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,7168,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,10240,0.011923555698659686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,1024,0.027043556173642475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,1024,0.01724799970785777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,3584,0.1555360025829739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,128,0.014579556054539151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,10240,0.12761599487728542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,3584,0.031088001198238794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,8192,0.13857421610090467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,256,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,6144,0.1528328922059801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,6144,0.050880889097849526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,768,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,16384,0.18914222717285156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,7168,0.01925511161486308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,3584,0.008550222549173567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,8192,0.012412444584899478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,3584,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,1536,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,3072,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,10240,2.9460870954725475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,256,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,10240,0.06665155622694227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,128,0.026941332552168105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,4096,0.052483555343416005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,10240,3.4737875196668835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,4096,0.10336177878909641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,16384,0.15690578354729548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,256,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,10240,0.016051555673281353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,16384,0.030012445317374334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,7168,0.05488444368044535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,2048,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,2048,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,12288,0.10622221893734402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,5120,0.00904177791542477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,128,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,10240,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,3072,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,7168,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,3072,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,10240,0.03412177827623155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,2560,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,65536,0.021929777330822412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,65536,0.6848764419555664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,32,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,1536,0.01827911039193471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,3584,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,12288,0.22528711954752603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,12288,0.017818666166729398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,5120,0.031143112315071955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,3072,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,2048,0.00903733323017756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,1024,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,128,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,2560,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,8192,0.026199999782774184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,2048,0.004555555681387584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,8192,0.021389333738221064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,256,0.021983999345037673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,3584,0.024100444383091394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,2048,0.2605022324456109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,128,0.014694222145610385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,3584,0.04619910981920031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,51200,0.4487742318047418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,12288,0.14145067003038195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,16384,0.01620444489849938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,1024,0.014470222923490735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,65536,0.0654382242096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,51200,0.17931644121805826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,65536,0.2927164501614041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,2048,0.004208000169859992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,1024,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,8192,0.3097146617041694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,1024,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,10240,0.0339955555068122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,2048,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,3584,0.27772710058424205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,768,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,2048,0.007329777710967594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,3584,0.06867021984524198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,3584,0.004554666578769684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,5120,0.04394577940305074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,6144,0.018741332822375827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,32,0.009460444251696268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,128,0.030373334884643555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,2048,0.010664000279373594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,2048,0.021029333273569744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,2560,0.04012444284227159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,1536,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,51200,0.05787110990948147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,7168,0.01979555520746443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,3072,0.025256888733969793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,512,0.046952890025244824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,256,0.01609511176745097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,6144,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,3072,0.043037331766552396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,64,0.020672889219390023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,10240,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,32,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,6144,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,6144,0.028676443629794653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,3584,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,1536,0.01923644377125634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,64,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,128,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,5120,0.009791110952695211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,16384,0.38937332895067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,256,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,10240,1.3227910995483398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,1024,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,2560,0.007719111111429002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,256,0.009124444590674506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,1536,0.05498399999406603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,4096,0.014988443917698331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,2048,0.15020533402760824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,7168,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,128,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,768,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,51200,0.05429066552056206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,1024,0.05389777819315592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,1024,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,4096,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,4096,0.07973155710432264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,768,0.03801155421468947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,8192,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,10240,0.025435555312368605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,2560,0.007003555695215861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,65536,0.44967291090223527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,65536,0.43087999025980633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,51200,0.7088933520846896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,10240,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,1536,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,1536,0.024067555864651997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,6144,0.00963199966483646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,65536,0.0597297747929891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,51200,0.18169600433773467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,32,0.0018853334089120228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,256,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,256,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,3584,0.044414222240448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,3584,0.022669333550665114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,128,0.03712711069318984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,6144,0.032271110349231295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,4096,0.09586844179365371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,256,0.017228444417317707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,16384,0.05198755529191759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,2560,0.04599022203021579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,256,0.019895111521085102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,2048,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,5120,0.013952889376216464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,5120,0.02693155573474036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,51200,0.04838488830460442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,2048,0.013385777672131857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,6144,0.4608951144748264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,8192,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,12288,0.0605857769648234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,2560,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,10240,0.008170666793982187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,128,0.00851999968290329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,768,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,3584,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,51200,0.0925333301226298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,51200,1.1310266918606229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,51200,0.678168879614936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,256,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,5120,0.029683556821611192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,6144,0.034563554657830134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,64,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,7168,0.052800890472200185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,128,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,3584,0.02269511090384589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,2560,0.017986666825082567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,128,0.008369777765538957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,2560,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,3584,0.1273662249247233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,3072,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,65536,0.365955564710829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,65536,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,2560,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,256,0.010967111421955956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,3072,0.009191999832789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,6144,0.01386755539311303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,51200,0.09191644191741943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,256,0.05425511134995354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,10240,0.11543644799126519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,4096,0.310332457224528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,64,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,3072,0.06168711185455322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,1536,0.20792978339725068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,32,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,768,0.008112000094519721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,256,0.006456888798210356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,64,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,768,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,2048,0.11533689498901367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,6144,0.03961866762903001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,6144,0.03195022212134467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,51200,1.0543173684014213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,32,0.029767112599478826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,256,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,2048,0.022984888818528917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,1024,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,10240,0.03650133477316962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,2048,0.015178667174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,5120,0.00904177791542477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,512,0.004905777672926585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,1024,0.025630222426520452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,4096,0.042169776227739125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,768,0.009118222528033787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,1536,0.019983111156357657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,3584,0.028396444188223943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,2048,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,5120,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,768,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,16384,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,128,0.006034666465388403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,1024,0.016167999969588388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,12288,0.015999111864301894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,256,0.007048889166778988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,6144,0.06257155868742201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,5120,0.08902044428719415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,1536,0.08885155783759223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,2560,0.041915555795033775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,3584,0.014736000034544202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,3584,0.04283466604020861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,4096,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,4096,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,512,0.013840888937314352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,12288,0.030795554320017498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,2048,0.03954844342337714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,5120,0.0214711113108529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,51200,0.01676177812947167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,512,0.00699733363257514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,128,0.01332800007528729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,64,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,12288,0.14663555887010363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,768,0.022379555635982092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,64,0.0018880000958840053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,2048,0.030198223061031763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,2560,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,256,0.012055111428101858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,128,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,5120,0.023141332798533972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,256,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,12288,0.07653955618540446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,3584,0.022967999180157978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,3072,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,768,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,64,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,32,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,256,0.023696889479955036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,1024,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,65536,0.0628604425324334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,7168,0.017648888958825003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,3584,1.2685911390516493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,12288,0.08045155472225614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,3584,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,768,0.016142222616407607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,256,0.01832799944612715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,16384,0.13226755460103354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,3072,0.29116177558898926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,1536,0.019242667489581637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,8192,0.015417777829700045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,8192,0.5152293311225044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,3584,0.040455109543270536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,256,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,128,0.005399111244413588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,10240,0.1232444445292155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,128,0.011001778145631155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,256,0.0051893335249688895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,5120,0.09301688936021592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,51200,11.136616177029081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,10240,0.043250666724310986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,64,0.0018275556051068837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,5120,0.035066667530271746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,1536,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,65536,0.4875955581665039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,12288,0.031346665488349065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,65536,0.2367102305094401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,16384,0.4066933261023627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,5120,0.024380443824662104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,2048,0.010977778169843884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,65536,0.04933422141604953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,6144,0.07482133309046428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,6144,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,128,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,6144,0.007429333196745978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,2560,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,256,0.026758222116364375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,12288,0.0424924459722307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,12288,0.06263733572430082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,16384,0.03784888982772827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,256,0.011873777541849347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,51200,0.051219556066724986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,5120,0.029665778080622356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,2560,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,512,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,12288,0.0199635558658176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,64,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,3584,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,128,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,2048,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,2048,0.09394666883680557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,65536,0.2783217695024279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,2048,0.0418542226155599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,5120,0.018936889039145578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,1536,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,128,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,1024,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,65536,0.1482737726635403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,16384,0.039953778187433876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,3072,0.028350220786200628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,12288,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,768,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,2560,0.04026577870051066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,12288,0.08789333369996812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,5120,0.02537333303027683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,128,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,51200,0.11848089430067275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,128,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,6144,0.022350221872329712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,128,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,65536,6.266009860568577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,1536,0.01587733295228746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,16384,1.0135013792249892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,12288,1.327294243706597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,128,0.03267200125588311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,1024,0.020336889558368258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,8192,0.018248000078731112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,128,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,7168,0.19272800286610922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,6144,0.17439644866519502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,2560,0.869525326622857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,128,0.01401511165830824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,65536,1.9272649553087022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,6144,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,16384,0.02473955518669552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,1536,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,64,0.003866666720973121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,128,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,256,0.022993778189023335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,51200,0.12841333283318415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,8192,0.01924533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,3072,0.023198222120602924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,8192,0.04271466533342997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,4096,0.012911111116409302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,256,0.0052986666560173035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,4096,0.01824355622132619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,1024,0.09852355718612671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,2048,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,16384,0.05797866980234782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,51200,1.3388373057047527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,10240,0.034662223524517484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,51200,0.07682400279574923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,1536,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,7168,0.03382933470937941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,16384,0.03948355714480082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,12288,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,8192,0.032687998480266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,1536,0.014999111493428549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,16384,2.2948409186469183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,3072,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,10240,0.016931555337376065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,12288,0.03976533479160733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,5120,0.053378668096330434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,256,0.0824506680170695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,12288,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,12288,0.11247377925448948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,51200,0.05224800109863281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,7168,0.0069955554273393415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,128,0.004932444542646408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,6144,0.07553066809972127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,64,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,16384,0.06622133652369182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,12288,0.09651377465989854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,3584,0.007551111280918121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,6144,0.2194239960776435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,10240,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,10240,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,32,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,12288,0.01921422282854716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,51200,0.4418897893693712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,512,0.01999022232161628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,10240,0.17848977777693006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,3584,0.25305067168341744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,3584,0.027570666538344488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,3072,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,65536,3.0808294084337025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,16384,0.020783111453056335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,768,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,1536,0.07075288560655382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,8192,0.029383109675513372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,512,0.02474400069978502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,16384,0.049730665153927274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,5120,0.0991191135512458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,10240,0.23256354861789277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,512,0.006896889044178857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,1536,0.010773333410422007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,32,0.04634755518701342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,51200,1.2139689127604167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,2560,0.029159112109078303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,16384,0.012711111042234631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,12288,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,256,0.020690666304694284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,512,0.011352000137170156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,768,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,5120,0.20156889491611055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,3584,0.14533244238959417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,3584,0.1416071123547024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,512,0.0032355555643637977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,128,0.016474665866957772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,768,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,128,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,3584,0.6292639838324653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,10240,0.11817422178056504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,65536,1.1265377468532987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,2560,0.010245333115259806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,12288,0.02511111067401038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,256,0.0025439999169773525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,3584,0.02892177634769016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,768,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,4096,0.009178666604889764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,12288,0.06527200010087755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,2048,0.02473955518669552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,768,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,65536,0.06581777996487088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,3584,0.03559644354714288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,3584,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,5120,0.27786133024427623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,5120,0.011618666350841522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,3072,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,128,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,64,0.012757333616415659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,8192,0.09296266900168525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,12288,0.02626666592227088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,1536,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,256,0.021313778228229944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,128,0.027786665492587622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,256,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,12288,0.017658667431937326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,128,0.012456888953844706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,1536,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,1536,0.030273778571022883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,8192,0.10348888900544907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,10240,0.16052444775899252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,512,0.02167555524243249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,10240,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,256,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,7168,0.01076533314254549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,3072,0.027094221777386133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,3072,0.05893333090676201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,51200,0.19578666157192656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,1024,0.01089244418674045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,3072,0.015288000305493673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,3584,0.08981421920988296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,512,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,7168,0.06742133034600152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,10240,0.19392889075809053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,51200,0.12923999627431235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,5120,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,768,0.017492444978819955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,7168,0.09411111142900254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,256,0.16548621654510498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,3584,0.05802222092946371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,16384,0.02130399975511763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,12288,0.04760533240106371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,256,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,2560,0.014544000228246054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,1536,0.31465066803826225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,2048,0.031946665710873075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,768,0.07561688952975802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,768,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,10240,0.03392622205946181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,12288,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,8192,0.0988337794939677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,65536,2.8010391659206815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,51200,0.07500622007581922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,10240,0.02644444505373637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,512,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,3072,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,51200,0.10703555742899577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,51200,0.041709333658218384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,10240,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,8192,0.03014755580160353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,128,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,2560,0.009591999981138442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,3072,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,1024,0.0070222218831380205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,2560,0.02395733363098568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,65536,0.08471555842293634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,128,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,3072,0.05393777622116936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,12288,0.19870489173465303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,6144,0.18235733773973253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,6144,0.011417778001891242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,256,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,64,0.07151377863354154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,16384,0.01957155598534478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,3584,0.06977422369851007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,4096,0.01608444419172075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,7168,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,1024,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,16384,0.30253156026204425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,12288,0.013832888669437833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,2560,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,51200,0.17863644493950737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,7168,0.027954667806625366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,4096,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,1024,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,6144,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,16384,0.5663626458909776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,16384,0.05656533108817207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,256,0.02096533278624217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,7168,0.014006222287813822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,2560,0.01367111090156767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,256,0.0181368887424469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,5120,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,2048,0.005839111076460944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,1024,0.0343795551194085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,7168,0.04439199964205424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,8192,0.020256888535287645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,1024,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,8192,0.028632889191309612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,128,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,768,0.11791377597384983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,8192,0.05908711089028252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,1024,0.01590844492117564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,16384,0.043346666627460055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,3584,0.37505332628885907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,2048,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,65536,0.2879893249935574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,7168,0.05460177858670553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,768,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,10240,0.02642400066057841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,65536,0.60115999645657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,1024,0.04163199994299147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,51200,0.06334133280648126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,2560,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,256,0.013786666923099093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,16384,0.25000090069240993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,2560,0.02020711037847731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,1536,0.01018311083316803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,1536,0.02268444498380025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,1024,0.007243555453088548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,2048,0.07041600015428331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,768,0.010634666515721215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,65536,2.0816622840033636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,32,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,3072,0.013389333254761167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,256,0.023169777459568445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,8192,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,128,0.17391467094421387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,4096,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,256,0.009328888522254096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,51200,0.015955555770132277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,10240,0.017976888351970248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,64,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,32,0.0042035554846127825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,51200,0.3346213234795465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,128,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,256,0.023167999254332647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,3584,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,1024,0.0396088891559177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,16384,0.20965066221025255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,512,0.02089511189195845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,1536,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,32,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,512,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,10240,0.024002666274706524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,512,0.026086222794320848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,768,0.016513778103722464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,2560,0.03347999850908915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,5120,0.0989733338356018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,2048,0.24857778019375273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,2048,0.03228266702757941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,10240,0.026836444934209187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,32,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,3584,0.04483200112978617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,3072,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,512,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,1024,0.0038968887594011095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,16384,0.08151288827260335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,1024,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,512,0.08965600199169582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,128,0.014854222536087036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,2560,0.05647199683719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,4096,0.015614221493403116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,512,0.017840888765123155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,51200,2.4448346032036676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,128,0.014018666413095264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,3584,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,3584,0.02165866725974613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,10240,0.059260441197289355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,2560,0.02405866649415758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,6144,0.00675022229552269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,256,0.01670222149954902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,10240,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,7168,0.03052977720896403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,16384,0.016976000534163583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,16384,0.016138666205936007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,6144,0.022831110490692988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,768,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,51200,0.06373244524002075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,10240,0.01648177703221639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,12288,0.02609599961174859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,32,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,64,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,8192,0.12357600529988606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,12288,0.011354666617181567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,8192,0.02404088940885332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,3072,0.20760977268218994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,5120,0.033944000800450645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,4096,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,512,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,7168,0.058784888850318066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,7168,0.3123733202616374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,51200,0.03972088959481981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,32,0.0018097776919603348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,4096,0.009527111219035255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,64,0.05775022506713867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,5120,0.019640889432695176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,4096,0.023911999331580266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,1024,0.02676711148685879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,768,0.017151110702090792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,32,0.0038337777886125776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,2560,0.07463910844590929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,256,0.015614221493403116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,256,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,3072,0.3779573440551758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,64,0.027931554449929133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,8192,0.07574933105044894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,2048,0.05123022198677063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,768,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,3072,0.09287466605504353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,6144,0.014665777484575907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,1536,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,768,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,32,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,64,0.008214222060309516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,128,0.05928977992799547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,2560,0.032891554964913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,5120,0.018206222189797294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,7168,0.027383999692069158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,512,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,8192,0.16695111327701143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,3584,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,8192,0.12743200196160212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,128,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,1024,0.02088800072669983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,64,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,3584,0.022456889351209004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,1536,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,65536,2.003041797214084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,6144,0.043748444981045194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,256,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,2048,0.010055111514197456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,768,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,10240,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,12288,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,2048,0.02546577817863888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,2560,0.06268178092108832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,7168,2.868605295817057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,768,0.014733332726690503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,6144,0.030976888206270006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,1024,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,3584,0.019871110717455547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,12288,0.040045334233178034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,6144,0.034992890225516424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,8192,0.02994666828049554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,768,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,6144,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,2560,0.027855998939938013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,768,0.006396444307433234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,128,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,51200,1.0450258255004883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,4096,0.016295111841625638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,7168,0.08247733116149902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,1024,0.06847111384073894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,128,0.004137777619891697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,65536,0.008426666259765625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,5120,0.012999110751681857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,4096,0.23523465792338052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,256,0.022777777579095628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,768,0.009966221948464712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,12288,0.042183111111323036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,16384,0.09897955258687337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,768,0.010398222340477837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,16384,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,1024,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,2560,0.06124266650941637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,3072,0.015999111864301894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,64,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,6144,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,1024,0.022780444886949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,512,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,2048,0.05031733380423652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,2560,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,3072,0.00944977750380834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,3072,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,256,0.0081386665503184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,6144,0.08522933059268528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,512,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,8192,0.020650666620996263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,5120,0.00905866672595342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,16384,0.02274311085542043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,1024,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,7168,0.01962933275434706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,2048,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,2560,0.010652444428867765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,1024,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,5120,0.029707554313871596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,1536,0.01389244364367591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,1024,0.02648977769745721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,51200,0.06171022521124946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,3072,0.07112889157401191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,1536,0.03648444347911411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,10240,0.7430062294006348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,16384,0.4123413297865126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,3072,0.023220444718996685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,65536,5.90029059516059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,32,0.01800355480776893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,16384,0.020602666669421725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,51200,0.9013466305202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,5120,0.01165066659450531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,512,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,8192,0.052707556221220225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,3584,0.04151022103097703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,12288,0.027239110734727647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,12288,0.2786906560262044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,2560,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,65536,0.07804444763395521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,3584,0.03353955679469638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,6144,0.032624890406926475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,8192,0.03204177816708883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,4096,0.024319110645188227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,16384,0.018567111757066514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,8192,0.035439110464519925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,256,0.01552977826860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,51200,0.04212622178925408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,128,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,2560,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,2560,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,768,0.03741066654523214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,1024,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,3584,0.05059022373623318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,6144,0.023352000448438857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,65536,0.0626595550113254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,32,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,65536,0.07801066504584418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,768,0.015294222368134392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,512,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,51200,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,768,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,51200,0.05403377612431844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,65536,2.036475499471029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,256,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,512,0.004552000098758274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,3584,0.05430577860938179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,7168,0.079876442750295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,128,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,4096,0.08540266752243042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,2048,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,7168,0.022954665952258643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,768,0.028198222319285076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,7168,0.02641511129008399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,256,0.018178666631380718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,6144,0.026693332526418898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,256,0.002995555599530538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,1024,0.04141689009136624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,5120,0.008225777910815345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,1536,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,7168,0.01752800080511305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,1536,0.04231199953291151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,1024,0.022312000393867493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,65536,0.12876178158654109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,768,0.028093334701326158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,8192,0.293250666724311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,2560,0.02368444369898902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,2560,0.0273324449857076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,3072,0.018199111024538677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,3584,0.12473244137234157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,5120,0.012791111237472959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,12288,0.023698666029506262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,10240,0.018936889039145578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,4096,0.03613333238495721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,16384,0.3887431091732449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,65536,0.24956711133321127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,512,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,12288,0.046965334150526256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,2560,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,65536,2.3197858598497176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,12288,0.05196088883611891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,5120,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,128,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,64,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,128,0.005021333280536863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,256,0.047056890196270414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,32,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,2048,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,768,0.011723555624485016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,65536,0.05988888608084785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,2048,0.00462311092350218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,512,0.011400889191362592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,10240,0.7836417622036405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,32,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,2048,0.008219555848174626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,128,0.17382489310370552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,2560,0.050501333342658154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,32,0.02166044380929735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,51200,0.020920889245139226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,8192,0.04308266772164238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,10240,0.05297866794798109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,8192,0.286373323864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,12288,0.06649866369035509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,1024,0.013472889032628802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,65536,0.811921755472819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,2048,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,65536,0.6951502164204916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,2560,0.023261333505312603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,5120,0.022481777601771887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,256,0.019941333267423842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,256,0.021627555290857952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,16384,0.29300088352627224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,12288,0.05336266756057739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,5120,0.06447111235724555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,8192,0.016710221767425537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,1536,0.017830222845077515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,2560,0.020632889535692003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,6144,0.013372444444232516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,6144,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,3584,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,3584,0.04329866502020094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,64,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,6144,0.13470666938357884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,7168,0.07797511418660481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,64,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,10240,0.11970399485694037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,1536,0.006451555424266391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,7168,0.019535111056433786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,3072,0.1881155570348104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,4096,0.11115377479129368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,16384,0.06395200226042005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,3584,0.25290576616923016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,64,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,512,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,3584,0.08098755280176799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,128,0.00666133314371109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,5120,0.012993777791659037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,128,0.001790222194459703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,3072,0.15592532687717015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,128,0.006040888941950268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,3072,0.014517333772447375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,128,0.005533333453867171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,768,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,256,0.007798222204049428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,2560,0.017924444542990792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,51200,0.052389333645502724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,65536,0.07093333535724215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,512,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,3072,0.34642399681939023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,6144,0.015198222464985318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,256,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,1536,0.13173333803812662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,512,0.011112000379297467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,2560,0.015191111299726697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,6144,0.03169155451986525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,1536,0.01980533368057675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,1024,0.02353777819209629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,768,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,1024,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,768,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,128,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,51200,0.454187552134196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,6144,0.02420711186197069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,16384,0.06017688910166422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,1536,0.070360885726081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,10240,26.421671549479168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,7168,0.03530755639076233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,12288,0.035522666242387556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,2048,0.013380444712109037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,8192,0.4044533305697971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,16384,0.042344888051350914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,1024,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,7168,0.01109333336353302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,10240,0.029927998781204224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,5120,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,256,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,10240,0.02752711044417487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,2560,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,5120,0.04505955510669284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,1536,0.018616000811258953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,4096,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,5120,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,768,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,5120,0.03067644437154134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,768,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,2560,0.041145778364605375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,32,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,3584,0.06726488802168104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,128,0.002945777856641345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,12288,0.013175110850069257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,128,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,7168,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,16384,0.02773688899146186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,128,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,512,0.020948444803555805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,2048,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,4096,0.08893244134055243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,8192,0.014646222194035849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,5120,0.020611556039916146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,3584,0.04414133230845133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,12288,0.024123556084103052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,1024,0.008040889269775814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,1024,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,256,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,32,0.009717333647939894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,3072,0.0651822222603692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,128,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,51200,0.12214844756656223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,256,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,7168,0.027279110418425664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,10240,0.22716887791951498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,4096,0.16871110598246256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,512,0.02438844409253862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,2048,0.03648444347911411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,16384,0.2806088924407959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,3584,0.1459724505742391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,768,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,32,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,32,0.07376088698705037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,6144,0.14383110735151503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,12288,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,16384,0.021011556188265484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,5120,0.0705804427464803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,12288,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,51200,0.05765599674648709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,64,0.0017928888814316855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,256,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,5120,0.05215377940071953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,2560,0.007085333267847697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,1536,0.014787554740905762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,3072,0.00907199995385276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,3072,0.03494222296608819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,7168,0.03881599836879306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,65536,0.17015911473168266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,1536,0.010159999959998662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,4096,0.015585778488053216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,3072,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,7168,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,4096,0.08922666973537868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,512,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,512,0.0028728888266616394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,65536,0.9469110700819227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,12288,0.15777511066860622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,4096,0.018025777406162687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,3072,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,4096,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,768,0.008780444661776224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,512,0.0025519999778932994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,65536,1.0825422075059679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,4096,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,8192,0.13092177444034153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,8192,0.15395288997226292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,32,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,2560,0.008785777621799046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,4096,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,16384,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,10240,0.035996443695492215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,5120,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,32,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,12288,0.030846221579445735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,65536,0.06560177935494317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,3072,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,768,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,7168,0.024505777491463557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,128,0.006743999818960826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,128,0.006314666734801398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,6144,0.07267466518614027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,7168,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,5120,0.14514577388763428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,2560,0.030220445659425523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,2048,0.03643911083539327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,1536,0.09753600094053481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,51200,0.5205235481262207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,5120,0.012302222351233164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,2560,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,768,0.006190222170617845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,1024,0.008231999973456064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,65536,0.0993466642167833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,256,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,4096,0.3947048981984456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,65536,0.14156799846225315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,65536,0.6612435446845161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,8192,0.05668533510631985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,65536,2.3794631958007812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,10240,0.7607999907599555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,16384,0.025999110605981614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,5120,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,7168,0.025454221500290766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,7168,1.1432889302571614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,3584,0.10478755500581528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,2560,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,16384,0.02405333353413476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,3584,0.1757733292049832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,64,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,64,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,65536,1.9032346937391493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,8192,0.08134133285946317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,65536,0.07406844695409139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,1024,0.008831111093362173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,65536,0.030960887670516968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,4096,0.04909422331386142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,3072,0.043008890416887074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,6144,0.03709244396951463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,512,0.08847110801272923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,65536,0.6284151077270508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,7168,0.9477973514133029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,16384,0.28238844871520996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,51200,10.230884975857204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,5120,0.012167111039161682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,65536,0.0958906677034166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,2048,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,51200,0.06573777728610568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,1536,0.00646577775478363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,128,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,5120,0.03146400054295858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,16384,0.039615111218558416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,5120,0.03163911236657037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,3584,0.00871022211180793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,32,0.05449688765737745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,4096,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,2560,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,2560,0.015423110789722867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,10240,0.43433154953850645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,64,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,1024,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,3072,0.012809777425395118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,3072,0.013796444568369122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,2048,0.006492444624503453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,65536,0.23129955927530924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,10240,1.152342266506619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,16384,0.05605955256356133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,5120,0.012538666526476542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,768,0.041546665959888034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,1536,0.004905777672926585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,768,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,768,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,2560,0.029492444462246362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,2560,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,512,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,10240,0.048342221313052706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,32,0.0058648888435628675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,8192,0.02477866742346022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,65536,0.13383377922905818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,4096,0.030710223648283217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,512,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,1024,0.010205333431561789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,2048,0.025228444072935317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,4096,0.018440888987647164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,51200,0.10374489095475938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,2560,0.004934222333961063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,12288,0.02075111038155026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,2560,0.014455111490355598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,64,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,2560,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,12288,0.7029991149902344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,12288,0.22072711255815294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,51200,0.08037955231136747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,5120,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,768,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,10240,1.2292764451768663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,1536,0.03070488903257582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,5120,0.08393866486019558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,3584,0.024681778417693243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,10240,0.04128799835840861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,7168,0.3223297860887316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,12288,0.4227084318796794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,12288,0.08012977573606703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,128,0.11429156197441949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,2048,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,32,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,3072,0.016061334146393668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,3584,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,1536,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,768,0.01351199961370892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,6144,0.31536799006991917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,512,0.011651555697123209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,4096,0.03088888857099745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,512,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,512,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,51200,1.355619536505805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,12288,0.5462986628214518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,512,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,6144,0.07970044347974989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,3072,0.2730871041615804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,51200,1.7973236507839625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,512,0.018630221486091614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,7168,0.7672515445285373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,2048,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,7168,0.03728266557057699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,4096,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,12288,0.06226844257778592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,256,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,2048,0.04882399903403389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,16384,0.033181332879596286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,7168,0.019846222466892667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,8192,0.022181333767043218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,3584,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,1536,0.018921777606010437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,8192,0.01276266657643848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,4096,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,512,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,8192,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,2560,0.09194755554199219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,7168,0.03067644437154134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,6144,0.8450186517503527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,8192,0.018862222631772358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,3584,0.10849333471722072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,3584,0.16592266824510363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,8192,0.03018933203485277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,16384,0.01646133263905843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,7168,0.013777777552604675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,256,0.007697777615653143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,256,0.004865777575307422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,7168,1.0060542424519856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,8192,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,2048,0.4991031222873264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,256,0.003535110917356279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,65536,0.08176000250710382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,12288,0.055649777253468834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,3584,0.1995404428905911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,768,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,65536,0.07739644580417208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,8192,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,2560,0.041308446062935725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,4096,0.06483199861314562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,8192,0.592540423075358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,10240,0.045732445187038846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,3072,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,6144,0.02656800051530202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,256,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,1536,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,2560,0.01738666660255856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,1024,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,3584,0.031160000297758315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,1024,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,7168,0.04347822070121765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,256,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,512,0.0058204444746176405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,512,0.04219377703136868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,256,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,10240,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,51200,0.11611199378967285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,2560,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,2560,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,51200,0.05456088980038961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,65536,1.264635509914822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,1024,0.025441777375009324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,5120,0.03626311156484816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,3584,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,12288,0.1036542190445794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,768,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,7168,0.0229413327243593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,256,0.014441778262456259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,256,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,3072,0.12064533763461643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,8192,0.04515555500984192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,64,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,1536,0.5981946521335179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,64,0.04333600070741442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,7168,0.03251555562019348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,10240,0.08997778097788493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,32,0.0024782222592168385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,32,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,65536,0.3133422268761529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,1024,0.019350222415394254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,128,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,3072,0.007363555332024892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,256,0.07407910956276788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,768,0.03763644562827216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,3584,0.009415110780133141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,128,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,10240,0.04704622096485562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,768,0.005730666634109285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,128,0.0075866662793689305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,2048,0.023520888553725347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,6144,0.033922665648990206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,1536,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,768,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,51200,0.0960693359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,65536,0.9714275995890299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,3072,0.022800000177489385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,12288,0.3289742204878065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,65536,1.5542870627509222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,12288,0.03832888934347365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,4096,0.03237955437766181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,5120,0.15835732883877224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,10240,0.054061333338419594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,768,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,1536,0.21600266297658285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,12288,0.489024904039171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,2560,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,3584,0.020448888341585796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,512,0.02365511159102122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,512,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,7168,0.010073777702119615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,3584,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,12288,0.01187999960449007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,6144,0.010527999864684211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,512,0.11069866683748032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,8192,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,6144,0.016571554872724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,12288,0.021967111362351313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,7168,0.12377599875132243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,4096,0.04984622200330099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,12288,0.8439884185791016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,5120,0.20150489277309844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,2048,0.006211555666393704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,4096,0.09809422492980957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,3072,0.013040888640615674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,12288,0.04133955637613932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,6144,0.35401421123080784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,768,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,51200,0.0380328893661499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,2560,0.02607466611597273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,51200,0.0726755526330736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,64,0.005970666805903117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,7168,0.053751998477511935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,512,0.09348889191945393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,512,0.0399617784553104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,10240,0.017845332622528076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,7168,0.017916444275114272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,16384,0.06621066729227702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,51200,0.04875466558668348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,10240,0.08020355304082234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,2560,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,6144,0.027779555983013574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,128,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,3584,0.023960888385772705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,4096,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,512,0.006035555568006303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,10240,0.017261332935757108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,2560,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,2048,0.014783999986118741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,8192,0.11978044774797227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,7168,0.09302666452195908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,768,0.01313777764638265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,8192,0.025723555021815833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,3072,0.008467555873923832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,1024,0.05770488580067953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,12288,0.148707562022739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,65536,0.05949155489603678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,51200,0.9136906729804145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,3584,0.32533245616488987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,256,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,7168,0.13827821943495008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,16384,0.06024266613854302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,10240,0.16184622711605495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,12288,0.3302035596635607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,3584,0.31095290184020996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,8192,0.024851555625597637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,3584,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,12288,0.24306045638190377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,512,0.007680888805124495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,3584,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,1536,0.00721244431204266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,512,0.02867199977238973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,3072,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,12288,0.0236817780468199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,10240,1.0249919891357422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,12288,0.035911109712388777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,2560,0.02679022153218587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,12288,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,2048,0.039408889081743025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,65536,0.07573688692516752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,128,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,3072,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,1536,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,5120,0.05095022254519992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,3584,0.021740444832377966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,1024,0.04241333405176798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,2048,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,16384,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,51200,0.1949297719531589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,8192,0.062209778361850314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,64,0.0032871110985676446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,65536,0.08528977632522583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,2048,0.008076444268226624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,16384,0.3717919985453288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,4096,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,65536,0.05913955635494656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,7168,0.03556266758177016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,768,0.028024888700909082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,1536,0.12700088818868002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,1024,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,7168,0.0432222220632765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,7168,0.034252444903055825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,7168,0.02720444401105245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,256,0.003938666648334927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,256,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,8192,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,51200,0.17097955279880098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,2560,0.6725751029120551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,128,0.002785777880085839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,6144,0.041999111572901406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,16384,0.1302151150173611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,64,0.003559111307064692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,16384,0.06533155838648479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,12288,0.06557866599824694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,7168,0.044812444183561534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,10240,0.023375999596383836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,4096,0.11338578330145942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,7168,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,1536,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,64,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,2048,0.03239999877081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,1536,0.16685244772169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,7168,0.02453600035773383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,1536,0.06159466505050659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,2560,0.009391999906963771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,10240,0.09342310825983684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,3584,0.141628450817532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,2048,0.004568888909286923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,2048,0.07985599835713704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,512,0.007154666715198093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,4096,0.022693332698610093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,3584,0.018139556050300598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,128,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,3072,0.014749333262443542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,2048,0.025245333711306255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,5120,0.019886222150590684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,512,0.007327111230956183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,65536,1.3706533643934462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,51200,0.048539555735058255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,10240,0.020480889413091872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,1024,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,768,0.029867556360032823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,16384,0.3445342116885715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,3072,0.02094399929046631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,1536,0.003859555555714501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,51200,0.37419999970330137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,3072,0.008414222134484185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,51200,0.3997653325398763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,128,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,768,0.02092355489730835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,5120,0.01348000019788742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,2048,0.014468444718254937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,256,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,256,0.016878222425778706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,1536,0.02623111175166236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,5120,0.16040978166792128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,51200,0.10753777954313491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,7168,0.10477155447006226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,1024,0.01926399932967292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,1536,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,1024,0.0034933334423436057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,16384,0.12244178189171685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,64,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,1536,0.055757330523596875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,4096,0.007829333345095316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,12288,0.011287111375066968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,512,0.024725332856178284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,1024,0.05633599890602959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,1536,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,512,0.013713777893119387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,7168,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,2048,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,128,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,64,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,12288,0.03489955597453647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,8192,0.025202666719754536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,8192,4.054645538330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,7168,0.5539475546942817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,512,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,256,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,64,0.016013332539134555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,7168,0.016393777396943834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,65536,0.5657484266493056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,51200,0.11858310964372422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,512,0.004959110998445087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,12288,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,4096,0.03165955675972833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,1536,0.05055555701255798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,8192,15.559961954752604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,5120,0.5674231317308214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,8192,0.029712888929578993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,128,0.013412444127930535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,65536,0.4665110905965169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,51200,2.233018663194444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,65536,0.0592764417330424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,6144,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,512,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,1024,0.2235297891828749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,4096,0.10557422373029922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,1536,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,3584,0.03518844313091702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,7168,0.009941333697901832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,16384,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,768,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,768,0.024813334147135418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,16384,0.03645777702331543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,512,0.009287111461162567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,1536,0.004022222426202563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,65536,13.250249226888021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,2048,0.01423733267519209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,16384,0.05937777625189888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,10240,0.640359984503852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,12288,0.05442400111092461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,512,0.005766222046481238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,51200,3.336664835611979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,768,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,6144,0.024696888195143804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,256,0.010991999672518836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,3584,0.006062222023804982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,4096,0.5540062056647407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,10240,0.05327999922964308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,10240,0.023981332778930664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,512,0.01921777758333418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,5120,0.044545776314205594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,8192,0.02279377811484867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,128,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,256,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,12288,0.07116533650292291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,1536,0.026123555170165166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,65536,0.2145013279385037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,3072,0.014160888062583076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,3072,0.03821600145763821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,1024,0.02067466742462582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,1024,0.01665244499842326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,128,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,768,0.0034826666944556763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,3584,0.011241777903503843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,2560,0.01613066593805949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,4096,0.039043555657068886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,1536,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,64,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,256,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,512,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,8192,0.08142222298516168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,768,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,128,0.0063457778758472866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,256,0.014174222946166992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,5120,0.052057776186201304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,1536,0.020023110840055678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,5120,0.2959973282284207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,4096,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,2560,0.030244443151685927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,4096,0.02870133188035753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,5120,0.10926133394241333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,12288,1.0621200137668185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,8192,0.025309332542949255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,64,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,3072,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,3584,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,2048,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,2560,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,4096,0.5307333204481337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,32,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,51200,0.14049332671695286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,2560,0.07426133420732287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,768,0.011000889043013254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,3584,0.01612622208065457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,512,0.18016978104909262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,512,0.023249778482649062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,3584,0.009896888501114314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,6144,0.3504435486263699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,4096,0.18515555063883463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,7168,0.06780888636906941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,2560,0.03552622265285916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,12288,0.04796088735262553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,6144,0.05484799875153435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,16384,0.32758667733934194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,768,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,7168,0.08591466479831272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,65536,0.9237741894192166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,1024,0.016925333274735343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,1024,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,12288,0.08629777696397568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,6144,0.04863288998603821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,3072,0.01180622229973475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,8192,0.019368888603316415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,12288,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,256,0.01609422266483307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,65536,0.0684808890024821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,128,0.0027520000520679685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,7168,0.0657164454460144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,2560,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,256,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,2560,0.0817813343471951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,10240,0.029269334342744615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,7168,0.01995822290579478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,256,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,12288,0.2968133290608724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,51200,0.12888622283935547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,12288,0.21671999825371635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,64,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,512,0.006080889039569431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,16384,0.11883644262949626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,32,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,12288,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,12288,0.007675555845101674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,1024,0.01866044435236189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,32,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,256,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,12288,0.20639288425445557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,4096,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,65536,1.6370070775349934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,5120,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,7168,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,512,0.026241777671708003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,6144,0.021752888957659405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,3072,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,256,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,8192,0.020303999384244282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,65536,1.2204248640272353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,16384,0.07232444153891669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,16384,0.04055644406212701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,6144,0.012094222009181976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,512,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,768,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,5120,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,1536,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,768,0.027526222997241553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,16384,0.029351999362309773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,5120,0.022109333011839125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,512,0.010068444742096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,10240,0.02092622220516205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,16384,0.03994133406215244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,10240,0.026844445202085707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,768,0.019648000597953796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,32,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,768,0.0041404445138242514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,3584,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,1024,0.0031306667046414483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,3072,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,65536,0.0649813347392612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,2048,0.006787555499209299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,768,0.008355555435021719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,16384,0.04056088791953193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,768,0.18904444906446669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,1024,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,12288,0.013629333012633853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,1024,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,16384,0.028862221373452082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,1536,0.0041848888827694785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,4096,0.017640888690948486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,6144,0.018728888697094388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,10240,0.12795289357503256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,65536,0.1693377759721544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,51200,0.04567644331190321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,6144,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,256,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,256,0.005180444568395615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,8192,0.02015999952952067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,512,0.02202400068442027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,4096,0.026944888962639704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,7168,0.09306933482487996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,768,0.01666933298110962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,1024,0.019665777683258057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,2048,0.007028444773621029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,8192,0.029924445682101782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,16384,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,10240,1.3559990988837347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,16384,0.06110577450858223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,6144,3.8439814249674478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,256,0.011392888923486074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,4096,0.44545867707994247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,7168,0.1522808869679769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,7168,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,12288,0.0450817777050866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,1024,0.03767733441458808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,4096,0.04325955443912082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,768,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,2560,0.007188444336255391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,8192,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,65536,0.4172817866007487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,51200,0.08634310960769653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,2560,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,6144,0.05912711222966512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,8192,0.2396622233920627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,3584,0.01054133309258355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,4096,0.07959289020962186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,768,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,6144,0.06785066922505696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,2560,0.11450578106774224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,32,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,4096,0.009383111364311641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,7168,0.12689599725935194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,8192,0.8210248947143555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,65536,0.15044711695777044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,3072,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,512,0.005967999911970562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,5120,0.057983113659752734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,768,0.03861777649985419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,16384,0.14408177799648708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,6144,0.025761778155962627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,32,0.003976888954639435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,2560,0.00907555553648207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,1536,0.006393777827421824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,7168,0.05232622226079305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,1536,0.009748444788985783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,16384,0.06614666514926486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,5120,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,65536,6.739015791151259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,65536,0.17008177439371744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,3584,0.008399110701349046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,32,0.11790488825903998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,2048,0.007120889094140794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,3072,0.024715556038750544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,12288,0.033599111768934466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,2560,0.008198222352398766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,32,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,32,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,256,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,10240,0.26250576972961426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,51200,0.20412533813052705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,128,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,768,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,2560,0.017478222648302715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,256,0.010856889188289642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,512,0.0603333314259847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,6144,0.07200621896319918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,51200,0.2481466664208306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,1536,0.022282666630215112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,64,0.03808355662557814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,32,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,64,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,64,0.0177502218219969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,64,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,10240,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,3072,0.027962668074501887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,16384,0.024804444776640996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,6144,0.03489066825972663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,4096,0.009377777576446533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,3584,0.026808000273174707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,65536,0.14236977365281847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,12288,0.10786400238672893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,2048,0.07104355759090848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,128,0.006795555353164673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,128,0.03573777940538194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,1024,0.004639111045334074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,32,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,5120,0.025028443998760645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,256,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,3584,0.148054215643141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,5120,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,5120,0.05831288629108005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,256,0.015171556009186639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,4096,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,768,0.021728888154029846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,65536,0.25358578893873424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,256,0.029320889049106177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,256,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,64,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,256,0.026163554853863184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,7168,0.27314665582444936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,65536,0.07228355275260077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,768,0.11149688561757405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,65536,0.07105155785878499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,2560,0.026639110512203638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,8192,0.04278399878078037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,1024,0.01917155583699544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,3584,0.008061333662933773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,1024,0.14453066719902888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,2560,0.024358222881952923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,1536,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,3072,0.04050577680269877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,512,0.02028888960679372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,64,0.005515555540720622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,5120,0.05307199888759189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,768,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,3072,0.08828622102737427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,3072,0.014284445179833306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,51200,0.07257599963082208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,3072,0.0170471105310652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,65536,0.7905644310845269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,12288,0.02163644466135237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,8192,0.13565422428978816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,128,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,51200,1.734134250217014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,3584,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,1024,0.03400444322162204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,65536,0.06456710894902547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,1024,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,6144,0.017280888226297166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,1024,0.11552800072564019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,4096,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,32,0.0032213332338465583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,7168,0.023856888214747112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,51200,0.22256888283623588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,1024,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,12288,0.06192977560891045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,65536,0.13797599739498562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,7168,0.01128088848458396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,6144,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,128,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,3584,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,65536,1.2391199535793729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,16384,0.3235609001583523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,12288,0.023010666171709698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,256,0.019633778267436557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,7168,0.03820266657405429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,8192,0.04086666636996799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,128,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,3584,0.021220443977249995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,512,0.0038746665749284956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,256,0.002459555657373534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,7168,0.33495201004876035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,2048,0.0068408888247278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,8192,0.05163200034035576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,12288,0.12302222516801621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,7168,2.075002670288086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,8192,0.03475644522243076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,256,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,1536,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,10240,3.161784913804796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,4096,0.49009153578016496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,64,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,1536,0.025018667181332905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,12288,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,16384,0.020208888583713107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,1536,0.15789333979288736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,1024,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,5120,0.21010311444600424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,128,0.003590222034189436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,10240,0.013063111239009433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,128,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,12288,0.1751368840535482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,65536,1.039886262681749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,1536,0.023154666026433308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,1536,0.014610666367742749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,3584,0.018665777312384713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,64,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,10240,0.025270221961869135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,10240,0.3093297746446398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,51200,0.04751999841796028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,128,0.01759288873937395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,6144,0.027834667099846735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,7168,0.011976888610257043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,51200,0.4294133451249864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,6144,0.027087110612127516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,2048,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,8192,0.014746665954589844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,4096,0.07508977916505602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,256,0.013634666800498962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,8192,0.20148889223734537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,512,0.023917333947287664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,2560,0.2696133454640706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,1024,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,12288,0.02846755584081014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,768,0.051199999120500356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,51200,0.8845342000325521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,64,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,7168,0.0458826654487186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,16384,3.7778061760796438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,1024,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,16384,0.02681155502796173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,3584,0.03642222285270691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,256,0.008734222087595198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,768,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,1536,0.007511110769377814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,7168,0.06766044431262545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,3072,0.07061511278152466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,16384,0.2532595528496636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,8192,0.03213511241806878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,256,0.04061955544683669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,51200,0.32379643122355145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,6144,0.010577778021494547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,768,0.07912977536519368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,2560,0.02849066588613722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,8192,0.011878222227096558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,7168,0.013625777430004544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,65536,0.3509439892239041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,12288,0.03403466608789232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,10240,0.031275557147132024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,768,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,51200,0.5641164249844021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,12288,2.4563865661621094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,32,0.0032133333798911837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,51200,0.10290488931867812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,65536,0.562553776635064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,65536,0.08157866530948214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,32,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,7168,0.027760000692473516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,768,0.0031840000301599503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,16384,0.039255999856524997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,51200,0.19195732805464003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,4096,0.01813155578242408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,7168,0.023373333944214716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,256,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,5120,0.011090666883521609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,32,0.0026942222482628296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,1536,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,2560,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,7168,0.08736177947786118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,65536,0.08595289124382867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,2560,0.026418666044871014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,16384,0.11249866750505234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,10240,0.7213599946763781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,2048,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,4096,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,512,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,10240,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,4096,0.015601777368121676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,2048,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,3584,0.01149511088927587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,2048,0.27772532569037545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,1536,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,51200,0.5391839875115289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,8192,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,12288,0.017656000124083627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,3072,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,768,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,4096,0.027549333042568628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,65536,0.0684337748421563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,12288,0.0513315565056271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,4096,0.024578667349285547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,768,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,65536,0.0865395532713996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,2048,0.019949333535300363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,2048,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,2048,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,64,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,4096,0.07502666446897718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,2560,0.03270666797955831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,2048,0.026766222384240892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,256,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,2048,0.13340622848934597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,128,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,64,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,1536,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,6144,0.01871200071440803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,2560,0.037993778785069786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,65536,0.07617777585983276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,16384,0.013632888595263163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,2560,0.08231021960576375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,1536,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,2048,0.07250310977300008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,8192,0.010619555910428366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,12288,0.0378115541405148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,12288,17.756549411349827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,256,0.002459555657373534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,3584,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,64,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,512,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,10240,0.04089511103100247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,65536,0.009133333133326637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,128,0.03875289029545254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,32,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,7168,0.010440888504187265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,128,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,3584,0.11323822869194879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,65536,0.14739021990034315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,2048,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,1024,0.05748799774381849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,512,0.020294222566816542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,768,0.11404621601104736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,4096,0.02397866712676154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,4096,0.010623111493057676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,1024,0.030181331766976252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,8192,0.029890666405359905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,512,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,256,0.01642666591538323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,5120,0.1576897833082411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,8192,0.15554400285085043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,128,0.003946666502290302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,51200,1.1469946967230904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,51200,0.06365155511432223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,32,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,6144,0.013555555707878537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,3072,0.05298133360015022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,4096,0.019528888993793063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,12288,0.056108441617753774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,32,0.01404888927936554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,4096,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,65536,0.26078134112887913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,8192,0.15936622354719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,512,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,12288,0.028208000792397395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,51200,0.09476621945699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,10240,0.12926756011115179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,51200,0.1291617817348904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,1536,0.03199377655982971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,1536,0.022786666949590046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,3072,0.006713777780532837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,7168,0.014870221416155497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,8192,0.14384178320566812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,256,0.0038444445365005066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,51200,0.5119804276360406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,12288,0.01443911095460256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,1536,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,3072,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,3072,0.04053689042727152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,1536,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,512,0.03135910961363051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,12288,0.0605564448568556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,3072,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,3584,0.019004444281260174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,512,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,51200,2.9083928002251516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,10240,0.0737146668963962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,512,0.006767999794748094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,4096,0.28713777330186635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,8192,0.0884560015466478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,2560,0.027064000566800434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,16384,0.056755556000603564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,4096,0.05323466658592224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,3072,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,2048,0.32421689563327366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,51200,0.06298933426539104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,2560,0.010021333065297868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,16384,0.08039200305938721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,256,0.01688799924320645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,256,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,5120,0.01107022249036365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,7168,0.023414222730530634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,5120,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,1536,0.026664889521068994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,12288,0.9541493521796333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,64,0.03738844394683838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,1536,0.012162666353914471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,10240,0.11814399560292561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,3584,0.010151999692122141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,65536,0.089629332224528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,65536,0.1113999949561225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,512,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,12288,0.04550933175616794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,10240,0.012982221941153208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,2048,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,5120,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,3072,0.039270222187042236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,1536,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,12288,0.045363555351893105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,6144,0.08334133360120986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,65536,0.14669422308603922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,16384,0.024446222517225478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,3584,0.007799111306667328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,768,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,65536,0.09592800007926093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,32,0.005650666852792104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,64,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,3072,0.3743075529734294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,2560,0.05375733309321933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,64,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,1536,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,2048,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,7168,0.03755911191304525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,4096,0.023728888895776536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,1536,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,10240,0.1318968931833903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,128,0.015525332755512662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,1024,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,51200,0.20813422732883027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,65536,0.06746844450632732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,4096,0.05916533205244276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,5120,0.008786666724416945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,6144,0.03204266561402215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,16384,0.2483910984463162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,512,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,6144,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,1024,0.1585386726591322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,7168,0.017955554856194388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,6144,0.0588888857099745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,12288,3.653008778889974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,65536,1.222405327690972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,3072,0.12060622374216716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,7168,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,512,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,768,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,3584,0.03499555587768555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,7168,0.032756444480684065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,32,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,10240,0.034062223301993474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,128,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,2560,0.008369777765538957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,51200,0.24062487814161512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,2560,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,768,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,6144,0.009604444106419882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,3584,0.008839999636014303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,256,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,51200,0.009464888936943477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,2560,0.010592000352011787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,16384,0.030116445488399927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,10240,0.02639466689692603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,3584,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,65536,0.06247555547290378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,3584,0.016908443636364408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,3584,0.01928711103068458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,2560,0.03853955533769395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,16384,0.04053422146373325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,51200,0.05296355485916138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,4096,0.03988622294531928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,5120,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,128,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,12288,0.1091537806722853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,6144,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,51200,0.15697956085205078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,512,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,256,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,16384,0.038318223423428006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,3072,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,10240,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,2560,0.022379555635982092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,1536,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,3072,0.027765333652496338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,1024,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,3584,4.5717998080783415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,51200,0.21690221627553305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,5120,0.01607644392384423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,64,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,65536,0.37160621749030215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,3072,0.008241777618726095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,12288,0.20376266373528373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,6144,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,4096,0.18367200427585176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,32,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,32,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,6144,0.06284088558620878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,4096,0.0162791113058726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,3584,0.02440799938307868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,12288,0.03774133324623108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,12288,0.01255022237698237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,8192,0.03940533267127143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,768,0.06715199682447645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,128,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,10240,0.01643466618325975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,1536,0.026509332987997267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,10240,0.21197688579559326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,768,0.0833395587073432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,512,0.01369244439734353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,768,0.06512266397476196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,512,0.022580444812774658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,512,0.002591111179855135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,2048,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,12288,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,5120,0.00885511106914944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,512,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,1536,0.006753777878151999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,1024,0.06185511085722181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,3072,0.007733333441946242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,12288,0.018331555856598746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,1024,0.03373422225316366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,256,0.035531553957197405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,2560,0.01891999940077464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,64,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,10240,0.03905066516664293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,1536,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,1024,0.03569333420859443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,5120,0.01145600030819575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,12288,0.2555146747165256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,12288,0.02401155564520094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,4096,0.04139288928773668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,10240,0.061392002635531955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,64,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,8192,0.015726221932305228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,3584,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,5120,0.030319998661677044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,12288,0.08826844559775458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,1536,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,1024,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,1536,0.04607111215591431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,128,0.03016622198952569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,1536,0.003657777690225177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,3072,0.028208000792397395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,10240,0.026144888665941026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,7168,0.03463110989994473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,32,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,5120,0.022602667411168415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,3584,0.007788444558779399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,3584,0.015182221929232279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,256,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,2048,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,6144,0.0537431107627021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,2560,2.561424891153971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,2560,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,6144,0.018271999226676095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,7168,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,8192,0.07310577895906237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,4096,0.043226665920681424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,256,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,256,0.01514933341079288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,32,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,6144,0.0196115556690428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,3584,0.010056888891590966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,12288,0.1111217737197876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,256,0.02203911046187083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,7168,0.021536888347731695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,8192,0.015229332778188916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,3584,0.004613333278232151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,8192,0.014888889259762235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,64,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,8192,0.014862222803963555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,4096,0.010460444622569613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,65536,2.705351935492622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,1024,0.009759999811649323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,4096,0.06174311372968885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,5120,0.03606577714284261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,2560,0.007400888535711501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,1536,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,1536,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,7168,0.09740089045630561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,4096,0.03280355532964071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,12288,0.013743110828929476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,1024,0.00544799988468488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,1536,0.060214221477508545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,6144,0.006506666541099548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,1024,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,4096,0.010647999743620554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,2560,0.04126844472355313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,2560,0.038647111919191145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,32,0.0038284444146686127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,2048,0.011718221836619906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,10240,0.9406479729546441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,5120,0.03320711188846164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,256,0.007360888852013483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,10240,0.043566223647859365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,5120,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,7168,0.01313066648112403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,256,0.0637520021862454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,2560,0.058450665738847524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,5120,0.016199999385409884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,32,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,6144,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,6144,0.10916355583402847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,5120,0.05382400088840061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,768,0.01552977826860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,12288,0.034950223233964704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,5120,0.13234221935272217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,3584,0.05151999990145365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,4096,0.03740888833999634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,5120,0.07884622282452054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,65536,0.4039955668979221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,10240,0.17873333560095894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,2048,0.01984444426165687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,7168,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,7168,0.09787644280327691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,128,0.0173999998304579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,768,0.017682666579882305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,4096,0.027847111225128174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,256,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,2048,0.020495999190542433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,2048,0.006837333242098491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,7168,0.038763556215498186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,16384,0.08508977625105117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,4096,0.008020444048775567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,10240,0.04734844300482008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,6144,0.09857244624031915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,12288,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,8192,0.005413333161009683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,4096,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,12288,4.170160081651476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,128,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,3072,0.0127368892232577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,5120,0.08008444309234619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,256,0.022991110881169636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,8192,0.018254222141371835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,5120,0.024714666936132643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,3072,0.01756444407833947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,1024,0.004543111142185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,7168,0.35629600948757595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,32,0.01643733349111345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,7168,0.018435556027624343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,65536,2.5268383026123047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,10240,0.05207555492719015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,6144,0.0069955554273393415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,6144,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,6144,0.04538222153981527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,2048,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,8192,0.02825066778394911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,768,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,6144,0.020428443948427837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,1024,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,2048,0.02812622321976556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,3072,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,51200,0.05792355537414551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,16384,0.43188977241516113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,65536,0.18058044380611846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,4096,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,65536,0.06330577532450359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,12288,0.15880444314744738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,16384,0.04403555393218994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,12288,0.28155467245313853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,10240,0.023510222633679707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,8192,0.10073244571685791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,768,0.01586666703224182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,12288,0.013517333401574029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,2048,0.026826666461096868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,5120,0.012947555217478009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,3584,0.018811555372344125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,12288,0.13410045040978327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,3072,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,5120,0.02528444429238637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,5120,0.006305777778228124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,12288,0.26382221115960014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,7168,0.020630222227838304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,3072,0.05445333321889242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,4096,0.020170667105250888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,32,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,2560,0.2042133278316922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,4096,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,10240,0.037224001354641385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,10240,0.0787644452518887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,7168,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,512,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,4096,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,16384,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,10240,0.18158133824666342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,1024,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,256,0.0499946673711141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,128,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,64,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,8192,0.009444444543785518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,1536,0.02036977807680766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,32,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,128,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,128,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,3072,0.026367111338509455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,65536,0.17152799500359428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,12288,0.06601688596937391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,6144,0.016197333733240765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,256,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,12288,0.326911105049981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,65536,3.7881982591417103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,6144,0.00907466643386417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,51200,0.08489422003428142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,7168,0.10928177833557129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,12288,0.23295378684997559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,7168,0.01093777765830358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,768,0.03343910972277323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,12288,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,7168,0.042803555727005005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,16384,0.03412266572316488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,65536,0.16105688942803278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,65536,0.07277244329452515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,4096,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,128,0.012381333443853589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,3584,0.028443555037180584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,2560,0.5131413141886393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,8192,0.024059555596775476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,4096,0.01625066664483812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,12288,4.779927147759332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,6144,0.10204799969991048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,256,0.0169831116994222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,2048,0.23044888178507486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,768,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,7168,0.04135466615358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,512,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,2048,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,3584,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,5120,0.008719999757077958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,32,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,4096,0.012332444389661154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,7168,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,6144,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,512,0.01788444485929277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,65536,0.012723555167516073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,32,0.0019253332995706135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,7168,0.059134225050608315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,3072,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,64,0.02145599987771776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,51200,0.7049386766221789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,2560,0.015296889675988091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,10240,0.12477421760559082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,10240,0.07497511307398479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,65536,0.1748595502641466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,10240,0.1400960021548801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,12288,0.18950755066341826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,16384,7.62374284532335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,3072,0.054879112376107105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,2560,0.012357333468066322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,16384,0.00928177767329746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,128,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,16384,0.02165599995189243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,12288,0.052336888180838696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,16384,0.1000444425476922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,65536,0.09084800216886733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,5120,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,12288,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,32,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,512,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,3072,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,5120,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,3072,0.029057777590221826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,51200,0.07279377513461642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,12288,0.036935110886891685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,32,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,4096,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,16384,0.011767999993430244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,32,0.008919110728634728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,16384,0.041963554090923734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,7168,0.1742088927163018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,16384,0.018082666728231642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,65536,0.15540888574388292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,2048,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,2048,0.04254400067859226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,12288,0.04952088991800944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,12288,0.055883553293016225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,16384,0.19099199771881104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,2560,0.5875288645426432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,51200,0.1569004456202189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,1536,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,6144,0.029983109898037378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,65536,0.058881779511769615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,2560,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,128,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,16384,0.023911111884646948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,3584,0.051656001143985324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,32,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,128,0.04492622282769945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,3584,0.02186222208870782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,2048,0.011094222466150919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,5120,0.10502311256196763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,4096,0.029044446018007066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,1024,0.02231377859910329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,51200,0.006791111081838608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,51200,0.12206489510006374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,2048,0.01178311142656538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,7168,0.12888977262708876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,10240,0.033610665135913424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,128,0.006728888799746831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,512,0.09367999765608047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,128,0.01514933341079288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,2048,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,10240,0.1333973275290595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,128,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,128,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,1024,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,5120,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,7168,0.08508444494671291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,2560,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,768,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,65536,0.07221066951751709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,2048,0.11062933339012994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,1024,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,6144,0.476668463812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,1536,0.0381279985109965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,3584,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,5120,0.09473333093855117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,2048,0.09237511290444268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,65536,0.020231111182106864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,3072,0.028596444262398615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,5120,0.12597155570983887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,3584,0.05528622203403049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,51200,0.09888444344202678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,8192,0.041343109475241765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,65536,1.1170667012532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,2560,0.007334222396214803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,128,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,10240,0.04826844400829739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,6144,0.027151111099455092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,51200,0.44582578870985246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,5120,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,2560,0.09534133142895168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,3584,0.07693599992328219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,51200,0.06583999925189547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,2048,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,3584,0.007903999752468532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,7168,0.024624889095624287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,51200,0.04653244548373752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,1536,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,512,0.013437333206335703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,8192,0.012344888514942594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,4096,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,7168,0.03872444563441806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,16384,0.022164444128672283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,512,0.006068444500366847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,512,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,2048,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,10240,0.017653332816229928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,3072,0.09756266408496433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,51200,0.12592089176177979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,32,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,3584,0.020247111717859905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,65536,0.24338934156629774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,8192,0.20513421959347197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,32,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,65536,0.2729449007246229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,51200,0.7287022272745768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,16384,0.0322382218307919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,8192,0.06889777713351779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,10240,0.1784364382425944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,10240,0.32747554779052734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,2048,0.028516444895002577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,4096,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,5120,0.06547911299599542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,1536,0.021003555920388963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,8192,0.1402533319261339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,3072,0.5235946443345811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,512,0.028412444723976985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,3584,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,2048,0.019997333486874897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,1024,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,768,0.0987502204047309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,65536,0.3760995599958632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,768,0.017356443736288283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,51200,0.519136905670166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,16384,0.06492888927459717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,65536,0.07400621970494588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,3584,0.030057777961095173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,16384,0.03353955679469638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,1024,0.01314399970902337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,768,0.04578488734033373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,65536,1.008991135491265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,10240,0.016920889417330425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,12288,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,65536,0.27817511558532715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,4096,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,768,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,768,0.004548444516128964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,51200,0.03401066528426276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,512,0.0469688872496287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,4096,0.0329440004295773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,128,0.0033760000434186724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,768,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,3072,0.5480462180243598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,5120,0.02753244505988227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,65536,0.048804445399178394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,768,0.029612445169025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,3584,0.02180622186925676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,16384,0.35880443784925675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,8192,0.038753777742385864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,128,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,4096,0.030327998929553564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,8192,0.012172444827026792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,3072,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,12288,3.5254444546169705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,8192,0.042837331692377724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,768,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,12288,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,6144,0.4183982213338216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,512,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,5120,0.041435556279288396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,4096,0.02553422252337138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,32,0.002238222294383579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,5120,0.00925511121749878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,16384,0.05832622448603312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,16384,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,3584,0.023929778072569106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,65536,0.05487644341256884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,1024,0.0041333333485656315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,8192,0.23795999421013725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,12288,0.10140266683366563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,64,0.0747706691424052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,2560,1.9240409003363714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,3584,0.013747555514176687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,2560,0.4490044381883409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,4096,0.02238222294383579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,7168,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,7168,0.05860444572236803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,128,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,1024,0.016338666280110676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,16384,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,16384,0.04959377646446228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,128,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,3584,0.0742115577061971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,65536,0.24620887968275282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,3584,0.02712533374627431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,16384,0.2938408851623535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,12288,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,4096,0.020863110820452373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,32,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,65536,0.08671289020114475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,16384,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,768,0.038726223839653864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,5120,0.09080177545547485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,128,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,2048,0.026012443833880957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,12288,0.018630221486091614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,512,0.014073777529928418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,768,0.0240328891409768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,16384,0.058522668149736196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,128,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,7168,0.01983644399378035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,51200,0.05361422234111362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,8192,0.046612444851133555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,4096,0.05761155817243788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,5120,0.020267556111017864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,1024,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,12288,0.050269332196977407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,7168,0.045569777488708496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,16384,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,10240,0.7367431322733561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,2560,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,1536,0.021039111746682063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,768,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,1024,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,256,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,4096,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,12288,0.019689778486887615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,16384,0.45461686452229816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,1536,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,3584,0.051649779081344604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,3072,0.005680888891220093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,12288,0.007556444240940942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,2560,0.0685102211104499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,6144,0.010645333263609143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,1536,0.008023999631404877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,512,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,16384,0.19444621933831108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,256,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,3584,0.04235288831922743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,51200,0.01940266622437371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,4096,0.08389777607387966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,1536,0.028382221857706707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,2048,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,3584,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,5120,0.008592888712882996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,4096,0.04409689042303297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,6144,0.010315555665228102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,12288,0.2917484442392985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,3072,0.1632968849605984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,65536,0.06493955850601196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,2560,0.020982222424613103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,51200,0.11614933278825547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,3584,0.035066667530271746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,8192,0.022276444567574397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,16384,0.10624444484710693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,128,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,1536,0.010296000374688042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,8192,0.025608888930744592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,2048,0.015106666419241162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,768,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,3072,0.037681778271993004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,6144,0.021652445197105408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,2560,0.008751110898123847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,256,0.018868444694413077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,12288,0.01677688956260681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,65536,0.3985955450269911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,768,0.06742044289906819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,128,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,3072,0.020573332905769348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,4096,0.1215662161509196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,12288,0.01677688956260681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,128,0.017995556195576985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,51200,2.546023898654514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,65536,3.8842078314887156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,3072,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,2560,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,128,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,6144,0.009880888793203566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,10240,0.14613955550723606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,6144,0.009345778160625035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,16384,0.1682622167799208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,8192,0.04449333416091072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,3072,0.05523911118507385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,8192,0.04721600148412916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,768,0.002963555562827322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,5120,0.048807111051347524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,1024,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,2048,0.17387821939256456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,16384,0.042697777350743614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,6144,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,512,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,12288,0.03314933180809021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,1024,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,6144,0.01869777838389079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,8192,0.11583467324574788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,1536,0.03535288903448317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,4096,3.19234127468533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,1536,0.028182221783532038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,1024,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,5120,0.019479110836982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,512,0.020627554919984605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,32,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,3072,0.008803555534945594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,51200,0.02033066584004296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,512,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,768,0.005992888990375731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,768,0.01828622155719333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,8192,0.037852442926830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,6144,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,16384,0.09213599893781875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,128,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,2048,0.012539555629094442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,51200,0.05228888988494873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,7168,0.17382755544450548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,4096,0.05581955777274238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,12288,0.019567110472255282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,512,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,10240,0.0166666673289405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,128,0.010281778044170803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,6144,0.02686400049262577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,768,0.01683200067943997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,2048,0.018223111828168232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,3584,2.192168977525499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,16384,0.03131999903255039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,256,0.07081244389216106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,768,0.004595555365085602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,512,0.005325333525737126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,10240,0.0162764439980189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,256,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,256,0.01259200026591619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,10240,0.058924443191952176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,256,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,4096,0.008392888638708327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,12288,0.13317867120107016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,2560,0.008554666406578487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,256,0.027055111196306016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,2560,0.01628800067636702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,2560,0.0255422227912479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,2048,0.029115554359224107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,256,0.03236000074280633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,6144,0.43642576535542804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,7168,0.027678221464157104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,2560,0.015629332926538255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,5120,0.03072355522049798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,5120,0.03735911183887058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,3072,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,16384,0.08104177978303698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,32,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,3072,0.11006577809651692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,512,0.03397866752412584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,65536,0.09705777963002522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,2048,0.048531555467181735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,16384,0.14506488376193577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,1024,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,10240,0.08673866589864095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,3072,0.03132266799608866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,4096,0.08043111032909818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,65536,0.0589688883887397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,1536,0.025800888737042744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,512,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,128,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,4096,0.014509333504570855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,3584,0.023899555206298828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,12288,0.02955555584695604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,256,0.0026986667265494666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,128,0.010404444403118556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,64,0.005396444350481033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,10240,0.04318133327696058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,768,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,64,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,1024,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,2560,0.017867555220921833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,3072,0.017200888858901132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,16384,0.024452444579866197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,51200,0.10375289122263591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,1536,0.04126755396525065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,10240,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,64,0.11334755685594346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,1024,0.19436800479888916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,10240,0.026663111315833196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,2560,0.0074622225430276655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,32,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,768,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,2560,0.008759111166000366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,7168,0.1355111069149441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,5120,0.10769155290391709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,3072,0.463880909813775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,51200,0.1400888893339369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,768,0.009721777505344814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,128,0.001999999913904402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,6144,0.05195555422041151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,16384,0.04937600096066793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,256,0.0023502223193645477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,8192,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,3072,0.005416000054942236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,7168,0.028637333048714533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,10240,0.2786497804853651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,7168,0.02403822210099962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,7168,0.023895111348893907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,7168,0.04277777671813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,3584,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,10240,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,7168,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,128,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,2560,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,8192,0.601618660820855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,3584,0.009550222092204625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,5120,0.07947288619147407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,128,0.01825777855184343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,16384,0.02458488941192627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,2560,0.010670222342014313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,65536,0.32108977105882436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,3072,0.03631733192337884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,12288,0.050033777952194214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,6144,0.11162132687038845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,8192,0.16308977868821886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,3072,0.008554666406578487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,51200,0.06036533249749077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,12288,0.036002665758132935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,4096,0.3792888853285048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,4096,0.03863199883037143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,3072,0.23111377822028267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,51200,0.19078132841322157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,10240,0.0428435570663876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,768,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,4096,0.09571199946933323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,256,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,2560,0.028415110376146104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,6144,0.40225866105821395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,4096,0.009872888525327047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,512,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,512,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,5120,0.02833777666091919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,3072,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,51200,0.14556888739267984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,2560,0.04532355401251051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,1024,0.07503200239605375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,2560,0.08652355273564656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,2560,0.020256888535287645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,51200,5.151759253607856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,12288,0.035754667388068304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,51200,0.047213332520590894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,32,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,256,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,1536,0.03650577863057455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,6144,0.05014666583802965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,3072,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,2560,0.013380444712109037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,6144,0.0890257755915324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,3584,0.015847999188635085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,7168,0.034272889296213783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,256,0.012429333395428128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,64,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,7168,0.09283822112613255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,32,0.059002664354112416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,2048,0.008088889221350351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,8192,0.01963911122745938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,6144,0.02015822297996945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,1024,0.04924622178077698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,2560,0.11249600516425239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,2560,0.023707555400000677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,32,0.014961777461899651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,128,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,10240,0.02570488883389367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,2560,0.04790755444102817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,10240,0.09020000033908421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,16384,0.3311759895748562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,5120,0.0949395563867357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,51200,0.3643475638495551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,12288,11.86842261420356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,2560,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,64,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,768,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,3072,0.05937688880496555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,1536,0.023375999596383836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,512,0.015463110473420886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,6144,0.14961866537729898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,2048,0.007426666716734569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,16384,0.5894453260633681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,5120,0.029899554120169744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,1536,0.008787555827034844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,51200,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,2560,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,256,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,512,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,65536,0.17460178004370797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,65536,6.069917466905381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,2048,0.006874666859706243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,512,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,7168,0.26062578625149196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,10240,0.02908710969818963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,7168,0.03647733396954007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,1536,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,128,0.021282666259341772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,3584,0.02678933408525255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,1024,0.004553777890072929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,32,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,128,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,12288,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,3072,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,5120,0.03287200133005778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,512,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,2048,0.035417778624428645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,32,0.003165333428316646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,16384,0.04225333200560676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,3072,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,2048,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,2560,0.05013066530227661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,1024,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,3584,0.030084444416893855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,10240,0.025974222355418738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,1536,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,12288,0.028185778194003638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,768,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,4096,0.014933332800865173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,65536,2.6033058166503906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,8192,0.13026489151848689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,1536,0.02808711263868544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,2560,0.004600888739029567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,2048,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,2048,0.027969777584075928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,51200,0.2675946553548177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,4096,0.008648888932334052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,768,0.008649778034951951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,12288,0.04283377859327528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,2048,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,5120,0.026791999737421673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,768,0.021020443903075323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,2560,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,51200,0.500732421875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,2048,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,10240,0.037047111325793795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,256,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,2048,0.026039999392297532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,12288,0.049139555957582265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,16384,0.08663911289638943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,16384,0.5348221990797255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,256,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,3072,0.013047999805874295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,3584,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,10240,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,65536,0.20032444265153673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,512,0.007076444725195567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,768,0.007009777757856581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,51200,0.131315549214681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,12288,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,8192,0.04152889053026835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,5120,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,5120,0.047613332668940224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,256,0.00906755526860555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,4096,0.013212444053755866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,2560,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,16384,0.04426755507787069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,512,0.02123466630776723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,256,0.01309866706530253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,16384,0.04195822278658549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,12288,0.10123378038406372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,32,0.006071110980378256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,16384,0.06913244724273682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,51200,0.006714666469229593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,5120,0.12389866511027019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,51200,0.026891556051042344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,51200,0.32944178581237793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,10240,0.032024890184402466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,51200,0.08216977781719632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,3072,0.08168444368574354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,64,0.0018826667219400406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,6144,0.16917867130703398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,8192,0.04792444573508369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,1536,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,7168,0.009480000370078618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,5120,0.008651555412345463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,5120,0.04059999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,16384,0.01755822201569875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,8192,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,768,0.068904889954461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,7168,0.050272001160515674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,256,0.013403555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,51200,10.043870713975695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,512,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,5120,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,5120,0.030601776308483545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,16384,0.021947556071811255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,2560,0.01976088848378923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,10240,0.0490551127327813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,12288,0.11903999911414252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,6144,0.028613332245084975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,16384,0.5599457952711318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,65536,0.1466239955690172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,2048,0.006664888726340399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,7168,0.16902132829030356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,65536,2.2425431145562067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,2048,0.02158755560715993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,10240,0.027951998843087092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,5120,0.025472889343897503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,2560,0.00720088928937912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,768,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,8192,0.02276800076166789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,10240,0.019479110836982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,1536,0.06879644261466132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,4096,0.007920000288221572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,10240,0.034196443027920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,4096,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,16384,0.022643556197484333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,256,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,16384,0.021242666575643752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,64,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,8192,0.011384888655609555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,7168,0.10478666755888198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,32,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,4096,0.007910221815109253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,512,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,10240,0.021128889587190416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,1024,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,16384,0.02786488996611701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,768,0.04170044594340854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,1024,0.022423999177085027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,1536,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,128,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,128,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,128,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,256,0.015115555789735583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,2560,0.11915555265214707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,3072,0.01309866706530253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,51200,0.10562311278449164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,12288,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,10240,0.08264711168077257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,128,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,16384,0.07580533292558458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,3584,0.030489779180950586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,512,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,1024,0.1745955546696981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,65536,0.42780089378356934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,5120,0.026690666874249775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,65536,0.1315982209311591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,2560,0.042280889219707914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,4096,0.013408888545301227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,1024,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,7168,0.042133334610197276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,768,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,2560,0.4436951213412815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,51200,3.5295689900716147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,10240,0.1948444445927938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,1024,0.006384000182151794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,7168,0.032151109642452665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,8192,0.09579200214809841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,6144,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,4096,0.04042311178313361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,6144,0.010798222488827176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,65536,0.12259022394816081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,10240,0.04420800010363261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,7168,0.023558222585254248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,16384,0.03704977697796292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,51200,0.10838399993048774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,256,0.07312266694174872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,2560,0.0814631117714776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,12288,0.22511288854810926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,16384,0.04892622099982368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,16384,0.26190490192837185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,3072,0.021330666210916307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,6144,0.08653244707319473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,10240,0.09211377965079413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,12288,0.15363111760881212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,256,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,10240,0.06979644298553467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,2560,0.023153778579499986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,8192,0.1253546608818902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,5120,0.053399999936421715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,7168,0.09318577580981785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,16384,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,12288,0.021908443835046556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,1024,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,3072,0.030975110001034204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,7168,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,12288,0.13188800546858045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,8192,0.01071466671095954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,4096,0.00835111074977451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,768,0.005671999934646819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,2048,0.03801155421468947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,1024,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,5120,0.00980799976322386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,8192,0.013953778478834363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,64,0.01696622206105126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,4096,0.2986782126956516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,12288,1.5716293123033311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,10240,0.08151288827260335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,1536,0.010647999743620554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,1536,0.022314666046036616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,3584,0.029896888468000624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,3584,0.014019555515713163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,256,0.049866666396458946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,2560,0.06875289148754544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,12288,0.0380639996793535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,32,0.18332978089650473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,6144,0.05829866727193197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,4096,0.009947555760542551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,2048,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,256,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,2560,0.034698665142059326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,64,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,7168,0.0088764445649253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,3584,0.007713778151406183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,256,0.05594578054216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,128,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,8192,0.07891644371880426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,12288,0.1622862286037869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,1024,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,768,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,16384,0.2167395618226793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,1536,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,1536,0.024373332659403484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,3072,0.019944889677895438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,1024,0.0047155556579430895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,3072,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,2560,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,4096,0.08513689041137695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,8192,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,3584,0.00758133331934611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,5120,0.048193779256608754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,256,0.016083555089102853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,4096,0.06055911382039388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,256,0.09488888581593831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,4096,0.03426488902833726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,256,0.019877332780096266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,65536,0.525677310095893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,1536,0.015453333655993143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,4096,0.03689333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,2048,0.006738666858938005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,256,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,10240,0.02104622291194068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,1024,0.0092275556590822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,7168,0.06231821907891167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,3072,0.007489778101444244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,51200,0.22728265656365287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,65536,0.43471378750271267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,51200,0.054252445697784424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,128,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,1024,0.017878222796652053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,1024,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,12288,2.5332747565375433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,768,0.02606933315594991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,2048,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,16384,0.02070844504568312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,6144,0.05468533436457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,512,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,3072,0.012695111334323883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,128,0.037074665228525795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,12288,0.1026044421725803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,16384,0.021874666213989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,1536,0.017675555414623685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,1024,0.016710221767425537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,1536,0.019623999794324238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,4096,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,12288,0.07651822434531318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,65536,0.07215555508931477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,32,0.011352000137170156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,12288,0.09443644682566325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,4096,0.02809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,3072,0.1050257749027676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,6144,0.03889066643185086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,51200,17.606313069661457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,128,0.005402666827042897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,1024,0.015811555915408664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,65536,0.08908889028761123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,7168,0.021720000439220007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,6144,0.07688089211781819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,128,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,1536,0.04211289021703932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,7168,0.01424088908566369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,2560,0.015106666419241162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,65536,13.335491604275175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,65536,0.3158160050710042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,768,0.00583022211988767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,10240,0.03382222188843621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,3584,0.021409778131379023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,7168,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,1536,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,128,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,128,0.005304000029961268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,2560,0.014174222946166992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,2560,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,2560,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,12288,0.04230933388074239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,256,0.01612444387541877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,10240,0.03343910972277323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,6144,0.1253404484854804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,12288,0.05279555585649279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,12288,0.6873955726623535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,8192,0.09014755487442017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,6144,0.019254222512245178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,12288,0.015413333972295126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,256,0.015560888581805758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,768,0.020805334051450092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,12288,0.03051111102104187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,512,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,16384,0.04407199886110094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,2560,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,65536,0.09905422396130031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,1024,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,10240,0.14264355765448675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,10240,0.038140445947647095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,128,0.01519377695189582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,12288,0.07406400309668647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,12288,0.6034933196173774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,2048,0.2744622230529785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,8192,0.015020444989204407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,5120,0.012690666649076672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,4096,0.5864248805575901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,5120,0.0119982221060329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,8192,0.04718044400215149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,5120,0.0719555550151401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,3072,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,8192,2.3391990661621094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,2560,0.03086400032043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,51200,0.6065742174784342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,51200,0.22431556383768717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,256,0.20954577128092447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,5120,0.47015020582411027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,10240,0.02511288887924618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,6144,0.14075911045074463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,51200,0.18982666068606904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,512,0.015549333559142219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,10240,0.07598400115966797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,65536,2.0051903194851346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,3072,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,6144,0.06160622172885471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,10240,0.028816001282797918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,6144,0.05530222256978353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,512,0.00572533326016532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,1024,0.008452444440788692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,8192,0.016511110795868766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,2048,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,768,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,3072,0.026127111580636766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,1024,0.0028853333658642242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,65536,0.24791555934482148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,256,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,5120,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,16384,0.04187555445565117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,2560,0.0344124436378479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,3584,0.0158942225906584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,768,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,5120,0.908375104268392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,1024,0.07861155271530151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,768,0.020100444555282593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,16384,0.04541066620084974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,51200,0.07360533210966322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,3584,0.06308622492684259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,512,0.01960088809331258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,3584,0.029431111282772485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,6144,0.06339466571807861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,16384,0.0258577780591117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,51200,0.05480800072352091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,64,0.0069013333155049225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,3584,0.023336889015303716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,12288,0.18913155131869844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,16384,0.13500444094340006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,10240,0.032857778999540545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,65536,0.629017776913113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,12288,0.5997609032524956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,256,0.007872000336647034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,65536,0.06650844547483656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,2048,0.015473778049151102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,1536,0.023880889018376667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,12288,0.015751111838552687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,64,0.010382222632567087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,8192,0.011078221930397881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,51200,0.07037333647410075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,2048,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,128,0.022262222237057153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,512,0.0222382214334276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,768,0.03014755580160353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,12288,0.12781422668033177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,32,0.016759999924235873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,3584,0.0188044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,12288,0.030096001095241968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,7168,0.20499555269877115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,4096,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,5120,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,1536,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,6144,0.02984444300333659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,8192,0.07745777898364596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,32,0.016723554995324876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,51200,0.26550043953789604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,128,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,512,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,6144,0.11814488304985894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,768,0.01885955532391866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,128,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,12288,0.02790488799413045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,512,0.005487999982304043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,2560,0.05974666939841377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,51200,0.02074399921629164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,5120,0.039320889446470476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,2560,0.02645066711637709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,65536,8.707901848687067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,10240,0.11130311754014756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,3584,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,7168,0.14330489105648464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,3072,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,4096,0.06276355849372016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,12288,0.012357333468066322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,3072,0.1034604443444146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,7168,0.004944889081848992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,16384,0.06630755795372857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,7168,0.010226666927337646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,2560,0.015432889262835184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,2560,0.030960887670516968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,2048,0.012768000364303589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,4096,0.026721777187453374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,5120,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,768,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,12288,0.040475553936428495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,256,0.020097777247428894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,128,0.0031884445084465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,16384,0.20691022608015272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,16384,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,3584,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,768,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,8192,0.021126222279336717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,51200,0.1200595564312405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,4096,0.008396444221337637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,8192,0.051474667257732816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,3584,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,1536,0.02550044490231408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,4096,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,6144,0.047704888714684375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,5120,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,4096,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,1024,0.018742221924993727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,6144,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,128,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,4096,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,10240,0.03824088970820109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,2048,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,1536,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,7168,0.020311110549502902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,3584,0.09982933600743611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,5120,0.01702666613790724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,32,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,512,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,3072,0.19473332828945586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,65536,5.395292494032119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,8192,0.023569777607917786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,6144,0.013502221968438892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,4096,0.010822222464614444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,128,0.02055022286044227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,2560,0.013050666285885705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,512,0.014714666538768344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,6144,0.0199297782447603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,16384,0.24409778912862143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,16384,0.3499608834584554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,10240,0.017549332645204332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,65536,0.062310225433773465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,16384,0.053567998939090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,3072,0.08739466799630059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,256,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,7168,1.3259876039293077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,7168,0.013286222186353473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,51200,12.80708482530382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,2560,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,65536,0.20314844449361166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,2048,0.17446045080820718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,32,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,7168,0.021065778202480737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,32,0.03561155663596259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,2048,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,10240,0.4171582327948676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,6144,0.22321777873569063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,32,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,3072,0.031120889716678198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,51200,0.28862934642367893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,2048,0.04309155543645223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,7168,0.0424924459722307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,4096,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,5120,0.012108444339699216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,12288,0.020564445190959506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,10240,0.022266666094462078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,8192,0.057935112052493624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,5120,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,3072,0.008381333616044786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,7168,0.010619555910428366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,128,0.010279999838935005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,128,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,10240,0.038324445486068726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,128,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,2048,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,7168,0.009945777555306753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,3072,0.01716088917520311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,8192,0.1385235521528456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,16384,0.03808710972468058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,2560,0.03815644317203098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,2560,0.01293155550956726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,2048,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,8192,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,3072,0.03590044379234314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,8192,0.042760888735453285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,2048,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,1536,0.030231999026404485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,3584,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,2560,0.05243555704752604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,12288,0.13389510578579372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,12288,0.04954666561550564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,10240,0.08891110950046116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,512,0.00565244464410676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,512,0.0840666691462199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,6144,0.011749332977665795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,5120,0.011657777759763928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,3072,0.029099557134840224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,2048,0.02427111069361369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,16384,0.12919999493492976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,512,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,2048,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,4096,0.030478222502602473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,128,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,128,0.0669599970181783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,10240,0.03615644574165344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,6144,0.11317866378360325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,5120,0.019270223047998216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,768,0.023804444405767653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,3584,0.06332177586025663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,128,0.01482311056719886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,3584,0.2318720022837321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,1536,0.022953778505325317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,10240,0.01705511079894172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,1536,0.012663111090660095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,12288,0.47185956107245547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,4096,0.01759999990463257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,1536,0.028528889020284016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,10240,0.030778666337331135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,7168,0.4206311172909207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,7168,2.187493430243598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,2560,0.035655998521380954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,10240,0.03291022115283542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,2048,0.037235554721620344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,2048,0.013004444539546967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,8192,0.012049777640236748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,3584,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,1536,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,64,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,1536,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,10240,0.034318221939934626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,65536,0.14794311258527967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,4096,0.024345777100986902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,1024,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,12288,0.08632977803548177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,5120,0.021212443709373474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,64,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,6144,0.055642664432525635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,2048,0.03525955478350321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,8192,1.0031635496351454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,8192,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,4096,0.09113688601387872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,1024,0.00925866680012809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,5120,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,16384,0.01593244406912062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,1536,0.03425777620739407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,1536,0.010991999672518836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,2560,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,3072,0.029712888929578993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,10240,0.05150844322310554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,51200,5.99211925930447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,3072,0.05465333329306709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,3072,0.019896888070636325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,64,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,1024,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,4096,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,3072,0.030219554901123047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,51200,0.9258658091227213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,6144,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,10240,0.020522667302025687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,12288,0.20341689056820342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,12288,0.12523378266228571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,768,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,5120,0.10724622011184692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,4096,0.03206933206982083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,16384,0.1428968906402588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,5120,0.044863111442989774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,768,0.012883555557992725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,3584,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,32,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,65536,0.16688888602786592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,12288,0.0662720004717509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,2048,0.022292445103327434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,10240,0.011958222422334882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,3072,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,6144,0.03403822249836392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,8192,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,768,0.005727111051479976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,4096,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,1536,0.0063368889192740125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,65536,0.43765778011745876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,512,0.03151199883884854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,12288,0.021328889661365088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,768,0.16658133930630153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,512,0.06596977843178643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,128,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,51200,0.4563280211554633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,6144,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,2560,0.023391111029518977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,5120,0.010656889114114972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,4096,0.009524444739023844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,65536,0.26996533075968426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,256,0.047694222794638745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,256,0.23579022619459364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,768,0.027705777022573683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,5120,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,5120,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,51200,0.6076951026916504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,6144,0.10550310876634385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,1536,0.06366399923960368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,2560,0.02328799996111128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,10240,0.023668444818920557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,2048,0.010867555936177572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,3584,0.2026968929502699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,6144,0.009730666875839233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,6144,0.00903911143541336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,1536,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,2048,1.0291164186265733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,6144,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,32,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,6144,0.01293155550956726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,1024,0.0032151111712058387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,2560,0.01366666704416275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,65536,0.06829244560665555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,512,0.08920088741514419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,2048,0.07429422272576226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,32,0.007357333269384172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,1536,0.0354364448123508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,51200,19.60162523057726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,65536,0.06951555278566149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,1536,0.020999110407299463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,2560,0.02943555514017741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,1536,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,3072,0.023411555422676936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,12288,0.03897866606712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,3584,0.01920355525281694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,4096,0.0177511109246148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,8192,0.027637332677841187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,10240,0.04363822274737888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,5120,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,2560,0.02420622275935279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,12288,0.10644888877868652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,51200,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,3072,0.09079911311467488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,2560,0.025791999366548326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,1536,0.4110186629825168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,6144,0.03229866756333245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,65536,0.08838044272528754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,512,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,4096,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,2048,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,65536,0.49187909232245547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,12288,0.09744089179568821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,32,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,10240,0.030405332644780476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,65536,0.09896710846159194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,5120,0.05365955498483446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,32,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,10240,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,128,0.002114666625857353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,16384,0.03156177865134345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,1536,0.006329777754015393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,8192,0.023504000571038988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,10240,0.015431111057599386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,1536,0.003921777837806278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,256,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,12288,0.11651110649108887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,7168,0.02055999967787001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,51200,0.23673423131306967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,1024,0.0496693319744534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,10240,0.011848889291286469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,32,0.06656266583336724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,2560,0.052116443713506065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,3584,0.00980177770058314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,512,0.011101333631409539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,6144,0.009720000127951304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,7168,0.8807964324951172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,3072,0.06851822137832642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,64,0.00351555562681622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,5120,0.10286400053236221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,10240,0.16906844245062935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,6144,0.169852442211575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,1536,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,1536,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,512,0.011672000090281168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,256,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,768,0.04876888791720072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,51200,0.06122044722239176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,10240,0.046916445096333824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,3584,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,16384,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,2048,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,5120,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,1536,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,16384,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,128,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,2560,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,512,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,768,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,65536,0.17393244637383354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,768,0.24027911822001138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,7168,0.11600711610582139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,256,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,51200,0.06222755379147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,5120,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,1024,0.009947555760542551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,1024,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,2560,0.26242756843566895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,10240,0.14104089472028944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,64,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,7168,0.6414462195502387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,3072,0.012628444366984896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,3584,0.022109333011839125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,512,0.05347555544641283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,2048,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,256,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,3072,2.2825963762071395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,10240,0.02512977851761712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,3584,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,128,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,3072,0.05520177880922953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,16384,0.0326551099618276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,7168,0.01124622258875105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,256,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,10240,0.005099555684460534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,512,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,2048,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,65536,0.12110843923356797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,768,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,65536,0.1103244423866272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,64,0.004201777693298128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,1536,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,10240,0.0125591109196345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,6144,0.19743733935885957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,5120,0.14904977215660944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,1024,0.002663111107216941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,128,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,2048,0.1030435562133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,51200,0.06204533576965332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,4096,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,6144,0.01849422189924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,6144,0.028142223755518597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,1536,0.009869332942697737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,10240,0.08915911118189494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,1024,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,6144,0.11060355769263373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,768,0.013030221892727746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,10240,0.02717422280046675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,4096,0.05712622404098511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,10240,0.023330666952663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,256,0.032927112446890935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,10240,0.1059946682718065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,128,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,1024,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,32,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,3072,0.09356888797548081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,5120,0.028145776854621038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,6144,0.02444177700413598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,768,0.027183110515276592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,768,0.029671112696329754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,1024,0.020611556039916146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,7168,0.07124533255894978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,1024,0.009760888914267222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,2048,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,256,0.008070222205585903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,3072,0.012801777985360889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,5120,0.03425155414475335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,10240,0.1029626660876804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,2560,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,4096,0.08793866634368896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,10240,0.022728888524903193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,12288,0.04850488901138306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,3072,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,16384,1.5440621905856664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,5120,0.017113778326246474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,3072,0.044030222627851695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,768,0.004969777746333016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,256,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,3584,0.07740888992945354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,3584,0.021307556165589228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,2560,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,51200,1.2750257915920682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,2560,0.021651556094487507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,1536,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,768,0.0058248887459437055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,12288,0.44750044080946183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,12288,0.02145066691769494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,1024,0.006175111151403851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,4096,0.05271200007862515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,10240,0.12158489227294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,3072,0.12334933545854355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,65536,0.41869510544670957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,65536,0.06333955791261461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,4096,0.0328453348742591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,6144,0.016740444633695815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,16384,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,12288,0.022384888596004907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,16384,0.015810666812790763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,32,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,3072,0.06365244256125556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,6144,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,64,0.0031235555393828284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,5120,0.10470133357577854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,16384,0.39199288686116535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,1024,0.015674667225943673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,7168,0.035135110219319664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,3072,0.01406133340464698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,65536,0.05708622270160251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,32,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,3072,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,512,0.03795199924045139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,64,0.06783644358317058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,7168,0.012289777398109436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,65536,13.489967346191406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,16384,0.2506462203131782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,2048,0.037725332710478045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,7168,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,3072,0.02995733420054118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,8192,0.04841511117087471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,64,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,4096,0.07233066691292657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,1024,0.018013333280881245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,7168,0.012882666455374824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,1536,0.04168000155025058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,7168,0.030304001437293157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,1024,0.14483377668592665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,1536,0.053370667828453906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,3584,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,6144,2.3127547370062933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,768,0.014742222097184924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,3072,0.019066666563351948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,65536,0.2737955517239041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,3072,0.012753778033786349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,51200,0.054281777805752225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,3584,0.007536888950400882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,6144,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,10240,2.0020133124457464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,10240,0.039585779110590615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,1536,0.0038782221575578055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,512,0.0032382222513357797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,5120,0.008886222210195329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,10240,0.032053334845436945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,5120,0.017522666189405654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,6144,0.06551822026570638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,1536,0.01850311126973894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,51200,0.015658666690190632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,7168,0.013747555514176687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,16384,0.372292439142863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,2048,0.020052444603708055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,51200,0.05069422059588962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,2560,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,1536,0.004247111164861255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,4096,0.07507644759284125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,768,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,256,0.005758222192525864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,2048,0.006324444380071428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,7168,0.05806044737497965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,51200,0.020624889267815482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,65536,0.09318933222028945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,10240,0.11149599817064072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,7168,0.13932622803582087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,2048,0.07223555776807997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,4096,0.023776888847351074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,1024,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,768,0.0030088888274298776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,12288,0.039741333987977766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,6144,0.05911199914084541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,1024,0.015477332803938123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,3072,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,2560,0.013194666968451606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,512,0.015656888484954834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,8192,0.16863466633690727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,1024,0.017714665995703805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,8192,0.050107555256949536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,7168,0.012253333297040729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,16384,0.1795626613828871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,128,0.014170666535695394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,2560,0.024671999944580927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,4096,0.040178666512171425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,3072,0.023731556203630235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,10240,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,2048,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,32,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,512,0.007000889215204451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,12288,0.15948622756534153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,1536,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,3072,0.04682666725582547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,65536,0.08044799831178454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,1536,0.024029332730505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,7168,0.007673777639865875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,6144,0.04921866787804497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,8192,14.382049560546875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,256,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,12288,0.04668622215588888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,6144,0.030917333232031927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,6144,0.08989510933558147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,12288,0.030756443738937378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,32,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,7168,0.024710221423043147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,10240,0.023542222049501207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,512,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,128,0.03477599885728624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,6144,0.013013333082199097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,1536,0.003903111235962974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,8192,0.040066666073269315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,65536,0.07678311400943331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,768,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,64,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,128,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,768,0.021700445148679946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,12288,0.08244533008999295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,64,0.02864711152182685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,16384,0.05852711200714111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,10240,0.016786666380034555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,1024,0.008206222620275285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,512,0.016542222764756944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,51200,5.705118391248916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,7168,0.18352888690100777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,16384,0.02148888839615716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,768,0.0034826666944556763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,3072,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,1536,0.019896888070636325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,2048,0.1248977714114719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,1536,0.01958666741847992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,2560,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,128,0.013366222381591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,2560,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,1024,0.016037333342764113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,7168,0.01496266656451755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,64,0.0014951111127932866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,768,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,7168,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,2560,0.10025066799587673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,3072,0.007575111256705389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,6144,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,1024,0.020053333706325956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,256,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,7168,0.01221333361334271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,512,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,3072,0.46947823630438906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,1024,0.04888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,8192,0.18791288799709746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,12288,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,5120,0.01012711144155926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,2560,0.026791110634803772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,2048,0.01495822270711263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,64,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,512,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,1536,0.01127466642194324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,768,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,1536,0.004168888760937585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,16384,2.94175296359592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,256,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,51200,1.2199768490261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,768,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,128,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,512,0.3525244394938151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,2048,0.013398222625255585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,8192,0.07241066959169176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,768,0.012778666284349231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,65536,0.8319466908772787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,12288,0.10726933346854316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,32,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,2048,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,32,0.014688000082969666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,512,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,4096,0.03228444523281521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,2560,0.012836444709036084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,256,0.009375111096435124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,128,0.019924443629052904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,768,0.013777777552604675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,256,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,5120,0.15602310498555502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,5120,0.0654639999071757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,768,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,768,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,12288,0.01956355571746826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,6144,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,7168,0.03194488750563727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,1024,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,12288,0.02366488840844896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,2560,0.029951999584833782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,4096,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,7168,0.01641688909795549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,1024,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,256,0.017619555195172627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,1536,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,51200,3.2113581763373484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,1536,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,1536,0.008854221966531541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,2048,0.028760890165964764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,512,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,1536,0.01959288948112064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,10240,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,3584,0.006783111227883234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,256,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,3584,0.05499822232458326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,128,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,768,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,768,0.018186666899257235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,64,0.012098666694429187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,1536,0.028348445892333984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,2560,0.06771022081375122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,12288,0.024996444582939148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,5120,0.004559999952713649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,3072,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,2048,0.16125067075093588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,64,0.038077334562937416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,2048,0.03202666507826911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,512,0.005159111072619756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,1536,0.006558222075303395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,256,0.02380888826317257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,3072,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,2048,0.08113333251741198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,1024,0.003165333428316646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,4096,0.03161155515246921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,10240,0.11573955747816299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,4096,0.017300445172521803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,65536,0.5837635464138454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,5120,0.162463108698527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,8192,0.10469422075483535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,7168,0.043671111265818276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,64,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,12288,0.05147733290990194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,6144,0.011027555498811932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,1024,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,4096,0.033823112646738686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,6144,0.011242667006121742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,10240,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,10240,0.07201155689027575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,7168,0.23332089847988555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,7168,0.036575112077924944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,10240,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,3584,0.02388800018363529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,12288,0.013180444637934366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,51200,0.03405422303411696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,12288,0.11179110738966201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,512,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,51200,0.13033955627017552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,1024,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,6144,0.07948178052902222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,8192,0.011639110743999481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,1536,0.023769777682092454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,512,0.014750222365061441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,512,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,2048,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,51200,0.15331199434068468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,1536,0.03239288926124573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,1024,0.030750221676296655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,1024,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,512,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,6144,0.014540443817774454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,65536,0.12207644515567356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,3584,0.14517333772447374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,10240,0.2892613410949707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,65536,0.2726159890492757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,16384,0.042021334171295166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,51200,0.0510408878326416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,10240,0.44950400458441836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,2560,0.0594959987534417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,3584,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,16384,0.03391822179158529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,256,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,8192,0.010451555252075195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,65536,0.1636479960547553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,2048,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,32,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,2560,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,7168,0.12306843863593207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,2560,0.015261333849694995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,2560,0.0698302189509074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,6144,0.8433529006110297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,65536,0.08214044570922852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,1024,0.06371021933025785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,512,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,16384,0.8954808976915148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,256,0.019133332702848647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,256,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,768,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,10240,0.019498666127522785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,256,0.03720266620318095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,1024,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,3072,0.15259110927581787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,8192,0.10523288779788548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,12288,0.021877333521842957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,5120,0.01462933255566491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,12288,0.019186667270130582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,128,0.05193688803248935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,7168,0.015344889627562629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,2560,0.008939555949634975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,256,0.005454222361246745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,128,0.04329244295756022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,5120,0.01389955480893453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,6144,0.012155555188655853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,256,0.040656887822681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,3584,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,16384,0.022421333524915908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,12288,0.01944977707333035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,2048,0.03131644593344794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,12288,0.08669955862892999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,2048,0.04119822382926941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,65536,0.8251111242506238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,3072,0.2845582167307536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,1536,0.020909332566791113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,10240,0.012453333371215396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,10240,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,256,0.01794488893614875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,2048,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,2560,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,768,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,256,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,768,0.024958221448792353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,3072,0.030960887670516968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,3072,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,1024,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,3072,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,768,0.017624000708262127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,64,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,12288,0.013594666288958656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,16384,0.049044443501366504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,256,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,64,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,2048,0.005126222140259213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,512,0.019331556227472093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,10240,0.017843556072976854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,2560,0.009054222040706212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,10240,0.1425244410832723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,2560,0.00905600024594201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,12288,0.051876445611317955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,65536,0.0903973314497206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,2048,0.01443911095460256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,10240,0.023676445086797077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,1536,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,8192,0.08683466911315918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,6144,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,4096,0.08790933423572117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,256,0.004996444616052839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,256,0.037119110425313316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,2560,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,128,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,2560,0.027133334014150832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,5120,0.010067555639478896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,12288,0.013976888524161445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,1024,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,16384,0.025898666845427618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,1536,0.027140445179409448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,3584,0.028027554353078205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,128,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,6144,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,5120,0.014872888724009195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,1536,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,768,0.042661332421832614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,1024,0.008728000024954477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,3072,0.025847999585999384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,7168,0.017087111870447796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,512,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,10240,0.03535999854405721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,10240,0.04562222295337253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,3072,0.008408889174461365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,128,0.0343777769141727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,65536,0.04952711198065016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,1024,0.1912764443291558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,4096,0.1421440045038859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,51200,0.794107543097602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,1536,0.6271111170450846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,512,0.01274400038851632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,6144,0.585860464307997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,32,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,256,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,1024,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,3072,0.025524444050259058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,512,0.007109333243634965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,1536,0.11237333880530463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,4096,0.025136889682875738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,8192,0.015041778484980265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,7168,0.023687111006842718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,10240,0.8495208952162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,65536,0.1852888928519355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,256,0.02363377809524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,256,0.1456408898035685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,3584,0.02507200009293026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,16384,0.2519520123799642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,1024,0.21626133388943145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,1536,0.41620177692837185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,2560,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,7168,0.24862578180101183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,16384,0.11404889159732395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,6144,3.5467635260687937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,128,0.017472000585661996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,65536,0.8487520217895508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,1024,0.017065778374671936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,51200,0.04929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,4096,0.008059555457697975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,32,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,128,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,768,0.008895999855465358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,51200,0.2364177703857422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,10240,0.09364889065424602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,1024,0.014175999495718213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,768,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,512,0.15088089307149252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,2048,0.0046924443708525766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,768,0.028036445379257202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,1536,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,10240,0.052341332038243614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,65536,0.3198915587531196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,51200,0.054442667298846774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,3584,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,256,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,4096,0.008105778031879002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,51200,10.603062099880642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,3072,0.06556533442603217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,64,0.003080000066094928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,2560,0.1831359995736016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,2560,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,768,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,64,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,16384,0.13830667071872288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,512,0.04633511106173197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,128,0.017636444833543565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,768,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,16384,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,768,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,65536,0.027190221680535212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,7168,0.022294221652878657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,2560,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,1024,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,12288,0.188756439420912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,128,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,7168,0.012146666646003723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,6144,0.021313778228229944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,10240,0.10018399688932632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,51200,0.06406133042441474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,2560,0.031201776531007554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,4096,0.015370666980743408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,6144,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,10240,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,768,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,64,0.0021031111892726687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,1024,0.043967998690075345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,12288,0.19122844272189668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,768,0.020720889170964558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,4096,0.11851733260684544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,2560,0.004921777794758479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,64,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,1024,0.014358222484588623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,512,0.01128355579243766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,8192,0.01107377807299296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,2048,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,1536,0.03672800130314297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,1536,0.015432000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,10240,0.1558471123377482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,4096,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,2560,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,1536,0.024898666474554274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,3072,0.027327999472618103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,51200,0.05794399976730347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,2048,0.05338577760590447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,3072,0.03622311022546556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,128,0.11137866973876953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,1024,0.022434666752815247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,12288,0.03013066781891717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,64,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,10240,0.13379289044274226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,3584,0.02677511175473531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,4096,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,5120,0.03127200073666043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,1024,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,256,0.03062044580777486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,10240,0.029127111037572224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,3584,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,7168,0.021205332544114854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,3072,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,768,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,2048,0.05250933435228136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,12288,0.6705830891927084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,768,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,65536,2.350834740532769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,1536,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,1536,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,3072,0.03809066613515218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,4096,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,5120,0.013391999734772576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,256,0.007242666350470648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,10240,0.0454346670044793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,10240,0.05782577726576063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,51200,0.06756266620424059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,2048,0.01751733322938283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,16384,0.03054311209254795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,32,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,10240,0.02914666798379686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,65536,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,51200,1.9772391849093969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,32,0.009568889107969072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,12288,0.028464890188641016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,512,0.03528533379236857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,7168,0.012656889028019376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,16384,0.14344888263278538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,2048,0.013061333033773633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,51200,1.018140474955241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,7168,0.08207555611928304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,768,0.008619555168681675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,65536,1.541585816277398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,3072,0.05189777745140923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,3584,0.16512621773613825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,12288,0.01828799976242913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,3584,0.010151999692122141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,128,0.005386666705211003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,10240,0.17546933227115205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,65536,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,65536,0.15512800216674805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,10240,5.171838972303602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,51200,0.06169422467549642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,51200,0.17156178421444365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,256,0.003624000069167879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,5120,0.013174222575293647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,3584,0.02972444560792711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,128,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,5120,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,2048,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,51200,0.05347111158900791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,64,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,12288,0.26275645362006295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,512,0.01734133395883772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,6144,0.00905955582857132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,2560,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,3584,0.020297777321603563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,51200,0.00720355576939053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,16384,0.08141333527035184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,4096,0.014323555760913424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,4096,0.026329777306980554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,2048,0.027843554814656574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,5120,0.022786666949590046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,65536,0.17001688480377197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,3072,0.024294222394625347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,5120,0.021299555897712708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,51200,0.07231289148330688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,3584,0.5138230853610569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,65536,0.49229155646430117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,3072,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,51200,2.6394596099853516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,768,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,256,0.007220444579919179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,51200,0.27081510755750865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,12288,0.025761778155962627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,128,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,768,0.024336889386177063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,8192,0.013222222526868185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,512,0.002921777880854077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,8192,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,16384,0.8701128959655762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,1536,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,51200,0.10951111051771377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,3072,0.029937777254316542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,5120,0.09538933303621079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,65536,0.05659377574920654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,128,0.0019351111518012153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,512,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,51200,0.5304444630940756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,4096,0.7679182158576118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,512,0.1719191074371338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,256,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,64,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,65536,0.04972266819741991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,64,0.00778311077091429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,10240,0.019765333996878732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,51200,0.04667644368277656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,2560,0.007156444092591603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,65536,0.0710373322168986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,4096,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,5120,0.01959911154376136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,12288,0.06462399827109443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,1536,0.009588444398509132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,3584,0.03764177693261041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,4096,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,256,0.005392000079154968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,512,0.04014133413632711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,512,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,32,0.01551555593808492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,512,0.007582222421964009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,512,0.03324888812171088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,6144,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,16384,0.05881244606441922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,10240,0.03291022115283542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,3072,0.036032888624403216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,4096,0.01755911111831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,3072,0.03446400165557861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,7168,0.05492622322506375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,51200,0.047296888298458524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,12288,0.01884711119863722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,2048,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,768,0.020976889464590285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,3584,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,1536,0.0384897788365682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,512,0.023023999399609033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,3584,0.7746177779303657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,512,0.02496355606449975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,768,0.14799022674560547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,1024,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,32,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,256,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,7168,0.016287111573749118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,128,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,4096,0.04506311151716444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,10240,0.02235288918018341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,1536,0.040031111902660794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,256,0.028731554746627808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,256,0.034196443027920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,6144,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,1024,0.005858666780922148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,4096,0.014736000034544202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,10240,0.07976355817582872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,3072,0.024450666374630396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,2048,0.7992613580491809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,12288,0.05453688899676005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,768,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,8192,0.050572444995244346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,7168,2.9878063201904297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,10240,0.11212000581953262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,32,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,1024,0.01607644392384423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,3072,0.008799999952316284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,3072,0.008069333102968004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,65536,0.07485600312550862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,6144,0.07678311400943331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,65536,0.13881688647800022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,128,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,51200,0.0503422220547994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,65536,0.533774216969808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,7168,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,2048,0.024600888291994732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,1536,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,12288,0.14320888784196642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,16384,0.1710204415851169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,2560,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,1024,0.014156444205178155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,10240,0.20869778262244332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,2560,0.007418666448858049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,2048,0.012705778082211813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,512,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,64,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,3584,0.047212445073657565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,1024,3.4067412482367625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,768,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,7168,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,2560,0.02200355629126231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,2560,0.04178222351604038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,128,0.0019208888212839763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,2560,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,6144,0.26611555947197807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,16384,0.6712017589145236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,2048,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,51200,0.042805333932240806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,8192,0.014012444350454541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,32,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,2560,0.022114665971861944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,3584,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,1024,0.04500266578462389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,2560,0.018792000081804063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,6144,0.019648889700571697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,51200,0.09741510947545369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,6144,0.05807999769846598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,512,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,1536,0.0395448903242747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,1536,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,12288,0.18480355209774443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,1536,0.03911822372012668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,768,0.05706044700410631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,4096,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,1536,0.028216001060273912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,256,0.01682400041156345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,1024,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,16384,0.07970399989022149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,128,0.012734221915404001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,2560,0.023119111855824787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,3072,0.017821333474583097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,32,0.030496888690524634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,1024,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,64,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,65536,0.021120000216695998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,512,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,3072,0.05685777796639336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,512,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,2048,0.03246577911906772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,5120,0.07496444384256999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,16384,0.13584089279174805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,32,0.0022408889813555614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,2560,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,3072,0.018910222583346896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,128,0.010767111347781287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,12288,0.023775999744733173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,3072,0.005842666659090254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,7168,0.013041777743233575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,3584,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,12288,0.08878578080071343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,1024,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,10240,0.31465599271986217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,16384,0.095396445857154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,6144,0.033992889854643084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,2048,0.01240888900227017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,128,0.013762666947311826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,128,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,16384,0.026094221406512793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,65536,1.6309057871500652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,16384,0.04165688819355435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,5120,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,1536,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,8192,0.4131191041734483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,128,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,65536,0.14800000190734863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,6144,0.0948266651895311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,16384,3.389084498087565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,4096,0.028374221589830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,4096,0.03814844290415446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,256,0.005704000178310607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,3072,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,10240,0.012145777543385824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,3072,0.05203555689917671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,16384,0.021698666943444148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,65536,0.425343116124471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,2560,0.11368444230821396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,6144,0.1730639934539795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,6144,0.02557955516709222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,16384,0.025828444295459326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,768,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,6144,0.004607111215591431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,1536,0.029482665989134047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,65536,0.18253510528140596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,3072,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,51200,0.08721866872575547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,10240,0.037978665696250066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,7168,1.1862249374389648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,3072,0.4519342316521539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,1024,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,65536,0.03403466608789232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,65536,0.2263395521375868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,16384,1.2324462466769748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,128,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,3584,0.34141156408521867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,512,0.1515537765291002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,5120,0.008937777744399177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,3072,0.02280355493227641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,2560,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,7168,0.03310577736960517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,4096,0.2444373236762153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,4096,0.03007555670208401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,32,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,3072,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,16384,0.021627555290857952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,3072,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,256,0.003903111235962974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,256,0.017222222354676988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,2560,0.17671733432345918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,51200,0.07433511151207818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,4096,0.07199822531806098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,7168,0.010108444425794814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,512,0.012910222013791403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,512,0.007358222372002072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,12288,0.18567466735839844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,8192,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,4096,0.008774222599135505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,512,0.03681866659058465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,1024,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,2560,0.05470755696296692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,51200,0.5254613558451334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,51200,0.5004924668206109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,3072,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,2048,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,7168,0.0270462218258116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,768,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,2048,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,16384,0.05641688903172811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,7168,0.04988977644178602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,7168,0.054639998409483165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,5120,0.24644621213277182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,1024,0.027089777919981215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,2560,0.022131555610232886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,3072,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,8192,0.21392177210913765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,128,0.05084977878464592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,65536,0.06865600082609388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,65536,0.05981333388222588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,3072,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,32,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,256,0.016505777835845947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,6144,0.043863998519049756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,512,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,3584,0.08715199761920506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,12288,0.033803555700514056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,10240,0.18139021926456025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,51200,0.09169689151975845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,2048,0.10061689217885335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,1536,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,128,0.02832711074087355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,1536,0.018491556247075398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,1024,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,32,0.007932444413503012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,3584,0.01735911104414198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,1024,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,6144,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,512,0.10045688682132298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,51200,0.050592889388402305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,768,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,8192,0.7221306694878472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,32,0.016397333807415433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,3072,0.11263644695281982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,5120,0.019691555036438834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,16384,0.040765331851111516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,12288,0.1539520025253296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,256,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,1024,0.05810578001870049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,7168,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,2560,0.01551822159025404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,128,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,8192,0.06412533256742689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,4096,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,1536,0.0038542221817705366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,768,0.0203146669599745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,3072,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,4096,0.0180711117055681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,5120,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,8192,0.020296888218985665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,3584,0.01868355605337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,8192,1.123887062072754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,4096,0.040822221173180476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,512,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,4096,0.02436977790461646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,1536,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,5120,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,8192,0.015607110328144498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,1536,0.022321777211295232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,512,0.23735645082261828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,51200,10.44240485297309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,3584,0.03122133347723219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,5120,0.23190132776896158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,3584,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,2560,0.07627999782562256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,16384,0.022358222140206233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,64,0.06543555524614122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,2560,0.05013422171274821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,512,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,5120,0.013271110753218332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,3072,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,768,0.018744889232847426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,6144,0.25393422444661456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,128,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,65536,0.018577777677112155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,128,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,65536,0.06429421901702881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,4096,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,2560,0.007695111135641734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,64,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,2560,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,1536,0.005015999906592899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,32,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,65536,0.561648898654514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,6144,0.0255084451701906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,6144,0.01760444376203749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,512,0.010448888772063784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,3584,0.020463999774720933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,2048,0.02658577760060628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,128,0.01552177800072564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,128,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,12288,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,1024,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,2048,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,10240,0.032229334115982056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,4096,0.0644657810529073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,2048,0.017551110850440133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,5120,0.02148977749877506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,2560,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,7168,0.012071111136012606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,12288,2.8181919521755643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,10240,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,256,0.00792711145348019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,256,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,2048,0.03938577903641595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,64,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,8192,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,16384,0.043903999858432345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,8192,0.021949332621362474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,7168,0.2572666539086236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,51200,0.44049866994222003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,16384,3.6595891316731772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,256,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,3072,0.021725333399242822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,1024,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,1024,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,8192,0.02535733415020837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,8192,0.01130844404300054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,256,0.002591111179855135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,768,0.01147911118136512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,32,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,2048,0.007528000407748752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,2560,0.4730648994445801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,16384,0.1356897751490275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,2560,0.021798221601380244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,128,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,2560,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,10240,0.042007111840777926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,5120,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,7168,0.030477331744299993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,512,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,2560,0.007443555527263218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,32,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,51200,0.020580444071027968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,512,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,1536,0.29319998953077525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,65536,0.16812977525922987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,8192,0.033279998434914485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,3072,0.010087110930018956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,1536,0.028941333293914795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,3072,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,10240,0.03464622298876444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,1536,0.13997066020965576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,51200,3.2650960286458335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,65536,3.3617341783311634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,2560,0.007232888705200619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,16384,0.02369511127471924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,2560,0.1551591157913208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,12288,0.010296000374688042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,8192,0.02271466619438595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,32,0.03369599911901686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,3584,0.027982221709357366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,32,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,65536,0.1488835546705458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,65536,0.0640631119410197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,5120,0.04348977737956577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,12288,0.0425600012143453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,3584,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,51200,0.08376533455318874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,1024,0.006070222291681502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,2048,0.033596442805396184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,6144,0.07639377646976046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,8192,1.5513875749376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,128,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,512,0.020232000284724765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,64,0.0018088887963030075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,65536,0.12575822406344944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,3072,0.020820443828900654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,12288,1.0993288887871635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,5120,0.12275377909342448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,5120,0.06218666500515408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,3584,0.021531555387708876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,3072,0.01200711147652732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,7168,0.014125333891974555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,4096,0.061547557512919106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,2048,0.028759111960728962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,768,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,1024,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,51200,0.04613066713015238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,6144,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,5120,0.026535999443795946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,512,0.008032889001899296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,1024,0.0041759999261962045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,32,0.03199022346072727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,4096,0.03140799866782294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,12288,0.22574133343166777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,10240,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,5120,0.01367644468943278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,256,0.02898755669593811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,10240,0.5557164616054958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,7168,0.028332445356580947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,2560,0.01681155628628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,1024,0.05383377936151293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,51200,0.04522577921549479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,10240,0.33850399653116864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,7168,2.4374542236328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,1536,0.03264177838961283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,65536,0.22787290149264863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,2560,0.010835555692513784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,3584,0.053151110808054604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,4096,0.023751111494170293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,2560,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,128,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,5120,0.021735999319288466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,2560,0.00666933341158761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,768,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,6144,0.1072488890753852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,12288,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,2560,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,4096,0.21417600578731963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,5120,0.12127021948496501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,10240,0.03517422080039978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,768,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,64,0.0038275557259718576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,2560,0.03348177671432495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,5120,0.012645333177513547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,16384,0.3650248845418294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,2560,0.06325333648257785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,12288,0.16856443881988525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,3584,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,3584,0.03564533260133531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,2048,0.3565039899614122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,8192,0.0537377761469947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,2048,0.017887999614079792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,128,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,12288,0.02546399997340308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,768,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,3072,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,4096,0.013401777380042605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,2560,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,2560,0.01646844380431705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,12288,0.023418666587935552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,768,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,2048,0.05167555477884081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,1024,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,8192,0.02327377763059404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,8192,0.011250666446155973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,1536,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,12288,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,1536,0.029672000143263075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,512,0.024829333027203877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,512,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,12288,0.04447911183039347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,1536,0.4005431069268121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,6144,0.009525333013799455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,3584,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,256,0.06535555256737603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,256,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,2048,0.4221679899427626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,7168,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,4096,0.4677768813239203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,1536,0.023170666562186346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,16384,0.18420533339182535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,2560,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,768,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,8192,0.01646844380431705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,768,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,2048,0.009566222627957663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,10240,0.2126613325542874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,512,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,10240,0.20942044258117676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,7168,0.03172711200184292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,8192,0.02626133296224806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,8192,0.30102401309543186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,12288,0.1314515537685818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,3072,0.01500888831085629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,7168,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,51200,0.16171199745602077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,8192,0.012283555335468717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,5120,0.02651111119323307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,3072,0.008173333273993598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,8192,0.014382221632533602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,64,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,5120,0.028344889481862385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,12288,0.20399999618530273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,2560,0.00962488849957784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,1536,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,7168,0.2218195597330729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,1024,0.01475822263293796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,10240,0.10030577580134074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,51200,0.051057779126697116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,128,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,5120,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,5120,0.04747733473777771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,3584,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,4096,0.03426577647527059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,2048,0.011051555474599203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,64,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,2048,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,64,0.03358488943841722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,4096,0.09790044360690647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,51200,0.3021768993801541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,64,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,16384,0.024453333682484094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,12288,0.025042666329277888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,5120,0.10751644770304362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,7168,0.024550222688251074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,51200,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,51200,0.11913066440158421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,3072,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,51200,0.3337911234961615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,8192,0.039614223771625094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,16384,0.03146400054295858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,5120,0.026569777064853247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,1536,0.012181333369678922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,2048,0.008790222307046255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,32,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,1536,0.011295110815101199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,1024,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,128,0.015787555111779105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,2048,0.010979555547237396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,5120,0.015100444356600443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,512,0.04140533341301812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,768,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,4096,0.019349333312776353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,7168,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,16384,0.0971413320965237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,51200,0.05000444584422641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,768,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,1024,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,3072,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,512,0.014858666393491956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,3584,0.011339555184046427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,8192,0.01402844488620758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,8192,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,7168,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,10240,0.04213688770929972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,2048,0.008346666892369589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,51200,0.037802666425704956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,4096,0.03453066613939073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,512,0.0033377777371141645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,7168,0.03707911239729987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,1536,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,64,0.0023484445280498928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,5120,0.06442311074998644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,5120,0.020735111501481798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,3584,0.022494221727053326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,6144,0.02899555696381463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,4096,0.02068800065252516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,10240,0.07322844531801012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,1024,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,65536,0.09168533484141032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,7168,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,12288,0.06817422310511272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,256,0.10567466417948405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,64,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,10240,0.03544800149069892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,3072,0.009287999735938178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,2560,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,4096,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,1536,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,768,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,10240,0.031253334548738264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,10240,0.11251110500759548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,8192,0.17404445012410483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,65536,0.17360088560316297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,768,0.049116445912255176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,768,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,6144,0.021670222282409668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,8192,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,2560,0.01420177850458357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,2048,0.045928001403808594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,12288,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,7168,0.20972977744208443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,4096,0.17803911368052164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,51200,0.3483875592549642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,3072,0.1238444381290012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,2048,0.007724444071451823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,7168,0.07580266396204631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,768,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,256,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,12288,0.11338310771518284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,3584,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,6144,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,512,0.014039110806253223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,3072,0.011203555597199334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,2560,0.020995555652512446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,2560,0.045158220662011035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,16384,0.1389253271950616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,4096,0.012852444416946836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,768,0.003559999995761447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,5120,0.02289955483542548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,4096,0.03829866647720337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,512,0.005476444545719359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,128,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,4096,0.015536889433860779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,16384,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,32,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,16384,0.16579643885294595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,64,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,1024,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,128,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,512,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,6144,0.022272000710169475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,12288,0.017290666699409485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,3072,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,128,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,3584,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,768,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,128,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,32,0.01238844460911221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,16384,0.028130667077170476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,10240,0.1874542236328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,12288,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,3072,0.040828443235821195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,4096,0.020879111356205408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,10240,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,65536,0.0885288914044698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,768,0.007582222421964009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,2048,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,8192,0.27528177367316353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,2560,0.17716711097293428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,128,0.005373333477311664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,5120,0.019669334093729656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,6144,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,1536,0.274437321556939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,1024,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,128,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,2560,0.10598933034472996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,8192,0.029832889636357624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,4096,0.017243555850452848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,128,0.006217777729034424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,8192,0.06354577673806085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,2560,0.00518844442235099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,512,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,128,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,7168,0.14589867326948378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,3584,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,64,0.019261333677503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,6144,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,2048,0.06395022074381511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,512,0.014103111293580798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,10240,0.0916995538605584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,3584,0.1593502230114407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,2048,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,5120,0.026139555705918208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,10240,0.02172800070709652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,7168,0.31560712390475804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,2048,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,256,0.03950488898489211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,16384,0.03030311067899068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,12288,2.649655024210612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,12288,0.10101511081059773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,5120,0.07549244165420532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,6144,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,1024,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,51200,0.03794044587347243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,3072,0.03517333335346646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,512,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,7168,0.013022222452693515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,1024,0.01741333305835724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,7168,0.11604621675279404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,768,0.01295377810796102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,1024,0.004571555389298333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,1536,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,65536,1.1215733422173393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,12288,0.06817866696251763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,2560,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,65536,0.20520799689822725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,1536,0.048877331945631236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,51200,0.23973777559068468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,3072,0.06850133339564006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,6144,0.01864977843231625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,12288,0.02147377696302202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,2560,0.045088890526029796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,8192,0.061894224749671094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,3072,0.04075110952059428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,7168,0.04186400108867221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,2560,0.0074373334646224976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,4096,0.03199111090766059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,64,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,4096,0.02347466680738661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,1536,0.00674933319290479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,12288,0.06457333432303534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,65536,1.1598676045735676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,65536,0.5669537650214301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,16384,0.02327199942535824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,51200,0.051784889565573804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,64,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,64,0.006659555352396435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,256,0.006755555669466655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,128,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,32,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,6144,0.050274666812684804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,512,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,768,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,51200,0.0585742195447286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,4096,0.03189511100451151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,12288,0.029840889904234145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,3584,0.017265778448846605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,2048,0.021382222572962444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,2048,0.338750229941474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,8192,0.01761155492729611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,12288,0.04293422235382927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,3072,0.04589244392183092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,2560,0.017782222893502977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,51200,12.57814195421007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,51200,2.5844222174750433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,8192,0.9380061891343859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,5120,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,51200,0.25007377730475533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,65536,0.4990880224439833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,8192,0.4062177870008681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,16384,0.06885244448979695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,7168,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,5120,0.021648888786633808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,51200,0.15591377682156032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,2048,0.031077331966824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,7168,0.026403556267420452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,768,0.025224000215530396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,512,0.03347999850908915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,2560,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,512,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,65536,0.394956456290351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,51200,0.08488177590899998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,1024,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,64,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,3584,0.01130311108297772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,3584,0.028505778974956934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,6144,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,512,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,7168,0.012865777644846173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,1536,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,5120,0.012315555579132505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,2560,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,5120,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,16384,0.10817155573103164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,3584,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,16384,0.012005333271291522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,1536,0.005467555589146084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,768,0.010120000276300643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,4096,0.010867555936177572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,2560,0.06132088767157661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,1024,0.013702222042613559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,7168,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,4096,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,12288,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,65536,0.304421345392863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,256,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,16384,0.017343110508388944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,51200,0.4010133213467068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,51200,0.1946320003933377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,5120,0.04951022068659464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,512,0.0038577777643998465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,4096,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,1024,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,2048,0.009177777502271865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,1536,0.006383111079533895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,256,0.010738666686746808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,4096,0.00870577742656072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,4096,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,2048,0.4865146742926703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,12288,0.06155111392339071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,3072,0.009097778134875828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,128,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,12288,0.22451999452379015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,2048,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,64,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,32,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,32,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,5120,0.010206222534179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,3584,0.015054222610261707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,1536,0.03629422187805176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,7168,0.01680177781316969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,768,0.0295466681321462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,512,0.0040835556056764396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,7168,0.15965867042541504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,16384,0.02112355497148302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,3072,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,16384,0.021645334031846788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,64,0.020230222079488967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,64,0.005201777650250329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,8192,0.032680001523759626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,16384,0.1053875552283393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,51200,1.0837991502549913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,10240,0.027412444353103638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,8192,0.02139644490347968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,128,0.01924444403913286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,12288,0.02123733361562093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,2048,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,65536,0.05950488646825155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,768,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,3584,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,12288,0.01642400026321411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,512,0.18791288799709746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,3072,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,512,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,2048,0.03586133321126302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,7168,0.09984266757965088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,65536,1.5641244252522786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,1024,0.014257777068350049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,6144,0.07509511046939425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,256,0.053789334164725415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,128,0.00849777791235182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,3584,0.022502221994929846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,768,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,65536,0.04828977915975782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,6144,0.03167555398411221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,10240,0.08728355831570095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,16384,0.5446702109442817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,2048,0.021951999929216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,128,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,7168,0.250328885184394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,2560,0.020717332760492962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,3072,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,768,0.012500444220172035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,4096,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,4096,0.10354933473798965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,8192,0.023016000787417095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,4096,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,12288,0.02053244411945343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,2560,0.015736889508035447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,64,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,3584,0.06520889202753703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,10240,0.02743822170628442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,7168,0.0683386656973097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,256,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,3072,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,768,0.0042462220622433555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,2048,0.006688888702127669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,65536,0.04861066738764445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,12288,0.13455643918779162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,6144,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,10240,0.013064888616402945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,16384,0.070196443133884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,7168,0.1459208859337701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,2048,0.005305777821275923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,2048,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,2048,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,12288,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,10240,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,512,0.0037991110649373797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,768,0.0034835553831524322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,6144,0.32034312354193795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,16384,0.022039999564488728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,3584,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,51200,0.8719973564147949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,8192,0.12182754940456814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,2560,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,128,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,5120,0.00962577760219574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,2560,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,5120,0.008836444053384993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,768,0.029326223664813574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,8192,0.07520088884565565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,2048,0.0041946665280395085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,4096,0.019221333993805777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,12288,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,1536,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,16384,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,8192,0.02718488872051239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,3072,0.023139556248982746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,1536,0.023588443795839947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,3584,0.7670506901211209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,65536,0.18216355641682944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,512,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,256,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,12288,0.037576887342664934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,6144,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,8192,0.1629013352923923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,768,0.015656888484954834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,51200,1.2318826251559787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,2560,0.017702221870422363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,7168,0.15151466263665095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,256,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,2048,0.012643555800120035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,128,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,12288,0.030593779351976182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,1024,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,2560,0.014767110347747803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,128,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,512,0.014975999792416891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,128,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,10240,0.3348711066775852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,12288,5.419656117757161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,2560,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,64,0.01166488892502255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,128,0.0019120000716712738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,768,0.003137777869900068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,2048,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,768,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,32,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,1024,0.0074720001882976955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,256,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,768,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,1536,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,1024,0.04792177677154541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,7168,0.020233778489960563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,2048,0.12939822673797607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,2048,0.018952000472280715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,1536,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,2048,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,3584,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,5120,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,2560,0.025637333591779072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,512,0.009144888983832465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,32,0.006130666782458623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,1024,0.006343110981914733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,256,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,3584,0.0539173318280114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,128,0.005395555661784277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,256,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,7168,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,10240,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,2048,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,256,0.0028924445311228433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,2048,0.011219555305110084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,512,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,12288,0.013577777478430005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,5120,0.00869777798652649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,32,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,10240,0.04533955454826355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,1024,0.019094222121768527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,65536,0.07630399862925212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,10240,0.01754399968518151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,3584,0.026233777403831482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,65536,0.061075554953681103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,256,0.02011022302839491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,65536,0.1195911169052124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,5120,0.008766222331258986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,256,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,6144,0.05603555838267008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,64,0.034641779131359525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,512,0.015167999598715039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,128,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,10240,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,5120,0.004893333133724001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,256,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,128,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,6144,0.023358222511079576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,6144,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,2560,0.007374222079912822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,5120,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,3584,0.029426667425367568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,1536,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,3072,0.014168888330459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,4096,0.592631975809733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,128,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,1024,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,10240,0.068067557281918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,7168,0.024155555499924555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,256,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,8192,0.011895111037625207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,128,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,1024,0.006134222365087933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,10240,0.0459591117170122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,16384,0.042837331692377724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,3584,0.03468088971243964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,6144,0.022983110613293115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,8192,0.11041866408454047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,8192,0.12992444303300646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,3072,0.026385777526431616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,1024,0.01961955593691932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,10240,0.0848204427295261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,4096,0.03587999939918518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,2048,0.02769688930776384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,512,0.0026906666656335196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,1024,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,10240,0.016764443781640794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,8192,0.024168888727823894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,5120,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,12288,0.3862631056043837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,3584,0.011964444484975604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,51200,0.21478577454884848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,10240,0.024143111374643113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,7168,0.1403955618540446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,5120,0.0868426693810357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,768,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,4096,0.33734843466016984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,4096,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,4096,0.027748444014125403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,65536,1.065776930914985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,5120,0.025765332910749648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,256,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,12288,0.044679999351501465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,512,0.0028897778441508612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,10240,0.6235075526767307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,5120,0.01533333294921451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,1536,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,768,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,7168,0.014313777287801107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,6144,0.03158844510714213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,2048,0.039481778939565025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,8192,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,16384,0.0588266650835673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,8192,0.021959111094474792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,512,0.08640711175070868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,512,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,8192,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,1536,0.03635911146799723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,128,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,10240,0.1250151130888197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,4096,0.47538042068481445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,6144,0.010055999788973067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,1024,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,768,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,64,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,32,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,2560,0.02862488892343309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,51200,1.9280880822075739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,6144,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,32,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,65536,13.49284447564019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,3072,0.01862133377128177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,2048,0.02993866801261902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,1536,0.039093332158194646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,4096,0.034240000777774386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,768,0.01664444473054674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,6144,0.03422133458985223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,65536,0.2210328843858507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,2048,0.5398569107055664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,51200,0.09534400039248997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,6144,0.018274666534529794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,3584,0.08047378063201904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,51200,0.04450577828619215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,51200,3.420197380913628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,16384,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,7168,0.06456444660822551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,1536,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,6144,0.00980711148844825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,65536,2.2343457539876304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,6144,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,7168,0.05804444683922661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,3072,0.5532728830973307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,128,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,3584,0.01127377814716763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,128,0.030833777454164293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,4096,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,1536,0.07201066944334242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,16384,0.09595111343595718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,3072,0.01311288856797748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,256,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,3072,0.9156213336520724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,2560,0.031324442889955305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,51200,0.06300089094373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,65536,0.08487021923065186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,16384,0.038189331690470375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,4096,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,12288,0.3961164421505398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,6144,0.021546666820844013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,128,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,7168,0.00544088871942626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,16384,0.03143822153409322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,6144,0.018009778526094224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,5120,0.009490666290124258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,6144,0.13380000326368544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,3072,0.010096000300513374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,16384,0.08565688795513576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,1536,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,51200,0.04721511072582669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,7168,0.012171555724408893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,10240,0.030015998416476782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,12288,0.2183928887049357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,1536,0.10138133499357437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,2560,0.06151377492480808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,5120,0.08293510807885064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,12288,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,512,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,512,0.00812533332241906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,2560,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,8192,0.07236178053749932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,12288,0.041590223709742226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,3072,0.028821332587136164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,6144,0.017445334129863314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,12288,0.017807111144065857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,12288,0.03279644582006667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,768,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,7168,0.02139644490347968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,7168,0.5009866820441352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,10240,0.6323528819613987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,7168,0.019457777341206867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,2048,0.050791111257341176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,4096,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,32,0.009284444153308868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,2560,0.013398222625255585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,256,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,256,0.017857778403494094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,2048,0.007052444749408298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,4096,0.02385422256257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,256,0.02870044443342421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,128,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,6144,0.012994666894276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,65536,0.243797328737047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,1024,0.0028773333049482773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,8192,0.010959111154079437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,2560,0.039247999588648476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,65536,0.3655173248714871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,8192,0.10336177878909641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,256,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,12288,0.016561778055297006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,5120,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,2048,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,512,0.04359377755059136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,2560,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,3072,0.005752000129885144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,64,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,2560,0.15723378128475612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,8192,0.34928533766004777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,768,0.0032924444725116095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,64,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,1536,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,32,0.005986666513813867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,2560,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,8192,0.11814933353000218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,51200,0.12210578388637966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,7168,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,3584,0.1221724483701918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,768,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,3072,0.02252799934811062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,1024,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,51200,0.04849955439567566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,64,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,1024,0.03642844491534763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,768,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,1024,0.017260443833139207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,6144,0.03585422370168898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,1024,0.016203555795881484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,512,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,8192,0.15197243955400255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,2048,0.08091466956668429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,1536,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,5120,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,768,0.005703111075692707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,10240,0.014115555418862237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,3072,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,3584,0.22051821814643013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,12288,4.684818691677518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,51200,0.04016977879736159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,32,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,65536,0.5628346867031521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,6144,0.028047998746236164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,256,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,64,0.0019706665641731685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,2048,0.014456000592973499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,8192,0.08395466539594863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,3584,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,512,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,5120,0.011219555305110084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,51200,0.21650132868025038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,256,0.09832888841629028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,65536,0.01611911091539595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,512,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,65536,0.18297955724928114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,3072,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,2560,0.054921779367658824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,2560,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,12288,0.08397777875264485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,768,0.009259555902745988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,2048,0.02810933192571004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,32,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,512,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,51200,0.4288231001959906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,1024,0.06306399901707967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,768,0.16688177320692274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,256,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,2560,0.04981244603792826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,128,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,3584,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,65536,0.12172622150844997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,5120,0.009528000321653154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,2048,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,512,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,51200,0.22621689902411568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,32,0.007205333146784041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,5120,0.02151022189193302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,768,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,7168,0.06790044572618273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,7168,0.07453422413931952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,64,0.0018088887963030075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,3584,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,12288,0.013353777428468069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,12288,0.011319110790888468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,5120,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,1024,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,1536,0.02784444557295905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,5120,0.06774044699139066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,1024,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,5120,0.02109333376089732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,5120,0.010494222243626913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,768,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,12288,0.04976977904637655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,8192,0.31589243147108287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,16384,1.8295154571533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,2048,0.030674666166305542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,6144,0.03888622257444594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,1536,0.028636445601781208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,64,0.0028773333049482773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,6144,0.28963110182020396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,4096,0.008151999778217739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,1024,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,128,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,64,0.19541777504814994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,6144,0.017047999633683097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,10240,0.04896266592873467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,3072,0.31313244501749676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,128,0.005480888817045424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,5120,0.018233777748213876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,6144,0.012683555483818054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,64,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,2560,0.2263031005859375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,128,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,12288,0.030270222160551283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,5120,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,3584,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,1024,0.01016177816523446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,128,0.035424000687069364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,7168,0.30665243996514213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,3072,0.00777333312564426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,64,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,3584,0.00757422215408749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,8192,0.010999999940395355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,1536,0.09876266453001235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,1024,0.0035031110876136353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,7168,0.05202044381035698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,10240,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,12288,0.03392088744375441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,65536,0.19309066401587593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,512,0.027385777897304956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,4096,0.019854222734769184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,256,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,64,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,65536,0.9156417846679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,768,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,7168,0.00998222248421775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,65536,0.13731466399298772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,768,0.004967999955018361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,2048,0.004857777721352047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,8192,0.9644746780395508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,256,0.1801848941379123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,10240,0.07818310790591769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,768,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,1536,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,6144,0.11444800429873997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,768,0.018883556127548218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,10240,0.05167288912667168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,51200,0.11733778317769368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,512,0.01553600033124288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,10240,0.02773688899146186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,3072,0.020124443703227572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,3584,0.0417555570602417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,32,0.011360888679822287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,512,0.006816888848940532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,1536,0.06605155600441827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,3072,0.009142222503821055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,2560,0.6260488828023275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,256,0.007436444362004598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,51200,0.055813332398732506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,2048,0.02256711158487532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,32,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,64,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,768,0.025352888637118872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,2048,0.012864000267452665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,51200,0.15617777241600886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,5120,0.2409306632147895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,512,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,5120,0.1101493305630154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,2048,0.02259822189807892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,768,0.010114666488435533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,8192,0.031102223528756037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,10240,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,32,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,768,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,32,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,10240,0.013944889108339945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,768,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,2048,0.018444443742434185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,512,0.003075555587808291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,16384,0.30049154493543834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,2560,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,7168,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,10240,0.28882310125562877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,768,0.06342133548524645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,2560,0.08844621976216634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,512,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,51200,0.0752426650789049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,8192,0.012138666378127204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,3584,0.028966221544477675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,512,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,65536,2.445720884535048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,10240,0.11610844400193955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,128,0.012999999854299756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,10240,0.021021333005693223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,2560,0.00815466625822915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,65536,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,6144,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,1024,0.0060159998635451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,51200,0.15597332848442927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,2048,0.019655111763212416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,3072,0.01056177748574151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,7168,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,2048,0.011178666518794166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,3584,0.32753777503967285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,8192,0.021459556288189355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,2048,0.025683555338117812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,12288,0.11010222302542792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,512,0.008543999658690559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,7168,0.021262221866183813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,2048,0.03296800123320685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,32,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,3584,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,12288,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,2560,0.014530667000346713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,2560,0.047472887569003634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,32,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,8192,0.02349066734313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,65536,0.05989866786532932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,3584,0.010307555397351583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,3584,0.013277333643701343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,2560,0.06537777847713895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,1024,0.019102222389645047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,16384,4.164531707763672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,6144,0.06724888748592801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,6144,0.056483553515540234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,512,0.014056000444624158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,12288,0.046169777711232506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,4096,0.007596444752481248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,7168,0.029676444000667993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,10240,0.18329066700405547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,2048,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,65536,0.06384266747368707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,512,0.014735110931926303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,4096,0.018215111560291715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,4096,0.051700443029403687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,64,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,2048,0.03253688746028476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,3072,0.014718222949239941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,16384,0.015784000356992085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,2048,0.31481067339579266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,2560,0.01332622187005149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,8192,0.07301244470808241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,65536,2.81620364718967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,3584,0.04983644353018867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,1536,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,3584,0.01906755566596985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,1024,0.013176889055305056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,65536,0.07934311363432142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,3584,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,3072,0.021868444151348535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,51200,0.06528977553049724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,6144,0.017041777571042378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,3584,0.008451555338170793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,7168,0.018568888306617737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,256,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,16384,0.07152089145448473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,3584,0.02107288936773936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,512,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,16384,0.02293955617480808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,512,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,6144,0.16313778029547796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,32,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,2560,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,256,0.017979555659823947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,768,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,51200,0.04970755510860019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,65536,0.07514933082792494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,10240,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,768,0.02128000060717265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,64,0.010400888820489248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,7168,0.03289422061708238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,128,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,8192,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,3584,0.034309334225124784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,2560,0.026263111167483862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,10240,0.06017511420779758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,768,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,1536,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,7168,0.026087111896938745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,512,0.0052879999081293745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,3584,0.023795555035273235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,8192,0.010479110810491772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,1024,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,1536,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,32,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,65536,0.08032266961203681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,3072,0.014166222678290473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,128,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,5120,0.026388444834285315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,256,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,51200,0.14963466591305202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,768,0.024288889434602525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,32,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,16384,0.05005333489841885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,1024,0.004883555488453971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,256,0.02537511123551263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,3584,0.07646844122144911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,10240,0.023406222462654114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,64,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,4096,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,2048,0.046704888343811035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,10240,0.08123377958933513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,8192,0.058265778753492564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,2048,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,8192,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,3584,0.03375733229849074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,128,0.019630221856964957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,1536,0.02107200026512146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,10240,0.06227733029259575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,1024,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,1536,0.00723555518521203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,4096,0.04073955615361532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,256,0.011135111252466837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,3584,0.019447111421161227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,3072,0.03711288836267259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,65536,0.2377075619167752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,5120,0.008919999831252629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,7168,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,4096,0.04979911115434435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,128,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,256,0.011139555937714048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,256,0.07668355438444349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,51200,0.6104257901509603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,2560,0.1237751113043891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,1024,0.012074666718641916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,128,0.01796533332930671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,128,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,1536,0.09676711426840888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,5120,0.018948444061809115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,4096,0.02453511125511593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,3072,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,10240,0.693086200290256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,7168,0.012849777936935425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,32,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,768,0.0031884445084465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,3072,0.040539556079440646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,7168,0.06575733423233032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,3072,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,16384,0.37821067704094785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,64,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,1024,0.017236444685194228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,256,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,16384,2.6787003411187063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,2048,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,5120,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,32,0.0063635553750726916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,8192,0.1790764464272393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,64,0.0027875554644399216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,1024,0.009278222090668148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,2560,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,5120,0.010630221830474006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,6144,0.027038221557935078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,3584,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,3584,0.04134577843878005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,1024,0.00886311133702596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,6144,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,64,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,3072,0.012071111136012606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,4096,0.09871821933322483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,7168,0.08767022026909722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,12288,0.02783999840418498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,5120,0.11812355783250596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,128,0.002385777731736501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,3072,1.4424222310384114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,10240,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,7168,0.256839116414388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,51200,1.2042071024576824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,6144,0.02158222264713711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,12288,0.053878221246931285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,8192,0.022104889154434204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,128,0.02405066622628106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,16384,0.02624799973434872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,3584,0.018591110905011494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,1024,0.007242666350470648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,256,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,256,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,51200,1.198486222161187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,4096,0.017435555656750996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,3072,0.056481778621673584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,3072,0.013800889253616333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,6144,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,64,0.030829333596759375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,12288,0.06121955315272013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,4096,0.018230222993426852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,64,0.009148444566461775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,4096,0.041965332296159535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,8192,0.051626665724648364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,512,0.019550222489568923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,768,0.018696889281272888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,768,0.011563555234008364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,1024,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,7168,0.019332443674405415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,6144,0.02664799988269806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,2048,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,1536,0.16199555661943224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,10240,0.06028088596132067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,768,0.004304888761705822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,65536,0.16592089335123697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,16384,0.03046577837732103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,1536,0.009085333181752099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,128,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,2560,0.007351111206743453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,12288,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,1536,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,12288,0.14918666415744358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,5120,0.10099199745390151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,8192,0.29566044277615017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,51200,0.12713778018951416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,3584,0.03393422232733832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,10240,0.026125333375400964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,8192,0.019865777757432725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,2560,0.02832088867823283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,65536,0.08644621902041966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,1536,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,7168,0.07517688804202609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,7168,1.2548187043931749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,10240,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,2560,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,2048,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,128,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,7168,0.0860897766219245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,256,0.042372445265452065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,3584,0.009218666288587783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,1024,0.08114666408962674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,2560,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,1024,0.01606666710641649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,2048,0.07973511351479425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,3072,0.03184177809291416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,6144,0.029120001527998183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,1024,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,3584,0.029743111795849268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,65536,0.14120177427927652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,128,0.03445422318246629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,768,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,16384,0.03779110974735684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,5120,0.01829777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,3584,0.06006844176186455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,1536,0.030560887522167627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,1024,0.02811466654141744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,51200,9.940450032552084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,2048,0.037872887319988675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,8192,0.3410951031578912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,64,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,5120,0.06563466787338257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,10240,0.028498666154013738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,768,0.010438222024175854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,128,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,768,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,51200,0.06644178099102445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,128,0.02625333269437154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,5120,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,51200,0.04673155479960971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,10240,0.017909333109855652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,65536,0.1150053342183431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,2560,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,3584,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,2560,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,51200,0.4482497639126248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,32,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,64,0.07424355877770318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,1024,0.018561777141359117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,51200,3.1809476216634116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,1536,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,1536,0.009039999710188972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,2048,0.03225422236654494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,5120,0.01089777797460556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,2560,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,65536,1.2206444210476346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,32,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,2560,0.0945884452925788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,65536,0.3353857729170058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,7168,0.053871999184290566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,1536,0.026366222235891554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,4096,0.011382222175598145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,12288,0.13893511560228136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,4096,0.0317795541551378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,2048,0.007473777565691207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,65536,0.06463111109203763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,16384,0.0399875541528066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,1536,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,2048,0.024012444747818842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,768,0.003979555434650845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,3072,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,8192,0.4774808883666992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,65536,0.8308409055074056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,65536,7.9272045559353295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,10240,0.18208355373806426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,1536,0.02497422198454539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,6144,0.023307555251651343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,51200,0.037532445457246565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,1024,0.018230222993426852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,10240,0.02572711143228743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,1536,0.010478222535716163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,8192,0.019254222512245178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,3584,0.07829155524571736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,3584,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,10240,0.16362044546339247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,51200,0.3740382194519043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,7168,0.03360177742110358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,2560,0.08489866389168634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,6144,0.03497155507405599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,2560,0.08158933454089694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,1536,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,5120,0.02959022257063124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,6144,0.12788177861107722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,768,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,1024,0.006436444405052397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,51200,0.1886613368988037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,2560,0.06470844480726454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,6144,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,512,0.012879110872745514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,6144,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,256,0.014456000592973499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,2048,0.010590222146775989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,768,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,65536,0.05797511339187622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,256,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,7168,0.12159021695454915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,128,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,2560,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,8192,0.03716088996993171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,12288,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,256,0.017500445246696472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,2048,0.023379556006855432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,4096,0.10664888885286118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,512,0.023074666659037273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,51200,0.05121866530842251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,8192,0.04924177792337206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,1024,0.12010133266448975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,65536,0.09612888760036892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,32,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,7168,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,6144,0.38199199570549863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,65536,4.605626848008898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,10240,0.022687999738587275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,12288,0.04694755540953743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,12288,0.015760888655980427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,5120,0.18982666068606904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,512,0.005631999837027655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,7168,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,256,0.010236444572607676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,3072,0.016674665941132438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,3584,0.043861332866880626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,8192,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,1536,0.0070293330483966405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,5120,0.0643084446589152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,10240,0.39335910479227704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,5120,0.02528800070285797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,5120,0.038000000847710505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,1024,0.3661697705586751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,32,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,1024,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,10240,0.01538755496342977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,2560,0.008008889026112026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,2560,0.12568888399336073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,256,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,32,0.014875556031862894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,3072,0.17181600464714897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,128,0.01240533341964086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,16384,0.0225608895222346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,65536,0.2667857805887858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,768,0.011382222175598145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,10240,0.11295289463467068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,12288,1.0092586941189237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,1536,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,1536,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,6144,0.00998222248421775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,3072,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,12288,0.15820088651445177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,64,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,4096,0.014888000157144336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,256,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,64,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,3584,0.026320000489552815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,16384,0.04578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,10240,0.1883528894848294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,5120,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,6144,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,32,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,2048,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,5120,0.0561386677953932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,256,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,8192,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,256,0.017151999804708693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,65536,0.1181466711892022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,10240,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,51200,0.42009954982333714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,10240,0.03173955612712436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,10240,0.05607199668884277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,2048,0.023217777411142986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,3584,0.008656000097592672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,51200,2.2783501942952475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,2048,0.0069057775868309875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,12288,0.07713066869311862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,5120,0.03776266508632236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,16384,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,1536,0.004521777646409141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,1024,0.020454222957293194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,8192,0.03155288762516446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,8192,0.2380382219950358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,65536,0.06661155488755968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,3584,0.022317333353890315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,6144,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,2048,0.04413511024581062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,2560,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,10240,0.010653333531485664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,3072,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,16384,0.026161778304311965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,4096,0.06037244531843397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,3072,0.025038222471872967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,16384,0.018089777893490262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,5120,0.03166044420666165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,2048,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,1024,0.024233778317769367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,1024,0.03014311194419861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,1536,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,2560,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,3072,0.007738666401969061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,128,0.004253333227501975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,10240,0.07553688685099284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,10240,0.06916355424457125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,10240,0.009200000100665623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,1024,0.021155556042989094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,4096,0.022648000054889258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,512,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,256,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,10240,0.1895057757695516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,2048,0.05816000037723117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,12288,0.25240710046556264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,512,0.0634960002369351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,6144,0.5098089112175835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,7168,0.010413332945770687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,4096,0.029783109823862713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,16384,0.3120488855573866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,65536,0.14132711622450086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,5120,1.0202319886949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,16384,0.026231999198595684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,1536,0.04673066735267639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,12288,0.1598328881793552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,2560,0.060495111677381724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,12288,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,256,0.027928001350826685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,51200,0.1220666699939304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,3584,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,65536,0.3832728862762451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,1536,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,1536,0.0035137778355015647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,1024,0.015538665983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,768,0.01587288909488254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,1536,0.018817777434984844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,128,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,3584,0.030988444884618122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,6144,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,3072,0.03309333324432373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,768,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,1024,0.013168888787428537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,51200,0.3168604373931885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,51200,15.887903849283854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,6144,0.013032889200581444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,768,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,4096,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,6144,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,2560,0.060086223814222545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,256,0.15723821851942274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,2048,0.01054933336046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,256,0.021203555994563635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,2560,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,8192,0.17781689431932238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,7168,0.01090844472249349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,8192,0.015433778365453085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,1536,0.3834382163153754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,10240,0.413224008348253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,12288,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,10240,0.13163644737667507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,3584,0.027856889698240492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,6144,0.050943109724256724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,32,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,1024,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,128,0.003007111036115222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,1536,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,10240,0.023510222633679707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,3584,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,512,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,16384,0.022464000516467627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,128,0.006815111057625876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,64,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,256,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,1024,0.030530667967266504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,6144,0.030061334371566772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,3072,0.032616890139049955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,16384,0.017260443833139207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,6144,0.19734933641221789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,12288,0.034429334931903414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,256,0.011549333731333414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,256,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,6144,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,3072,0.03551910983191596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,16384,0.17270044485727945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,51200,0.06252711349063449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,6144,0.036990222003724836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,65536,0.024914667010307312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,8192,0.010775999890433418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,2560,0.03305955396758185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,12288,0.37865421507093644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,12288,0.045957333511776395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,3584,1.0011635886298287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,64,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,128,0.008559111091825698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,2560,0.03257511059443156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,64,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,65536,19.125352647569446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,128,0.10846222109264797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,65536,3.152040057712131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,7168,0.045928888850741915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,3584,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,51200,0.16334133678012425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,16384,0.05474844574928284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,7168,0.06838755475150214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,64,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,7168,0.017671111557218764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,8192,0.14611732959747314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,6144,0.020299555526839364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,5120,0.0391591125064426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,7168,0.022473777333895367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,65536,0.017408889200952318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,2560,0.6266782018873427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,1536,0.31938399208916557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,128,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,512,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,32,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,12288,0.04249777727656894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,5120,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,1536,0.014175999495718213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,1024,0.025840888420740765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,65536,0.051185776789983116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,768,0.018239110708236694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,12288,0.03238044513596429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,65536,0.7520613140530057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,5120,0.004610666798220741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,256,0.017626666360431247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,768,0.00721155520942476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,128,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,5120,0.03246311015552945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,8192,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,8192,0.024375110864639282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,16384,0.044086221191618175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,1024,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,1536,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,64,0.0018133332745896445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,768,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,6144,0.18482666545444063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,65536,0.33269066280788845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,6144,0.02254399988386366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,12288,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,256,0.002324444345302052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,8192,0.01350044459104538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,4096,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,3072,0.028145776854621038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,4096,0.03592711024814182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,12288,0.08564444382985432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,512,0.006747555401590135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,10240,0.06231821907891167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,4096,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,4096,0.42764443821377224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,16384,0.10539911190668742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,512,0.015318221516079374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,512,0.027452444036801655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,1024,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,4096,0.01148888882663515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,65536,0.18912177615695527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,2048,0.004858666823969947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,512,0.01648533344268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,10240,0.37178399827745223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,6144,0.07206755876541138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,256,0.0035048888789282907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,64,0.0042257776690853965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,3584,0.02459555533197191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,51200,0.19862043857574463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,12288,0.017055999901559617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,256,0.010420444111029306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,4096,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,32,0.021670222282409668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,512,0.0074275558193524676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,6144,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,2048,0.0192248887485928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,2560,0.019015999303923715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,65536,1.1015208562215169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,12288,0.392027563518948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,4096,0.014157333307796054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,8192,0.0232968893316057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,7168,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,7168,0.13354044490390354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,1024,0.02382044494152069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,51200,0.9351217481825086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,1536,0.029866665601730347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,3072,0.0608160032166375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,8192,0.024756444825066462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,4096,0.12113066514333089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,768,0.02308711078431871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,256,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,256,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,7168,1.7368035846286352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,5120,0.0697279969851176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,128,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,1024,0.09451911184522842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,6144,0.011251555548773872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,8192,0.07269510957929823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,65536,0.09438133239746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,5120,0.027306665976842243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,3072,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,8192,0.0321084459622701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,10240,0.047304888566335045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,5120,0.02789777848455641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,4096,0.051580445633994214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,4096,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,128,0.01333688861793942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,7168,0.034480889638264976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,10240,0.0635955532391866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,3584,0.38570311334398055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,512,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,1024,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,4096,0.015240889456537036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,2560,0.016927111479971144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,6144,0.05281866590181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,256,0.010868444210953183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,1024,0.004839999808205498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,2560,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,65536,0.08925777673721313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,3072,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,6144,0.08220266633563571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,768,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,3072,0.016884444488419425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,3072,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,768,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,6144,0.00463733325401942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,2048,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,10240,0.09605333540174697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,32,0.001628444426589542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,2560,0.0358497765329149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,2048,0.024704888463020325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,3584,0.0094035557574696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,16384,0.0196435550848643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,3072,0.03588177760442098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,5120,0.14030222098032633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,8192,0.05857955747180515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,768,0.020544000797801547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,256,0.021028444170951843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,10240,0.02642755541536543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,1024,0.016128000285890367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,4096,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,16384,7.893461439344619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,768,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,5120,0.011122666299343109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,2560,0.019906666543748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,65536,0.019609777463806998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,10240,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,1024,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,32,0.0041786668201287585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,128,0.003903999924659729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,1024,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,2560,0.09068444702360366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,12288,0.44075822830200195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,512,0.05983822213278877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,2048,0.10177510976791382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,512,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,8192,0.12664800220065647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,512,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,3584,0.03498488995763991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,1536,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,6144,0.15332089530097112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,128,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,10240,0.007091555330488417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,64,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,5120,0.00849333322710461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,12288,0.017972444494565327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,256,0.4009671211242676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,2048,0.030253334177864924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,512,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,2560,0.022992889086405437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,51200,0.8264657656351725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,65536,0.8255875375535754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,768,0.01074399964676963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,3584,0.6078417566087511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,32,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,512,0.03162755568822225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,8192,0.011982222398122152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,3072,0.026558222042189703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,1024,0.022831110490692988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,512,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,6144,1.0807831022474501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,16384,0.021171554923057556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,3584,0.10159911049736871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,256,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,2560,0.0120319997270902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,12288,0.3433982266320123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,512,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,16384,0.5925662252638075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,1536,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,16384,0.09555821948581272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,16384,0.08254488971498278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,4096,0.01742844449149238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,2560,0.010774222513039907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,12288,0.019660444723235238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,10240,0.17211022641923693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,7168,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,51200,0.3847671084933811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,128,0.0022826666633288064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,12288,0.03155733479393853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,7168,0.029117332564459905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,51200,0.21893155574798584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,5120,0.07030310895707872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,512,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,10240,0.025294222765498694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,3072,0.08927200237909953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,16384,0.04661600126160515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,128,0.009102221992280748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,65536,0.019461333751678467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,128,0.009734222458468543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,1024,0.01716800034046173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,768,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,3584,0.007914666500356462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,12288,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,2560,0.12587200270758733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,512,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,4096,0.6618382136027018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,128,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,1024,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,8192,0.3182106547885471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,4096,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,5120,0.0177528891298506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,6144,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,4096,0.018610666195551556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,128,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,5120,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,1024,0.017634666628307767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,512,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,1024,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,128,0.01386666629049513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,5120,0.019711999429596793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,51200,0.10873688591851129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,6144,0.012736000120639801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,4096,0.04155644443300035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,1536,0.019880889190567862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,256,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,768,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,2048,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,4096,0.2901822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,512,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,2560,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,10240,0.08947022093666925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,3072,0.016886222693655226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,3584,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,10240,0.012084444363911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,65536,5.319097730848524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,16384,0.1443155606587728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,7168,2.4982872009277344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,7168,0.010853333605660332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,51200,0.11977956030103896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,128,0.03001333276430766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,12288,0.09487289190292358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,10240,0.021290666527218286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,1536,0.061231997278001576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,51200,0.1531182262632582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,2048,0.07141066922081842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,512,0.03867466582192315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,12288,0.2877279917399089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,4096,0.029135998752382066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,1536,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,16384,0.47261868582831484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,10240,0.034041778908835515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,512,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,10240,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,1536,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,12288,0.04498399959670173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,10240,0.012832000023788877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,7168,0.05434933304786682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,2048,0.018464888135592144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,16384,0.05403288867738512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,6144,0.01330311099688212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,6144,0.022781333989567224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,10240,0.029448890023761328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,7168,0.020632889535692003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,1536,0.031186666753556993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,4096,0.016346666547987197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,2560,0.03642933236228095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,2560,0.021720889541837905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,5120,0.02534488836924235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,65536,0.24396355946858725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,32,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,8192,0.011695110963450538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,2048,0.0471093323495653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,768,0.025228444072935317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,768,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,2048,0.006653333289755716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,65536,0.7458595699734158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,7168,0.013930666777822705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,65536,0.2514977720048693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,1024,0.0640231106016371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,256,0.006446222050322427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,4096,0.028806222809685603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,65536,0.05645777781804403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,2560,0.0469724436601003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,16384,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,10240,0.15390400091807047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,51200,0.11616710821787517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,16384,0.02604444490538703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,12288,0.13099288940429688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,3072,0.05102577805519104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,1536,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,1024,0.003391111062632667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,10240,0.018345778187115986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,2560,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,51200,0.21043733755747476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,51200,0.44164178106519913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,768,0.012894222305880653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,7168,0.05300177799330818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,768,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,12288,0.22476532724168566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,6144,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,8192,0.18853778309292266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,8192,0.7958266470167371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,3072,0.04200444287723965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,5120,0.008786666724416945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,12288,0.07413778040144178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,128,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,6144,0.009460444251696268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,256,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,6144,0.09039200014538235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,12288,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,5120,0.17870399687025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,10240,0.027935110860400732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,1024,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,8192,0.018051554759343464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,128,0.020896888441509668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,768,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,8192,0.015077332655588785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,128,0.06791466474533081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,7168,0.04418222109476725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,6144,0.02163644466135237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,3072,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,1024,0.013278221918476952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,512,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,3584,0.06827466355429755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,65536,0.19642489486270479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,6144,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,1024,0.003175111073586676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,51200,0.8349582354227701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,16384,0.07390399773915608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,2560,0.011200889117187925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,512,0.011063999599880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,512,0.012780444489585029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,16384,0.03421866562631395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,1024,0.020834666159417894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,51200,1.2547439999050565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,65536,0.08459644185172187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,51200,0.09552355607350667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,256,0.0029360000044107437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,512,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,3584,0.04169955518510607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,2560,0.028346667687098186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,768,0.2519386609395345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,32,0.005525333185990651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,3072,0.017103110750516255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,6144,0.008954666554927826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,12288,0.027618666489919026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,3072,0.08516177866193984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,10240,0.07076266739103529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,12288,2.5853182474772134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,3072,0.04522222280502319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,2048,0.020446222689416673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,256,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,16384,0.03132266799608866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,5120,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,3072,0.07459200090832181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,8192,0.015403555499182807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,3072,0.20900889237721762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,4096,0.03219555483924018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,2560,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,64,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,6144,0.1306462287902832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,1024,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,12288,0.1584764454099867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,64,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,10240,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,768,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,51200,0.12798666954040527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,7168,0.013334222137928009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,256,0.018215111560291715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,1024,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,2048,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,2560,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,3072,0.04353244437111748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,7168,9.104873657226562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,2560,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,10240,0.02516533268822564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,4096,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,6144,0.030136889881557886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,16384,0.05654311180114746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,12288,0.05346044566896227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,8192,0.0471093323495653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,6144,0.02525955604182349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,1024,0.006329777754015393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,16384,0.19393600357903373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,3072,0.019688889384269714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,3072,0.023694222172101338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,6144,0.02796799937884013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,512,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,2560,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,12288,0.013688888814714221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,512,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,7168,0.46169598897298175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,7168,0.09495911333296035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,1536,0.006675555474228329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,1024,0.027196443743175928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,768,0.01962844365172916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,12288,0.26412712203131783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,6144,0.03289511137538486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,12288,0.02259644369284312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,7168,0.013331555657916598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,2048,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,4096,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,51200,0.23300888803270128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,8192,0.02439466615517934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,2048,0.04663644565476311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,512,0.009121778110663096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,4096,0.03291911217901442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,1536,0.03747377792994181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,64,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,65536,0.7415057818094889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,6144,0.034853332572513156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,6144,0.06239910920461019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,3584,0.04950933323966133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,2048,0.014191110928853353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,2560,0.026663111315833196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,1024,0.0038444445365005066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,3584,0.04077422287729051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,6144,0.021407999926143225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,65536,0.251839107937283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,768,0.038768000072903104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,10240,0.031530665026770696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,768,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,6144,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,65536,16.21879408094618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,16384,0.04380888740221659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,8192,0.04611733224656847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,1536,0.21766666571299234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,1024,0.066849774784512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,1536,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,10240,0.04126311010784573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,16384,0.05607199668884277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,4096,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,7168,0.024988444315062627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,8192,0.22826843791537813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,2048,0.017893332574102614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,65536,0.05877688858244154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,1024,0.018153778380817838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,4096,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,16384,0.08200711011886597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,512,0.14120354917314318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,3584,0.02899822261598375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,5120,0.017533333765135873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,7168,0.04032444291644626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,51200,0.0609475572903951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,12288,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,6144,0.19576088587443033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,2048,0.019900444481107924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,128,0.01459822224246131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,12288,0.02566577825281355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,7168,0.026161778304311965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,5120,0.024538666009902954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,1536,0.005045333256324132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,8192,0.03276622295379639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,1024,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,4096,0.045329779386520386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,7168,0.13284088505638972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,5120,0.013175110850069257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,256,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,5120,0.0103511114915212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,51200,0.3681200080447727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,8192,0.027786665492587622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,7168,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,1024,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,3072,0.0074622225430276655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,768,0.012459555433856117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,8192,0.5388506783379449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,7168,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,1024,0.021759111020300124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,16384,0.010955555571450127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,8192,0.04097244474622939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,3584,0.012376888758606382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,2560,0.08372000190946792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,4096,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,7168,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,128,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,64,0.0022124445272816554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,3584,0.023447111248970032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,51200,0.09527377949820624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,1024,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,3072,0.02238399949338701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,16384,0.48197513156467015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,2048,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,16384,0.23898310131496855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,32,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,768,0.016582222448454965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,1024,0.046874665551715426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,8192,0.667534245385064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,8192,0.05440088775422838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,5120,0.05216266711552938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,2560,0.028894222444958154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,512,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,5120,0.627886242336697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,2560,0.02232622272438473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,64,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,12288,0.016583111551072862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,1024,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,256,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,6144,0.15813599692450628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,1536,0.019998222589492798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,1024,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,3072,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,4096,0.008389333056079017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,4096,0.01978400018480089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,16384,0.05978577666812473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,768,0.043916443983713783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,128,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,10240,0.028782222006056044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,6144,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,3584,0.07423822085062663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,1024,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,3584,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,3584,0.004849777867396672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,256,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,10240,0.021947556071811255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,2048,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,51200,0.2870604462093777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,768,0.003216888962520493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,768,0.015863110621770222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,5120,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,1024,0.0468426677915785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,12288,0.926820437113444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,768,0.029473778274324205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,5120,0.022039999564488728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,256,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,65536,0.33959823184543186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,65536,0.03707289033465915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,8192,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,12288,0.4219813346862793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,4096,0.07569955454932319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,12288,0.08254044585757785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,3072,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,3072,0.174599117702908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,8192,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,4096,0.03897422220971849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,12288,0.11750843789842393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,8192,0.02343911098109351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,4096,0.8530666563245984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,6144,0.03926755653487311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,3584,0.011876444849703046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,10240,0.013425778183672162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,5120,0.016007110476493835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,5120,0.028372443384594385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,1024,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,16384,0.059247109625074595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,16384,0.025897777742809717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,1024,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,65536,4.265462239583333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,512,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,512,0.006067555397748947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,64,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,3584,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,3072,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,1536,0.003567999849716822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,3584,0.041061331828435264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,16384,0.6631431049770778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,4096,0.05823466512891981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,4096,1.0955786175198026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,3584,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,65536,0.4944044219122992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,12288,0.03881333271662394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,3072,0.21220000584920248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,6144,0.29701778623792857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,65536,0.10671644740634495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,16384,0.07292266686757405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,65536,2.0488177405463324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,256,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,51200,0.20850577619340685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,10240,0.0688355565071106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,12288,0.0602933300866021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,3072,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,5120,0.034726222356160484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,512,0.016156444946924847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,10240,0.26697156164381236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,5120,0.11398577690124512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,1024,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,768,0.01127822200457255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,8192,0.026405332816971674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,6144,0.2123511102464464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,2560,0.004867555366622077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,128,0.02697155541843838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,256,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,51200,0.09745511081483628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,3072,0.05432800120777554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,51200,0.12268000178866917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,8192,0.19725422064463297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,256,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,1536,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,3072,0.021026665965716045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,5120,0.09968088732825385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,6144,0.033152000771628484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,128,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,51200,0.7647813161214193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,51200,1.001542197333442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,1024,0.04088622331619263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,6144,0.013516444298956128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,1536,0.0498577786816491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,3072,0.02132622235351139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,1536,0.03877422213554382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,128,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,65536,32.003255208333336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,32,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,1024,0.01776977711253696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,512,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,7168,0.04531110988722908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,7168,0.02364088926050398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,512,0.009860444400045607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,7168,0.13190755579206678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,51200,0.1572479936811659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,4096,0.07923555374145508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,64,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,1536,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,3584,0.028961777687072754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,8192,0.03848177856869168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,1536,0.013786666923099093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,7168,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,51200,0.05093333456251356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,768,0.016498666670587327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,7168,0.7998426755269369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,12288,0.02516266703605652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,1024,0.025054223007626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,32,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,256,0.012959111067983838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,8192,0.020279111133681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,8192,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,256,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,1024,0.23856533898247612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,3584,0.02186577849917942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,2560,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,768,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,256,0.01109955542617374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,6144,0.2472942140367296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,768,0.016040888097551134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,3584,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,12288,1.9143636491563585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,128,0.07050044669045342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,64,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,2048,0.008441777692900764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,1024,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,2048,0.013033777475357056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,5120,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,6144,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,4096,0.010247111320495605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,12288,0.12836000654432508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,65536,0.4514471160040961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,512,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,16384,0.0786000026596917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,128,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,16384,0.45253419876098633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,256,0.019313777486483257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,3584,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,32,0.005401777724424998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,512,0.0038240001433425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,2048,0.02198666665289137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,16384,0.29685955577426487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,1536,0.006055111272467508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,16384,0.009722666607962714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,6144,0.012968888713253869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,6144,0.030224889516830444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,3072,0.03158044483926561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,16384,0.5974355803595649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,16384,0.028005331754684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,256,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,7168,0.04619110955132378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,768,0.015435554915004306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,6144,0.029050668080647785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,12288,0.016167999969588388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,1024,0.02166044380929735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,5120,0.07487289110819499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,3584,0.01372977760103014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,7168,0.01514933341079288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,7168,0.05879466401206123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,1536,0.28256267971462673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,12288,2.4570630391438804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,6144,0.01351377781894472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,5120,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,51200,0.05225333240297106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,2560,0.008177777959240807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,16384,0.3824942111968994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,256,0.013631111217869652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,51200,0.11110044187969631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,1024,0.04888711041874356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,3072,0.011002666420406766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,2560,0.09365777836905585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,1536,0.027642667293548584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,8192,0.017826666434605915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,4096,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,4096,0.04855288730727302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,5120,0.012039110892348819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,3584,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,4096,1.3598986731635199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,5120,0.08712444702784221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,3072,0.20472355683644614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,16384,0.019868445065286424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,768,0.03192799952295091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,8192,0.0196079992585712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,32,0.006011555592219035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,768,0.04241066508822971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,128,0.0019315555691719055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,8192,0.021839999490314062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,7168,0.06810844606823392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,512,0.008568888737095727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,1024,0.016470222009552848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,4096,0.006215111249023014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,1024,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,7168,0.1532355546951294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,256,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,2048,0.006554666492674086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,3584,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,65536,0.12408444616529678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,12288,0.1893813345167372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,512,0.014956444501876831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,2048,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,32,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,6144,0.09052089187833999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,1536,0.010764444039927589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,1024,0.015903110305468243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,10240,0.016089777151743572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,3584,0.019512888458040025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,4096,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,1536,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,4096,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,3072,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,16384,0.016082665986484952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,2048,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,768,0.017697778013017442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,65536,0.21659199396769205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,5120,0.027354665928416785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,5120,0.008897778060701158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,12288,0.14427910910712347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,65536,0.08444888724221124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,1536,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,4096,0.07582310835520427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,256,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,1024,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,64,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,2048,0.08859822485182021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,512,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,768,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,65536,0.8090915679931641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,7168,0.01570133368174235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,12288,0.19705422719319662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,8192,0.20719288455115425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,4096,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,1024,0.00314666661951277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,2048,0.0236835562520557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,7168,0.09461777740054661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,256,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,65536,0.06244977977540758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,6144,0.023786667320463393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,128,0.16448533535003662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,2048,0.024527110987239416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,1536,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,7168,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,6144,2.4037662082248263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,2560,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,16384,0.02885155545340644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,4096,0.025470222036043804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,2560,0.012120889292822944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,32,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,5120,0.025662221842341956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,65536,0.17643733819325766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,32,0.004853333450025982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,512,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,256,0.37588000297546387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,128,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,256,0.06049866808785332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,5120,0.04156088829040527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,7168,0.08492177724838257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,512,0.01899466746383243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,3072,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,256,0.03891999853981866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,2048,0.06724177466498481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,5120,0.012642666697502136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,3584,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,3072,0.026399999856948853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,8192,0.011411555939250521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,65536,0.14516711235046387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,2048,0.021887999441888597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,768,0.0761715571085612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,256,0.003559999995761447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,7168,0.08425333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,6144,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,12288,0.10157422224680583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,1024,0.018765333626005385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,128,0.07823999722798665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,128,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,1024,0.02148088812828064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,65536,0.024005333582560223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,32,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,5120,0.11049333545896743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,2560,0.08830844693713719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,512,0.003995555556482739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,128,0.00685511115524504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,65536,1.6796578301323786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,256,0.0034977777136696708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,10240,0.3361457718743219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,6144,0.011300444602966309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,3072,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,65536,0.19242933061387804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,768,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,12288,0.06855199734369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,256,0.041800889703962535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,51200,0.08625866969426473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,65536,0.10659022463692559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,65536,0.16272089216444227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,5120,0.012396444049146442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,65536,0.20206844806671143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,2048,0.015026667051845126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,10240,0.05754488706588745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,7168,0.0737173358599345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,1024,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,512,0.020346666375796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,4096,0.008734222087595198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,12288,0.05347377724117703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,64,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,5120,0.09890577528211807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,6144,0.013723555538389417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,3072,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,1536,0.022361778550677832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,16384,0.1319626702202691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,7168,0.013776000175211163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,2560,0.06087377998563978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,128,0.019559999306996662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,1536,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,2560,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,6144,0.046684443950653076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,2560,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,7168,0.28572267956203884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,32,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,768,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,3072,0.008402666283978356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,16384,0.2849831051296658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,1536,0.022335999541812476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,2048,0.00943288869327969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,2560,0.008763555851247575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,512,0.015617777903874716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,5120,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,64,0.002571555475393931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,1024,0.05512711074617174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,768,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,128,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,65536,0.7961351076761881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,65536,1.8272791968451605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,3584,0.016574222180578444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,1536,0.013741333451535968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,1024,0.02163022259871165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,10240,0.01330844478474723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,1024,0.007148444652557373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,3584,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,256,0.0057946667075157166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,10240,0.0536444452073839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,1024,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,51200,2.182243559095595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,8192,0.018929777873886954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,4096,0.008445333275530074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,5120,0.012311999996503195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,256,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,1536,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,3584,0.014296000202496847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,512,0.014639111028777229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,768,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,3584,0.06318755282296075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,2560,0.14313599798414442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,10240,0.026351999905374315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,16384,0.027445332871543035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,12288,0.06740799877378675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,10240,0.025731555289692346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,6144,0.05867822302712334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,7168,0.1043742232852512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,51200,0.8561013539632162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,10240,0.06019733349482218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,4096,0.06764800018734403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,51200,0.014155555102560254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,4096,0.08518577946556939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,65536,0.08824711375766331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,1536,0.0032248888164758682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,256,0.013969777358902825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,2560,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,512,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,16384,0.00926755534278022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,3584,0.146905779838562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,1024,0.020557334025700886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,51200,0.17315288384755453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,12288,0.6944764455159506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,1024,0.2017813391155667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,2048,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,65536,0.0864293310377333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,7168,0.09942489200168186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,768,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,5120,0.11661333507961696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,65536,3.9098862542046438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,51200,0.06230044364929199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,3072,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,768,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,2560,0.021346666746669348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,65536,0.09569778045018514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,128,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,768,0.05860444572236803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,6144,0.03273777829276191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,12288,0.018279999494552612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,10240,0.5101582209269205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,5120,0.046725332736968994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,1536,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,8192,0.010841777755154504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,8192,0.03276622295379639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,5120,0.027405333187845018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,32,0.003842666745185852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,512,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,128,0.0031395554542541504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,2560,0.01162755572133594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,128,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,2560,2.100191116333008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,4096,0.030080888006422255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,5120,0.10120089186562432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,1024,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,6144,0.34806132316589355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,51200,0.0417795545525021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,4096,0.023192000057962205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,6144,0.20095377498202852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,6144,0.0243857784403695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,6144,0.02331200076474084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,2048,0.17949333455827501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,768,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,128,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,3072,0.2401804394192166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,3072,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,256,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,1536,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,5120,0.023012444376945496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,12288,0.12057599756452773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,8192,0.01314755529165268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,4096,0.17640444967481825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,16384,0.009266667068004608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,64,0.0052968888647026485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,2560,0.0046968890560997855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,7168,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,128,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,2560,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,3584,0.0332915551132626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,6144,0.018246221873495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,7168,0.014303111367755465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,2048,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,3072,0.022696000006463792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,3584,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,12288,0.2913297812143962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,4096,0.008955555657545725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,5120,0.10707822110917832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,1024,0.05609333515167236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,768,0.0029333333174387612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,2048,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,5120,0.21156356069776747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,16384,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,8192,0.016394666499561734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,7168,0.021567111214001972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,65536,0.26768000920613605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,2560,0.025404444999165002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,1536,0.01642577846844991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,6144,0.011191999746693505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,128,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,128,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,6144,0.09920533498128255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,1536,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,8192,0.014723555909262763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,12288,0.05885244740380181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,1024,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,51200,0.1489457819196913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,10240,0.018768888380792405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,8192,0.258499542872111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,8192,0.43921777937147355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,256,0.022703111171722412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,3072,0.5709297921922472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,3584,0.042471110820770264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,5120,0.024097777075237695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,8192,0.023800000548362732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,128,0.015448888142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,3072,0.028041776683595445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,2560,0.08740533060497707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,12288,0.35082043541802305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,10240,0.025153777665562097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,10240,0.010118222071064843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,10240,0.11813333299424912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,3584,0.029021332661310833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,51200,20.19446478949653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,256,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,5120,0.03190844588809543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,4096,0.09790933132171631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,8192,0.056313779619004994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,12288,0.07027377684911092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,7168,0.057160000006357826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,65536,0.5963359938727485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,3072,0.007342221836249034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,512,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,2048,0.028038223584493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,4096,0.009386666946940953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,2560,0.007436444362004598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,512,0.021980444590250652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,512,0.0036071112586392295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,128,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,3584,0.09189422263039483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,2048,0.04110844598876105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,512,0.010319111247857412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,128,0.010407999985747868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,1024,0.019916445016860962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,6144,0.010266666611035665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,3584,0.025423111187087163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,7168,0.21231022146013048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,12288,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,7168,0.014350222216712104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,6144,0.03063911199569702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,2560,0.009062222308582729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,3072,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,1024,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,65536,0.1510595613055759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,8192,0.0777768890062968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,65536,0.06546844376458062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,7168,0.08343022399478489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,2560,0.03566577699449327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,65536,0.06276444594065349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,5120,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,32,0.00906755526860555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,3072,0.010240000155236986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,7168,0.1905724472469754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,3584,0.08039200305938721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,65536,0.09164533350202773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,256,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,7168,0.06856355402204725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,256,0.01408622165520986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,3584,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,256,0.0022053333620230355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,10240,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,512,0.005636444522274866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,256,0.058818664815690785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,1024,0.02184355590078566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,4096,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,10240,0.01291911138428582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,6144,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,10240,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,2048,0.009172444542249044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,12288,0.0503422220547994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,1536,0.033952001068327166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,2560,0.012264000044928657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,10240,0.015199111567603217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,5120,0.46997690200805664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,3072,0.09261244535446167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,51200,9.705165439181858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,768,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,12288,0.06999822457631429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,2560,0.0671164459652371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,1536,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,768,0.00997688869635264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,4096,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,4096,0.02444177700413598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,2048,0.019281778070661757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,7168,0.03187822302182516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,3584,0.008567111359702216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,7168,0.15801066822475857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,768,0.057167112827301025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,3584,0.018246221873495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,128,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,16384,0.027060444156328838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,10240,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,1536,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,2048,0.0414648883872562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,32,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,3584,0.025295999315049913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,5120,0.03859022259712219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,3072,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,3584,0.04157333241568671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,7168,0.06213866339789497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,6144,0.024676443801985845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,6144,0.021588444709777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,1536,1.2741582658555772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,3584,0.043057776159710355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,7168,0.04957155717743767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,51200,0.14016444153255886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,4096,0.01995822290579478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,5120,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,65536,0.7322604391309949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,8192,0.012477333347002665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,7168,0.09770399994320339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,256,0.008419555922349295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,256,0.027042667071024578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,1024,0.020845333735148113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,12288,0.11139555772145589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,1024,0.006011555592219035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,512,0.02625688910484314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,128,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,3584,0.06269777483410306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,768,0.004239999999602635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,2560,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,2560,0.24139200316535103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,65536,0.05840355820126004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,3584,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,1024,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,10240,0.03800088829464383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,3584,0.020614221692085266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,64,0.002516444358560774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,7168,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,51200,0.05364711085955302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,2560,0.0750675532552931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,1024,0.015095111396577625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,5120,0.11215555667877197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,1536,0.04878488845295376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,8192,0.010743111371994019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,10240,0.016020443704393175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,128,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,128,0.37839200761583114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,7168,0.15682133038838705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,7168,0.07278488741980658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,5120,0.033556444777382746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,512,0.008032889001899296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,16384,0.022304889228608873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,2560,0.038747555679745145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,8192,0.04535911149448819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,32,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,512,0.03883822096718682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,1536,0.04398399922582838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,1024,0.007354666789372762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,10240,0.12314755386776394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,8192,0.0741893317964342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,12288,0.032568001084857516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,1536,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,16384,0.3412337832980686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,4096,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,128,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,3584,0.03605955508020189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,16384,0.020106666617923312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,32,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,768,0.00328177772462368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,32,0.010442666709423065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,32,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,10240,0.05974044402440389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,65536,0.11026133431328668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,51200,0.05288889010747274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,16384,0.040566222535239324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,64,0.0034968890249729156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,4096,0.039136889908048846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,32,0.023008000519540574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,8192,0.042769776450263135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,4096,0.02420355545149909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,128,0.020224889119466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,128,0.011712888876597086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,5120,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,512,0.01279022213485506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,2560,0.005431111074156231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,3072,0.023670223024156358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,51200,0.05127999848789639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,2560,0.04064711266093784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,8192,0.07788088586595324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,10240,0.09684711032443577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,1536,0.019508444600635104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,7168,0.12485955821143256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,3584,0.4001893202463786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,16384,0.022312000393867493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,6144,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,2560,0.011376000112957425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,2560,0.049385776122411094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,7168,0.012906666431162091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,16384,0.021506667137145996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,128,0.004938666605287128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,256,0.014945778581831189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,3584,0.02786933382352193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,8192,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,2048,0.11515111393398708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,768,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,10240,0.043931557072533496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,10240,0.14293866687350804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,2560,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,6144,0.05702933337953356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,2048,0.034049779176712036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,3072,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,2560,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,512,0.030582222673628066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,12288,0.03577155537075467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,51200,0.024163555767801072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,3584,0.01461155547036065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,6144,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,4096,0.20581155353122285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,1024,0.021307556165589228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,3072,0.05541777941915724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,3584,0.05647377835379707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,768,0.026729777455329895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,128,0.0018888889915413326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,1536,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,51200,0.08494666549894546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,2048,0.03400177756945292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,768,0.020278222031063504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,3584,0.0840266678068373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,10240,0.0889679988225301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,128,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,512,0.01756088932355245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,512,0.01716444392999013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,12288,0.016697777642144095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,51200,0.29475556479560006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,65536,0.07788355482949151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,10240,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,65536,0.07829155524571736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,2048,0.22631467713250053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,512,0.021262221866183813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,256,0.007490666376219855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,51200,0.9128124448988172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,65536,0.30819111400180393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,2048,0.031170666217803955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,10240,0.18274311224619547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,65536,0.20435110727945963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,2048,0.05659911367628309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,6144,0.11681155363718669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,6144,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,128,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,768,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,1536,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,4096,0.0270444436205758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,4096,0.2567608886294895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,16384,0.08557066652509902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,4096,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,6144,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,512,0.022011554903454248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,128,0.015096000499195524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,7168,0.20009333557552764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,6144,0.09608178006278144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,128,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,6144,0.07395288679334852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,768,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,2048,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,51200,0.05365511112742954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,5120,0.3307369020250109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,7168,0.04937244455019633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,768,0.12103288703494602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,128,0.011420444481902652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,1024,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,3584,0.028042667441897925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,3584,0.0074240002367231585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,2048,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,16384,0.25404267840915257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,1024,0.012166221936543783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,3072,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,128,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,16384,0.16474844349755183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,2048,0.004585777719815572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,2048,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,512,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,8192,0.16592355569203696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,51200,0.04480888777308994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,7168,0.05226933293872409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,7168,0.01812000075976054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,8192,0.06377689043680827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,128,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,65536,0.05560533205668131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,51200,0.38374310069613987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,1024,0.008047999607192146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,256,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,7168,0.019273777802785236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,3072,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,10240,0.030016889174779255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,64,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,10240,0.1357777780956692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,768,0.46013689041137695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,2560,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,256,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,2560,0.018937778141763475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,2048,0.0232924438185162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,1536,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,51200,0.06138222085105049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,16384,0.11894399589962429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,6144,0.06588266955481635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,10240,0.021423111359278362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,16384,0.12079111735026042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,5120,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,768,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,3584,0.15420356061723497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,5120,0.03397599856058756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,7168,0.15219644705454508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,3584,0.07260355684492323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,51200,0.10229866372214423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,128,0.03211733367707994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,16384,0.020224000016848247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,512,0.025232889586024817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,3584,0.06754577822155423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,16384,0.015953777564896476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,768,0.008700444466537898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,16384,0.01349244432316886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,1024,0.01127644462717904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,10240,0.20336443848080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,1536,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,768,0.012772444221708508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,2048,0.028541333145565454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,768,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,1536,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,1536,0.022668444448047217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,10240,0.21825777159796822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,12288,0.04916622241338094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,4096,0.030784888399971858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,768,0.009069333473841349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,64,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,1536,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,6144,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,6144,0.014206222361988492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,128,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,128,0.010334221853150262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,2560,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,7168,0.02810933192571004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,16384,0.030588444736268785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,512,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,7168,0.05981600284576416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,65536,0.6087697876824273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,2048,0.31953954696655273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,16384,0.22066666020287407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,256,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,16384,2.5904354519314237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,6144,0.2021226618025038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,1024,0.040969779094060264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,8192,0.04050666756100125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,7168,0.020592888196309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,10240,0.017456889152526855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,4096,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,2560,0.03152711192766825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,10240,1.3590479956732855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,3584,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,5120,0.017120000388887193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,4096,0.05415199862586128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,512,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,8192,0.15885867012871638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,6144,0.17603022522396514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,7168,0.010778666370444827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,64,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,65536,0.06717333528730604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,3584,0.06888266404469807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,256,0.025409777959187824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,5120,0.025418667329682246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,3584,0.028860443168216284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,768,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,128,0.016240888171725802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,2048,0.1930373377270169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,512,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,10240,3.32004886203342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,768,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,1024,0.019316444794336956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,5120,8.211745368109808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,12288,0.07760889000362821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,512,0.026014222039116755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,4096,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,768,0.008046222229798635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,32,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,256,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,8192,0.04929511083496941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,8192,0.0971475574705336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,7168,0.02777333392037286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,6144,0.030272000365787085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,4096,0.00960533320903778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,32,0.005707555760939916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,5120,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,51200,0.06639022297329374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,1536,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,65536,0.14983999729156494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,512,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,768,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,16384,4.913416968451606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,65536,0.10347021950615777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,8192,0.04121422105365329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,768,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,5120,0.028492444091373022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,12288,0.02643022272321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,3072,0.010141332944234213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,5120,0.04383733206325107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,1536,0.021145777569876775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,7168,0.12961244583129883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,32,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,128,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,2560,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,256,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,768,0.014783999986118741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,16384,0.01197866681549284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,6144,0.031353778309292264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,16384,0.02860355708334181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,768,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,64,0.003855111284388436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,51200,0.009526222116417354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,10240,0.037045333120558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,1536,0.015215999550289579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,16384,0.03810311026043362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,65536,0.21671822335984972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,2048,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,1024,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,51200,0.04996000064743889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,2048,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,256,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,64,0.001975111042459806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,65536,0.2931164370642768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,8192,0.25354666180080837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,4096,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,4096,0.03161333335770501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,1024,0.04325422313478258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,16384,0.06508266925811768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,8192,0.21998577647738984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,4096,0.019317333896954853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,256,0.005495999836259418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,6144,0.004687110996908612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,10240,0.2324319945441352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,3584,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,2048,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,12288,0.017197334104114108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,6144,0.026406221919589575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,1024,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,3584,0.016798221402698092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,4096,1.2741591135660808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,4096,0.014967110421922473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,7168,0.021292444732454088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,32,0.01238844460911221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,7168,0.16507911682128906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,256,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,12288,0.2823111216227214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,128,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,128,0.0024071110205517872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,6144,0.08472088972727458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,512,0.0214675549003813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,5120,0.03900711072815789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,32,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,5120,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,2048,0.022847111026446026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,4096,0.02304711110062069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,65536,3.0132355160183377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,8192,0.562197314368354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,1024,0.005046222358942032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,512,0.017166222135225933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,512,0.020658666888872784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,3584,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,3072,0.07481866412692599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,4096,0.026143999563323125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,2048,0.020678222179412842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,3072,0.011198221809334226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,51200,0.1398799949222141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,3072,0.028205331828859117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,32,0.0057760001056724125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,3584,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,7168,0.07264355818430583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,6144,0.625656869676378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,4096,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,12288,0.0962382223871019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,1536,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,10240,0.1063422229554918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,65536,0.10180799828635322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,128,0.01386844449573093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,2560,0.012216000093354119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,1536,0.05667733483844333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,256,0.0028924445311228433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,512,0.010493333141009012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,8192,0.07005244493484497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,128,0.12367110782199436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,64,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,4096,0.010832889212502373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,1536,0.02637422250376807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,12288,0.06482844220267402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,12288,0.16606667306688097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,256,0.024735111329290602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,16384,0.06166488594479031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,6144,0.016879111528396606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,12288,0.013764444324705334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,2048,0.006456888798210356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,32,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,51200,0.4379111131032308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,32,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,128,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,3072,0.007366221812036302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,10240,0.06320266591178046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,1536,0.03991466760635376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,65536,2.4983022477891708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,51200,0.0652666687965393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,8192,0.0827031135559082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,256,0.02889866630236308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,16384,0.8434631029764811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,32,0.00813244448767768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,7168,0.014408888088332282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,12288,0.1470764478047689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,32,0.002007999974820349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,6144,0.009569777382744683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,10240,0.030329777134789362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,1024,0.004493333399295807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,768,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,16384,0.017363554901546903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,128,0.027976890405019123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,8192,0.02984444300333659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,7168,0.13008444839053684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,5120,0.08976178036795722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,64,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,6144,0.017519111434618633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,3584,0.0074755557709270054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,8192,0.023730667101012334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,3072,0.05770311090681288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,5120,0.00850399997499254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,8192,1.4544817606608074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,768,0.009384888741705153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,8192,0.014416888356208801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,12288,0.012113778127564324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,51200,0.48612176047431094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,2560,0.06272711356480916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,12288,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,4096,0.01590399940808614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,4096,0.02608266638384925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,10240,0.011968000067604912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,8192,0.14333689212799072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,65536,0.049365331729253135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,10240,0.020229332976871066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,7168,0.12557689348856607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,3072,0.339413324991862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,4096,0.010283555421564314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,1536,0.040236443281173706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,128,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,2560,0.0075253330998950535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,6144,0.013521778086821238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,768,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,32,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,2048,0.011791999969217511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,2560,0.008566222257084316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,12288,0.028374221589830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,3072,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,5120,0.02013777858681149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,64,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,768,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,2048,0.02308977809217241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,3584,0.05711466736263699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,7168,0.027112000518374976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,6144,0.03664711117744446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,128,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,768,0.006679111056857639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,8192,0.0800328916973538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,1024,0.018833777970737882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,5120,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,256,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,6144,0.8296151161193848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,32,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,10240,0.09809244341320461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,16384,0.3261146545410156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,1536,0.050224890311559044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,2560,0.02642400066057841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,256,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,2560,0.08111200067732069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,16384,0.025292444560262892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,6144,0.019464888506465487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,1536,0.017470222380426195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,2048,0.017715555098321702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,5120,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,2560,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,7168,0.010837333069907295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,51200,2.817393832736545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,32,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,5120,0.05351466602749295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,51200,0.105785780482822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,8192,0.1630328893661499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,1536,0.0046880000995265115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,1536,0.04554311103290982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,64,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,512,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,7168,0.11604444185892741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,256,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,768,0.003698666476541095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,5120,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,51200,0.19089333216349283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,2560,0.004893333133724001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,65536,0.14088711473676893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,4096,0.01144177797767851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,8192,0.06795022222730848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,512,0.006536888993448681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,1536,0.02588266630967458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,12288,0.01702844434314304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,16384,0.05527466535568237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,2560,0.01498044364982181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,2560,0.01717866626050737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,5120,0.025705777936511572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,2048,0.025045333637131587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,256,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,16384,0.05117688907517327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,3584,0.1282657782236735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,8192,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,2048,0.5233404371473525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,2560,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,4096,0.6320026715596517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,768,0.003195555466744635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,6144,0.08348533180024888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,5120,0.01956888867749108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,51200,0.11224088403913711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,65536,0.176639993985494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,4096,0.7015422185262045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,768,0.03517599900563558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,8192,0.016521778371598985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,2048,0.032178666856553816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,2560,0.0388408899307251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,512,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,6144,0.08103022310468885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,128,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,512,0.00849422232972251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,256,0.014526221487257214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,5120,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,1024,0.0035048888789282907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,2048,0.1073404418097602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,4096,0.08462578058242798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,3584,0.012016889121797351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,32,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,5120,0.12850666046142578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,3072,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,7168,0.05080177717738681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,128,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,10240,0.012446222205956778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,32,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,2048,0.03149155444569058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,1024,0.00871733327706655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,512,0.02511288887924618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,10240,0.0415635539425744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,65536,2.76963742574056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,1536,0.31028710471259224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,65536,0.18182044559054902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,3584,0.018526222970750596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,7168,1.100404421488444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,6144,0.03225333491961161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,5120,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,51200,0.31497422854105633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,1024,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,65536,0.26359465387132436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,12288,0.043493333790037364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,10240,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,7168,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,128,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,16384,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,1536,0.03716355562210083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,5120,0.021615111165576514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,512,0.00565422202150027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,512,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,256,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,128,0.009918221996890174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,768,0.017516444126764934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,7168,0.07315022415584989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,512,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,1536,0.01477688882086012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,12288,0.013421333498424955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,768,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,8192,0.12799733214908177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,16384,0.596884462568495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,3072,0.03018310997221205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,16384,0.023504889673656885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,65536,0.16944800482855904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,6144,0.03529777791765001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,4096,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,128,0.004584889031118817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,7168,0.11379822095235188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,5120,0.1659653319252862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,4096,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,10240,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,2048,0.05140977766778734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,16384,0.06259733438491821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,8192,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,65536,2.7283200158013234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,768,0.029049777322345312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,7168,0.42528711424933535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,65536,0.15714489089118108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,256,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,6144,0.05669689178466797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,2048,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,3072,0.3184524377187093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,8192,0.17496089140574136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,2560,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,2048,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,16384,0.10369422038396199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,4096,0.014614222778214348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,3584,0.05537155601713392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,7168,0.015542222393883599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,2048,0.024007999234729346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,2560,0.01143200033240848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,3584,0.03746577766206529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,1024,0.03053511182467143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,1536,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,1024,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,65536,0.07840444644292195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,8192,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,65536,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,1536,0.043112887276543506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,5120,0.07675555679533216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,768,0.01995022263791826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,2560,0.043175998661253184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,51200,0.14598488807678223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,64,0.029140445921156142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,6144,0.4527617560492621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,10240,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,64,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,8192,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,16384,0.012671999633312225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,1536,0.010608889162540436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,3072,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,6144,0.12450488408406575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,5120,0.004603555632962121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,2048,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,4096,0.3285715579986572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,256,0.005043555465009478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,1024,0.0033653332955307434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,65536,0.060270223352644176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,6144,0.01990133358372582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,4096,0.009140444298585257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,128,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,65536,0.21858489513397217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,5120,0.0844871136877272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,5120,0.008553778131802877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,3584,0.009484444227483537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,1024,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,512,0.012446222205956778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,3072,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,768,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,128,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,65536,0.06275644567277697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,12288,5.3770294189453125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,65536,1.0315146976047092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,128,0.018868444694413077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,1024,0.03894844320085313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,2560,0.12122221787770589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,16384,0.020922667450375024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,4096,0.03898222247759501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,4096,0.01180622229973475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,6144,0.015038222074508667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,12288,0.032124443186653986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,5120,0.018758222460746765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,256,0.019596444235907663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,7168,0.011760888828171624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,32,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,512,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,2048,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,3584,0.03763377666473389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,3584,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,2048,0.020982222424613103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,2560,0.047469334469901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,4096,0.023561777340041265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,2048,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,10240,0.22052621841430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,16384,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,51200,0.15871466530693903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,16384,0.027610666222042505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,65536,0.9557182523939344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,16384,0.10569689008924697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,65536,0.1451546616024441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,6144,0.01845066746075948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,10240,0.03413777881198459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,768,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,16384,0.5589404635959202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,2048,0.3257662190331353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,1536,0.006480000085300869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,7168,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,65536,0.05731733640034994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,7168,0.0665137767791748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,16384,0.2125084400177002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,3584,0.019997333486874897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,12288,0.08150133159425524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,5120,0.0096133334769143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,4096,0.018575999471876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,256,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,32,0.0022391111900409064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,2048,0.33287734455532497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,8192,0.005173333403136996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,6144,1.2542728847927518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,128,0.07811111211776733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,256,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,7168,0.009714666340086196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,2048,0.41963556077745223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,1536,0.011391110718250275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,768,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,12288,0.6378586557176378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,5120,0.3364391061994765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,768,0.008535111116038429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,12288,0.07189777824613783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,2048,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,512,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,768,0.0041680000722408295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,65536,0.35780800713433164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,1024,0.05307555529806349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,16384,0.034950223233964704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,256,0.01756622228357527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,5120,0.004247111164861255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,6144,0.023090667194790308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,2560,0.08290666341781616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,65536,0.1316817733976576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,6144,0.06972977850172254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,8192,0.028395556741290625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,64,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,4096,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,51200,0.09681599669986302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,51200,0.05064622229999966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,1536,0.0042160000238153665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,5120,0.019862223002645705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,2560,0.017283555534150865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,10240,0.024703111913469102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,1024,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,2560,0.02741866641574436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,2048,0.031727999448776245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,10240,0.18997332784864637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,3072,0.007676444119877285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,6144,0.017228444417317707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,768,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,8192,0.41239200698004824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,6144,0.020670221911536325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,2560,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,512,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,3072,0.11955289045969646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,768,0.01995733380317688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,512,0.003145777723855443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,1536,0.01817244456874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,512,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,10240,0.19343466228908965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,8192,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,768,0.030598223209381104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,256,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,2048,0.036284443404939436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,3584,0.10348800155851577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,10240,0.04214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,768,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,4096,0.23521155781216094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,1024,0.20250044928656685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,256,0.014914666612943014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,65536,0.33135732014973956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,4096,0.02254844374126858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,2048,0.08897066778606838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,12288,0.05106044477886624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,128,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,768,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,768,0.04474044508404202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,51200,3.7874357435438366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,51200,0.11776178412967259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,768,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,64,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,3584,0.008174222376611497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,4096,0.026025777061780293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,3584,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,2560,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,51200,0.21556711196899414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,1024,0.09185955259535049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,51200,0.08994577990637885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,6144,0.04625244604216682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,2560,0.09974310795466106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,16384,0.04434222314092848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,256,0.016887111796273124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,2048,0.03140799866782294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,7168,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,1024,0.006044444524579578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,128,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,3072,0.05778577592637804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,512,0.04884888728459676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,1536,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,512,0.01197866681549284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,2560,0.06989510854085286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,2048,0.04840355449252658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,2560,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,768,0.003240888938307762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,2048,0.020939555433061387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,2560,0.07464800278345744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,10240,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,2560,0.01926399932967292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,65536,0.7887946764628092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,65536,2.2100978427463107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,5120,0.01974311139848497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,65536,19.93428209092882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,32,0.03608533408906724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,2048,0.1772444380654229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,51200,0.1360213359196981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,128,0.02177511155605316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,2560,0.01016444464524587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,1536,0.02325422234005398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,8192,0.024746666351954143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,512,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,12288,0.08509244521458943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,8192,0.02127733329931895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,8192,0.030071109533309937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,16384,0.04487733377350701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,1536,0.025352888637118872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,65536,0.05412888858053419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,12288,0.2187502251731025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,7168,0.017941332525677152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,6144,0.013559111290507846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,4096,0.007896000312434303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,65536,0.05537688732147217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,4096,0.025450666745503742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,2048,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,128,0.0018719999740521114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,32,0.012334222594896952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,2048,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,512,0.011438222395049201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,4096,0.05233333177036709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,1536,0.01367111090156767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,16384,0.027746667464574177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,512,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,16384,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,65536,0.10662221908569336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,16384,0.2040062215593126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,51200,0.2473226653205024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,256,0.010459555519951714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,16384,0.14218578073713514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,512,0.011367999845080905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,51200,0.05570844146940443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,10240,0.026713778575261433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,128,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,51200,0.08393066459231907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,6144,0.04150222076310052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,3584,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,5120,0.017487999465730455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,128,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,51200,0.05939110782411364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,5120,0.025362667110231187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,8192,0.011333333121405708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,51200,0.05181599987877739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,1024,0.015757333901193406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,256,0.006045333213276333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,512,0.0035217776894569397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,768,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,16384,0.09665244155459934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,32,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,2560,0.03411644366052415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,512,0.0057591112951437635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,5120,0.009345778160625035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,256,0.0019360000474585427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,1024,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,3584,0.020602666669421725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,1024,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,128,0.011787555283970304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,4096,0.008837333156002892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,8192,0.02363377809524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,16384,0.045250667466057666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,5120,0.010827555424637265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,2048,0.022030222747060988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,2048,0.012539555629094442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,768,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,2048,0.01016266644001007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,1536,0.04817422231038412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,6144,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,512,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,4096,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,4096,0.13351555665334067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,768,0.013591999808947245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,3584,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,65536,1.064736048380534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,1536,0.01835644410716163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,4096,0.009142222503821055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,2048,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,1536,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,768,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,3072,0.07347910934024386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,65536,0.17882311344146729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,7168,0.018173333671357896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,6144,0.011911110745535957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,10240,0.012698666916953193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,128,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,16384,0.20364177227020264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,5120,0.3913884427812364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,256,0.006046222315894232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,128,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,51200,0.30893066194322377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,6144,0.020623111062579684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,256,0.002354666590690613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,256,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,768,0.018623110320832994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,2048,0.01254755589697096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,65536,0.08467288812001546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,3584,0.010117332968446944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,4096,0.37808977233039004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,256,0.01833511061138577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,256,0.021591999464564856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,2560,0.05944977866278755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,4096,0.14862310886383057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,8192,0.1475484503640069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,65536,0.06161066558625963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,10240,0.338539547390408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,32,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,7168,0.19946577813890246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,7168,0.03405688868628608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,16384,0.0788631108072069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,8192,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,16384,0.23997865782843697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,5120,0.012317332956526013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,3072,0.015432889262835184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,7168,0.013425778183672162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,768,0.015108444624476962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,7168,0.13543200492858887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,3584,0.13311378161112467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,16384,0.12210222085316975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,6144,0.3699511157141791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,512,0.0047111109726958806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,16384,0.09372266795900132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,6144,0.008160889148712158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,3072,0.017260443833139207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,256,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,10240,0.006656000183688269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,4096,0.007806222471925948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,65536,0.07676533195707533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,768,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,2560,0.03357955482270982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,5120,0.10579378075069851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,1536,0.011063999599880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,51200,0.0547599991162618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,16384,0.40008089277479386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,3072,0.027867555618286133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,512,0.10886400275760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,2048,0.06619822316699557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,4096,0.008708444734414419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,12288,0.039474666118621826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,5120,0.09679288996590509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,3584,0.05867555406358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,8192,0.019283556275897555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,2048,0.12113332748413086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,128,0.03311022122701009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,512,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,5120,0.036791112687852644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,64,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,51200,1.1913590961032443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,8192,0.015232000086042615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,8192,0.016714667280515034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,128,0.015511110424995422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,1024,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,1536,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,32,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,768,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,2560,0.007735110819339752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,12288,0.03698666559325324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,2048,0.027758222487237718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,128,0.09325422181023492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,4096,0.009497777455382878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,768,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,64,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,1536,0.03553422292073568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,3584,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,5120,0.00851288851764467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,16384,0.11935822168986003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,2048,0.037046220567491316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,512,0.0028702221396896574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,1024,0.019395555059115093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,2048,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,6144,0.05158933334880405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,3072,0.013655111193656921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,1536,0.3105813397301568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,1024,0.008361777497662438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,3584,0.08382666773266262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,128,0.01622577839427524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,64,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,2048,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,2048,0.004942222187916438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,64,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,2048,0.04073511229621039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,3584,0.017920888132519193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,7168,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,1536,0.013682666752073499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,51200,0.09194666809505886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,7168,0.015677332878112793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,256,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,1024,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,10240,0.0355697770913442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,7168,0.016597333881590102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,4096,0.0663377775086297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,51200,0.05378311210208469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,512,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,1024,0.01574222246805827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,6144,0.009458666874302758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,1024,0.009572444690598382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,2560,0.019927110936906602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,51200,6.8222766452365455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,512,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,10240,0.4224008984035916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,2560,0.039911998642815485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,51200,0.11911111407809788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,1536,0.022966222630606756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,64,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,3072,0.00792711145348019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,65536,0.036907556984159685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,5120,0.17620978090498182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,7168,0.02914488977856106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,12288,0.10318844185935126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,3584,0.004885333279768626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,1024,0.014680888917711047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,6144,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,16384,0.24625688129001191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,51200,0.7082373301188151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,65536,1.2269422743055556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,2560,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,4096,0.01884355478816562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,1536,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,5120,0.010189333723651038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,2048,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,1536,0.015071110592948066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,1536,0.021469333105617102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,4096,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,512,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,16384,0.020803555846214294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,2048,0.12836266888512507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,3584,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,6144,0.021686222818162706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,768,0.03805955582194858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,51200,1.0453093846638997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,12288,0.005193777796294955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,768,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,3584,0.013535999589496188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,4096,0.029355555772781372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,256,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,5120,0.01626311077011956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,2560,0.0543431109852261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,6144,0.16381688912709555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,1536,0.006679111056857639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,8192,0.010607110957304636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,2560,0.038655999634000994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,3584,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,512,0.021980444590250652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,7168,0.01620533400111728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,3072,0.009173333644866943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,7168,0.024892444411913555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,1536,0.004743111216359668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,128,0.007753777835104201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,8192,0.01571733256181081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,10240,0.04725066820780436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,512,0.005330666485759947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,12288,0.22880800565083823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,2048,0.01889866590499878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,3584,0.036200887627071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,6144,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,16384,0.017445334129863314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,2560,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,16384,0.3618355592091878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,2560,0.006823110911581252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,2560,0.023590222001075745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,3584,0.06315200196372138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,768,0.008933333059151968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,1536,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,2048,0.016142222616407607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,4096,0.03877422213554382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,2048,0.37916000684102374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,256,0.009935999910036722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,4096,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,16384,0.14811821778615317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,4096,0.006859555426571105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,1536,0.00978577799267239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,7168,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,256,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,4096,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,2560,0.013745778136783175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,7168,0.19528710842132568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,12288,0.04971377717124092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,768,0.018889778190188937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,5120,0.029683556821611192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,1024,0.015596444408098856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,4096,0.015232889188660515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,3072,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,1024,0.01308444473478529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,2560,0.02536711096763611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,7168,0.023231110639042322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,51200,8.116470336914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,6144,0.02865511178970337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,768,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,2048,0.02717777755525377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,4096,0.01927288870016734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,1536,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,2048,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,256,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,32,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,256,0.01753866672515869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,51200,1.5907004674275715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,16384,0.02608088817861345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,6144,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,7168,0.1063271098666721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,512,0.05662577682071262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,64,0.030410667260487873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,3584,0.00757244477669398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,128,0.015325332681337992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,8192,1.0939679675632052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,65536,0.16843999756707084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,7168,0.08280711041556464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,2048,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,3072,0.011130666567219628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,51200,0.3233937687344021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,768,0.005777777896987067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,5120,0.05717600054211087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,32,0.012279999752839407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,512,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,1536,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,1536,0.03206933206982083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,4096,0.07701510853237577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,51200,0.12663643889957005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,4096,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,7168,0.012107555237081317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,2048,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,5120,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,64,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,8192,0.07923911015192668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,6144,0.06975999805662367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,7168,0.0329751107427809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,256,0.005523555394675996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,128,0.010595555934641095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,10240,0.038450668255488075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,16384,0.02055022286044227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,1536,0.02567377852069007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,512,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,6144,0.021783999270863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,5120,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,768,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,3584,0.05649155378341675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,7168,0.010315555665228102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,3584,0.010996444357766045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,65536,13.874278598361544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,2048,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,1024,0.0075351107451650835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,32,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,2560,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,16384,0.12323644426133896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,1536,0.01940711173746321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,16384,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,16384,0.26594577895270455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,10240,3.1210178799099393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,51200,8.064609951443142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,1024,0.021348444951905146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,5120,0.047282665967941284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,51200,0.07112711005740695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,4096,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,5120,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,512,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,2560,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,2048,0.24028356870015463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,7168,0.00945688866906696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,7168,0.027031110392676458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,7168,0.037477334340413414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,10240,0.09683644771575928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,65536,0.06970755259195964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,64,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,768,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,51200,4.258861541748047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,128,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,768,0.03142666816711426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,768,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,6144,0.032616890139049955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,5120,0.01776355504989624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,16384,0.23267555236816406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,3072,0.00885333369175593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,128,0.012582222620646158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,256,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,3072,0.02743822170628442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,3072,0.17154133319854736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,8192,0.03370133373472426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,8192,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,8192,0.013704000247849358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,32,0.002605333303411802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,128,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,1536,0.1260062191221449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,256,0.009864889085292816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,256,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,16384,0.06490044461356269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,8192,0.04621866676542494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,768,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,1536,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,2560,0.04598666561974419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,65536,0.059828446971045605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,1536,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,1024,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,128,0.003564444267087512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,5120,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,1024,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,3072,0.010775999890433418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,2048,0.007338666253619724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,2048,0.025177778469191656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,256,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,10240,0.012061333490742577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,12288,0.05758844481574165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,2048,0.06241244739956326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,512,0.020748444729381137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,65536,0.0259253333012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,1536,0.03231199913554721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,5120,0.012708444562223224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,256,0.03406044509675767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,8192,0.010089777410030365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,128,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,12288,0.028581334484948054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,768,0.020610666937298246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,16384,0.1955866681204902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,65536,0.0785964462492201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,768,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,1024,0.018920888503392536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,4096,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,5120,0.019558222757445443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,1536,0.006550222221348021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,5120,0.036192887359195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,256,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,10240,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,3072,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,51200,0.7125884691874186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,1536,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,16384,0.046111110183927745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,1536,0.011217777927716574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,12288,0.01258933295806249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,4096,0.04892444279458788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,8192,0.33580711152818465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,3072,0.03851200143496195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,6144,0.004971555537647671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,5120,0.04713777701059977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,10240,0.040919999281565346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,6144,0.1171120007832845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,12288,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,1536,0.011112000379297467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,7168,0.14242843786875406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,32,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,16384,0.02127911150455475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,3584,0.05029866761631436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,128,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,51200,0.15818133619096544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,768,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,512,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,16384,0.16362399525112575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,5120,0.017321777012613084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,16384,0.0635075569152832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,5120,0.014843554960356818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,1024,0.025937777426507738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,8192,0.010858666565683154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,12288,0.03645955522855123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,65536,0.07495911253823175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,5120,0.10340355502234565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,128,0.02299555473857456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,256,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,16384,0.1459217733807034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,128,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,16384,0.016210666961140104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,1024,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,7168,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,7168,0.18123466438717314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,2560,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,1024,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,7168,0.29191822475857204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,6144,0.051474667257732816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,51200,0.03379822108480666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,3584,0.04958488874965244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,6144,0.07793244388368395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,512,0.0027991111079851785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,65536,6.981544070773655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,51200,0.11813333299424912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,6144,0.05382933219273885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,3584,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,2560,0.03975111246109009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,3072,0.08899022473229302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,2560,0.022952889402707417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,768,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,12288,0.030215111043718126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,256,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,2560,0.038246221012539335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,8192,0.1209377712673611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,768,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,10240,0.035312000248167254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,2048,0.0701191094186571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,5120,0.026177777184380427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,4096,0.02683911058637831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,16384,0.04271022147602505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,3584,0.015058666467666626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,256,0.015236443943447538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,128,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,1024,0.253074672487047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,3584,0.16016444895002577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,1024,0.015849777393870883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,5120,0.09678844610850017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,65536,0.07249599695205688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,5120,0.043613334496816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,3072,0.037343111303117536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,256,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,16384,0.033922665648990206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,4096,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,2048,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,64,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,51200,0.7559537887573242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,512,0.011324444578753578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,5120,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,1536,0.004844444493452708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,32,0.008360889222886827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,32,0.0034453332838084963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,16384,0.2866506576538086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,4096,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,51200,3.268962648179796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,3584,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,6144,0.031055112679799397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,128,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,12288,0.16878222094641793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,2048,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,8192,0.16164622041914198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,4096,0.008067555725574493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,2048,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,5120,0.09582133425606622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,2560,0.012033777932325998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,1024,0.0035626664757728577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,3584,0.037773334317737155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,12288,0.02481777800454034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,3072,0.10642488797505696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,4096,0.01606666710641649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,6144,0.017344888713624742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,2560,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,3072,0.16178399986690944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,8192,0.01827022267712487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,8192,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,4096,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,5120,0.1012951135635376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,7168,0.017477333545684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,2048,0.02701688806215922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,2048,0.04327466752794054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,65536,0.058167113198174365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,51200,0.6583111021253798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,65536,0.10831111007266575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,512,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,768,0.020847999387317236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,7168,0.031630221340391375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,65536,0.9906977547539605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,128,0.09139021899965073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,512,0.0042204442951414315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,512,0.059472892019483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,10240,0.15982311301761204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,3072,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,10240,0.012532444463835822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,65536,0.05327644281917148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,512,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,2048,0.02159022291501363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,12288,0.22600534227159289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,8192,0.023747555083698694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,2048,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,1536,0.01763911048571269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,1024,0.0032275555034478507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,51200,0.20750310685899523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,128,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,1024,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,3584,0.013203555511103736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,768,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,768,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,256,0.0239982224173016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,256,0.015094222293959724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,16384,0.021792888641357422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,65536,0.1052924460834927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,51200,0.1505875587463379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,2048,0.018567999203999836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,256,0.024979554944568213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,6144,0.2559724383884006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,51200,0.09946755568186443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,6144,0.07029066483179729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,768,0.01740533279048072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,8192,0.06759466727574666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,1536,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,51200,0.12053510877821182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,5120,0.044792887237336904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,256,0.020974222156736586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,512,0.0057795556883017225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,128,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,128,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,32,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,8192,0.01650399963061015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,3584,0.012329777909649743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,16384,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,2560,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,2048,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,2560,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,3584,0.01923466722170512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,2048,0.0642204417122735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,2048,0.013384888569513956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,3584,0.051402668158213295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,7168,0.0783768892288208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,256,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,768,0.03314044409328037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,768,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,6144,0.04696977800793118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,10240,0.016187555260128446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,1536,0.011192888849311404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,256,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,51200,0.17816000514560273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,4096,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,64,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,512,0.046238223711649575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,8192,0.027377777629428442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,128,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,768,0.018415111634466384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,1536,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,3072,0.6119155353969997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,4096,0.038346668084462486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,768,0.22379554642571342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,1536,0.020445333586798776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,256,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,8192,0.031162665949927434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,2048,0.00941244430012173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,512,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,6144,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,6144,0.03274844421280755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,10240,0.012069332930776807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,10240,0.0143857780430052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,3072,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,4096,0.02401599950260586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,8192,0.7564755545722114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,3072,0.7798133426242404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,64,0.008436444732877944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,64,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,6144,0.032594667540656194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,4096,0.014167111780908374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,5120,0.012801777985360889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,16384,0.01330933305952284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,7168,0.15789066420661077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,65536,0.06333688894907634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,1536,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,3072,0.2756444348229302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,10240,0.03079644507831997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,64,0.01716444392999013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,128,0.009422221945391761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,768,0.12442666954464382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,51200,0.05398133397102356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,2048,0.02884888980123732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,16384,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,2048,4.990789201524523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,7168,0.04053599966896905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,65536,0.07462310791015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,5120,0.01776000029510922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,1024,0.025416000021828547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,128,0.0018862223045693503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,768,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,768,0.012774222426944308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,5120,0.010761777559916178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,512,0.02128622266981337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,512,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,8192,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,4096,0.25294044282701283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,1024,0.019840889506869845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,64,0.0018959999498393799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,3072,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,51200,0.05761155817243788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,3072,0.03646488984425863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,6144,0.06574044624964397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,12288,0.05167733298407661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,6144,0.026495110657480028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,1536,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,512,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,6144,0.018546667363908555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,5120,0.04957511027654012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,3072,0.0991742213567098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,3072,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,2048,0.0366231103738149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,1536,0.055162668228149414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,16384,0.02117777698569828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,1024,0.029115554359224107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,7168,0.09924622376759847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,256,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,768,0.0032435556252797446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,5120,0.01755199995305803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,7168,0.012921777864297232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,512,0.012060444388124677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,4096,0.03546133306291368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,8192,0.02949955397182041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,768,0.012509333590666452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,768,0.006476444502671559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,3584,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,51200,0.2774782180786133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,128,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,32,0.015674667225943673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,128,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,12288,0.03529511226548089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,8192,0.045533332559797496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,51200,0.05283822284804451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,256,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,768,0.014395554860432943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,8192,0.0652266674571567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,8192,0.12842578358120388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,65536,2.0087333255343967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,65536,13.748744540744356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,4096,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,256,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,2560,0.02100533412562476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,1024,2.574158138699002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,8192,0.0598942240079244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,8192,0.06895466645558675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,51200,0.27615287568834096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,2048,0.33361511760287815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,512,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,512,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,4096,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,256,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,512,0.025423111187087163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,10240,0.03616444600952996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,51200,0.5169315338134766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,2048,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,3584,0.014549333188268872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,3584,0.08611466487248738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,2048,0.013232000172138214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,1536,0.016559999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,768,0.009182222187519073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,3584,0.009783999787436591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,12288,0.024337778488794964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,2560,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,16384,0.020669332808918424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,256,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,51200,0.06600977977116902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,5120,0.03521155648761325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,8192,0.09531199932098389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,10240,0.07691822449366252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,65536,0.06262399752934773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,512,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,10240,0.023518222901556227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,8192,0.01556711064444648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,2560,0.020095111595259774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,3072,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,1536,0.04831644561555651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,12288,0.016328889462682936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,2048,0.032060444355010986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,51200,0.7534657584296333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,4096,0.04721777637799581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,32,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,5120,0.031167109807332356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,10240,0.01440177857875824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,8192,0.010964444114102257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,5120,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,65536,0.07636266946792603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,5120,0.030664887693193223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,32,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,12288,0.2412275473276774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,10240,2.399343066745334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,16384,0.02385777731736501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,512,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,8192,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,1024,0.011251555548773872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,1024,0.00720000018676122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,10240,0.10749866565068562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,5120,0.009894222021102905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,2560,0.004839999808205498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,768,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,5120,0.03937777876853943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,51200,0.20718222194247773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,65536,0.08515733480453491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,4096,0.011929777761300405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,51200,0.19435555405086943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,256,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,4096,0.04298311140802172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,6144,0.31447911262512207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,6144,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,768,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,512,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,3584,0.047862221797307335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,1536,0.008064889245563084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,256,0.016904888881577384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,8192,0.03370133373472426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,51200,0.5748026635911729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,768,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,65536,0.07639644543329875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,3072,0.026399110754330952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,65536,0.27813956472608775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,2560,0.22626309924655488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,2560,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,2048,0.006554666492674086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,64,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,3584,0.11746488677130805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,10240,0.042543109920289784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,6144,0.17381155490875244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,768,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,4096,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,7168,0.013402666482660504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,10240,0.1675351063410441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,128,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,10240,0.31054843796624076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,3072,0.03979466689957513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,256,0.002458666761716207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,1536,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,2560,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,1024,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,768,0.028359111812379625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,3072,0.07291466659969754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,10240,0.01259022206068039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,65536,0.06426666842566596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,256,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,4096,0.23213333553738066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,1024,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,3584,0.0092284447617001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,1024,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,768,0.04232977827390035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,3584,0.05100889007250468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,4096,0.02944355540805393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,32,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,65536,1.062744034661187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,8192,0.5953262117173936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,512,0.025358221597141687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,256,0.007236444287829929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,2048,0.012676444318559436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,10240,0.028288000159793433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,10240,0.2820853392283122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,16384,0.02198222279548645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,2048,0.021623111433453027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,16384,0.038535111480289035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,51200,2.7796372307671438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,128,0.06587911314434476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,3072,0.02962133288383484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,10240,0.05243644449445936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,6144,0.03939111034075419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,4096,0.01145866678820716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,256,0.009646221995353699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,3584,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,3584,0.017172444197866652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,256,0.09546133544709946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,2048,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,5120,0.012857777376969656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,4096,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,128,0.0053288886944452924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,32,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,4096,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,12288,0.016420443852742512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,16384,0.02609866691960229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,512,0.024296889702479046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,5120,0.08771733442942302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,4096,0.011386666860845355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,1536,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,3584,0.10025511185328166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,3072,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,1536,0.009431111315886179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,65536,0.2798231177859836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,128,0.016903999778959487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,768,0.015232000086042615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,2560,0.023708444502618577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,32,0.005787555542257097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,5120,0.009635555247465769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,65536,0.3522062301635742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,1536,0.01681066718366411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,6144,0.03289955523278978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,6144,0.024705777565638225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,10240,0.24321420987447104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,1536,0.006124444305896759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,1024,0.04389244318008423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,3072,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,3072,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,5120,0.013730666703648038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,8192,0.0107022225856781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,10240,0.01458577811717987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,7168,0.04678666591644287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,3584,0.0170462214284473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,3072,0.08905066384209527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,1536,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,10240,0.046747555335362755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,3072,0.01763911048571269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,32,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,64,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,51200,0.02853155467245314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,6144,0.08578666713502671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,3072,0.15185333622826472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,10240,0.04871288935343424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,2048,0.07045866383446588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,16384,0.07712178097830878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,8192,0.24572266472710502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,6144,0.01184266640080346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,6144,0.1821662187576294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,4096,0.025356445047590468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,768,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,65536,0.13719199763403997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,3072,0.3895280096266005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,7168,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,1536,0.02120977805720435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,3584,0.15084621641370985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,2560,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,7168,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,1536,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,6144,0.027076444692081873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,65536,0.0564604467815823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,16384,0.023727110690540735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,64,0.0019360000474585427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,512,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,128,0.009633777870072259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,3072,0.034725334909227155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,3584,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,12288,0.05318399932649401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,3072,0.07113422287835015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,256,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,7168,0.04621600111325582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,5120,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,12288,0.05079466766781277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,3072,0.029691557089487713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,1536,0.020979555116759405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,16384,0.023642667465739783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,6144,0.21776621871524385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,7168,0.0199626667631997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,7168,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,2048,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,1536,0.01183999992079205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,2560,0.8812133471171061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,3584,0.03559199968973795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,2048,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,51200,0.9530080159505209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,16384,0.08201599783367582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,12288,0.09951644473605686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,10240,0.01715555621518029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,512,0.011531555818186866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,16384,0.0810862249798245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,3584,0.10782933235168457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,7168,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,51200,0.043459554513295494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,51200,0.05647199683719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,512,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,2048,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,51200,2.9080933464898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,1536,0.029490666257010564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,1024,0.014168888330459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,2560,0.025581333372328017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,2560,0.06522755490409003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,512,0.016583111551072862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,10240,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,7168,0.016229333149062265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,5120,0.16742400328318277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,32,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,2560,0.1844222280714247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,3584,0.0693884425693088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,768,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,12288,2.7988683912489147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,512,0.017530666457282174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,1024,0.0032355555643637977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,7168,0.021235555410385132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,16384,0.07636977566613092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,1536,0.02377066678471035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,1536,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,128,0.17109155654907227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,2048,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,65536,0.08646755748324925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,32,0.018444443742434185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,5120,0.021993777818149988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,12288,0.024136889312002394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,3584,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,256,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,64,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,256,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,512,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,1536,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,3072,0.04103999998834398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,5120,0.009517333573765224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,2560,0.02480799953142802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,1536,0.003882666842805015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,65536,0.14726932843526205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,3072,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,3072,0.01385600037044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,1024,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,16384,0.3518017662896051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,512,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,5120,0.2637217839558919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,7168,0.07456355624728732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,2048,0.020152888364262052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,51200,0.21065955691867402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,768,0.019627556204795837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,128,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,4096,0.09079733159806992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,3072,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,12288,0.10382044315338135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,16384,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,512,0.01684533390733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,3584,0.039340443081325956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,16384,0.15628355079227024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,16384,0.029651555750105116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,16384,0.0674168864885966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,8192,0.023623999622133043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,10240,0.016023111012246873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,1536,0.04468355576197306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,512,0.054004444016350635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,1536,0.06655289067162408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,1536,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,5120,0.09488711092207168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,512,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,12288,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,1024,0.016221332881185744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,6144,0.01350044459104538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,51200,0.30894221199883354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,65536,0.06922399997711182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,2048,0.017952000101407368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,768,0.019280888968043856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,4096,0.009752000371615091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,256,0.0022435554613669715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,2048,0.024001777172088623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,7168,0.012815111213260226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,6144,0.03937600056330363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,3072,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,256,0.0063635553750726916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,32,0.039729777309629656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,1536,0.018896889355447557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,6144,0.06312444474962023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,10240,0.6340079837375218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,1536,0.16668000486161974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,32,0.002206222257680363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,3584,0.04820444517665439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,7168,0.009734222458468543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,768,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,5120,0.08577422300974528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,5120,0.20431199338701037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,5120,0.09840177827411228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,7168,0.9849199718899198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,2048,0.012429333395428128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,4096,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,3072,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,8192,0.02980888883272807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,32,0.0018959999498393799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,7168,0.011427555647161273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,2048,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,10240,0.18117245038350424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,3072,0.005482666608360078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,5120,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,4096,0.18841155370076498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,512,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,10240,0.01310755560795466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,1024,0.009296000003814697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,4096,0.024455999334653217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,7168,0.010208889014191097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,2560,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,512,0.008920888933870528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,2048,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,256,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,10240,0.10183022419611613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,64,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,7168,0.5299244456821018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,3584,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,768,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,768,0.01907111042075687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,10240,0.027003554834259882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,12288,0.03411733441882663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,2560,0.008058666355080074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,5120,0.004535111288229625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,32,0.006509333435032103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,6144,0.10444533162646824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,256,0.01923377811908722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,16384,0.7712906731499566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,1024,0.01015022231472863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,512,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,1024,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,2048,0.03334666623009576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,256,0.007010666860474481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,2560,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,1024,0.006664888726340399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,1536,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,51200,0.060119999779595264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,12288,0.004927111168702443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,5120,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,2560,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,65536,0.07744710975223117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,8192,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,12288,0.17338132858276367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,8192,0.02648266653219859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,2048,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,128,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,3584,0.03183822168244256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,3584,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,16384,0.04798755711979336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,10240,0.016157334049542744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,2560,0.046965334150526256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,128,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,768,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,1536,0.02425066630045573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,64,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,8192,0.012151111331250934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,128,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,2048,0.009876444107956356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,3584,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,7168,0.029639111624823675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,768,0.013815110756291283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,4096,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,3584,0.007777777810891469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,5120,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,5120,0.025006221400366888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,2560,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,4096,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,1536,0.008367111285527548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,128,0.015284443895022074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,1536,0.009730666875839233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,768,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,7168,0.034063110748926796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,1024,0.011360000405046674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,2048,0.07785955402586195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,8192,0.018556444181336295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,64,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,32,0.00903555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,6144,0.1664044459660848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,768,0.010483555495738983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,4096,0.06206755505667793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,51200,0.12745689021216497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,2560,0.032920890384250216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,3072,0.015581332974963717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,7168,0.028385778268178303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,2560,0.02534577747186025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,1536,0.010079999764760336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,512,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,6144,0.9557742012871636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,512,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,256,0.014775110615624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,768,0.05145777927504646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,128,0.09932533237669204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,2560,0.01347555551264021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,16384,0.06782844331529406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,5120,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,8192,0.02295644415749444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,51200,2.3206123775906033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,16384,0.25247732798258465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,32,0.09811911318037246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,4096,0.1676835616429647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,1536,0.024029332730505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,8192,0.09044177664650811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,12288,0.03020711077584161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,6144,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,10240,0.09262577692667644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,64,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,1024,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,128,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,16384,0.3332168791029188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,4096,0.0927084419462416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,16384,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,128,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,1536,0.024439111351966858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,128,0.014211555322011312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,3584,0.1487333377202352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,6144,0.018114666144053142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,12288,0.32377155621846515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,1024,0.1504711045159234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,10240,0.024879111184014216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,256,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,768,0.09709955586327447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,12288,0.025640888346566096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,6144,0.23739287588331434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,3072,0.0079297779334916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,128,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,51200,0.10166933139165242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,3072,0.012176888684431711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,512,0.01723377737734053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,3072,0.008911111288600499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,1536,0.025601777765485976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,2048,0.03186044428083632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,3072,0.04335911075274149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,1536,0.01791200041770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,1536,0.4012666543324788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,6144,0.015154666370815702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,6144,0.08163377973768446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,64,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,3584,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,16384,0.04264444443914625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,128,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,32,0.026749332745869953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,5120,0.45698753992716473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,1024,0.019543111324310303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,6144,0.00977600034740236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,32,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,768,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,2048,0.23653690020243326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,32,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,51200,0.3608871036105686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,16384,0.024488889508777197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,4096,0.010277333358923594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,6144,0.03360622127850851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,2048,0.011238222320874533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,2560,0.004524444540341695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,12288,0.036658667855792575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,8192,0.05344000127580431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,10240,0.08721511231528388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,12288,0.040906667709350586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,2048,0.008955555657545725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,4096,0.034747554196251765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,3072,0.04575555523236593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,12288,0.024882665938801233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,7168,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,4096,0.02863022353914049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,64,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,6144,0.00889955543809467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,12288,0.022640888889630634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,16384,0.19493510988023546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,5120,0.026353778110610113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,2560,0.04641244477695889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,2560,0.048688001102871366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,16384,0.2817395528157552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,1024,0.003927111211750242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,10240,0.01993777851263682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,128,0.021394666698243883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,5120,0.11093955569797093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,256,0.016694222887357075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,512,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,4096,0.026456000076399908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,51200,0.6101386811998155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,8192,0.7300800217522515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,2048,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,1024,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,128,0.020038222273190815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,3072,0.0236826671494378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,12288,0.023377777801619634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,8192,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,10240,0.026001777913835313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,8192,0.4796453581915961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,32,0.009474666582213508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,256,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,65536,0.21943467193179658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,51200,0.09677600198321873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,10240,0.21477777428097197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,4096,0.02029600077205234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,3072,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,51200,0.06296444601482816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,1536,0.0098213329911232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,7168,0.023731556203630235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,4096,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,768,0.019959110352728102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,1536,0.003933333274390963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,51200,0.19931111070844862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,256,0.025772444076008264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,32,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,64,0.04611911045180427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,65536,0.995382202996148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,1024,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,512,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,768,0.02829155657026503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,32,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,256,0.015461333923869662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,12288,0.10492799679438274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,8192,0.01568266749382019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,128,0.014583110809326172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,7168,0.033901333808898926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,12288,0.018567111757066514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,1024,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,3584,0.007117333511511485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,6144,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,1024,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,5120,0.04543377624617683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,1024,0.009450666606426239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,12288,0.0576711098353068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,3072,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,1536,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,512,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,3072,0.08344444301393296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,2560,0.021264889174037512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,256,0.2321324348449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,768,0.044158223602506846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,6144,0.009706666900051964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,768,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,7168,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,512,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,7168,0.0052888890107472735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,256,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,3072,0.08729510837131077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,32,0.0017760000709030365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,6144,0.10971199803882176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,51200,0.45266220304701066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,10240,0.019467555814319186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,3584,0.013040888640615674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,2560,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,7168,0.1345777776506212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,16384,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,1024,0.0053102220926019884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,12288,0.014582221706708273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,4096,0.03471022182040744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,6144,0.03927910990185208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,8192,0.03384977910253737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,256,0.011235555840863122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,32,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,768,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,32,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,64,0.08188800017038982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,768,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,128,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,51200,0.2395768960316976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,65536,0.0421982208887736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,3072,0.2962968879275852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,2048,0.0748151143391927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,4096,0.004516444272465176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,16384,0.034909334447648786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,2560,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,2560,0.0477128889825609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,7168,0.1250106626086765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,64,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,3072,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,1024,0.010803555448849997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,128,0.002100444502300686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,12288,0.22206221686469185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,7168,0.013311111264758639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,8192,0.16516355673472086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,2048,0.028460443019866943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,65536,0.07333955499860975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,12288,3.7115830315483938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,7168,0.03525333272086249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,768,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,32,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,65536,0.17692800362904867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,6144,0.04504799842834473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,3584,0.008882666627566019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,8192,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,16384,0.11539467175801595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,7168,0.7678284645080566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,2560,0.03345511025852627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,5120,0.01991555591424306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,3584,0.023339556323157415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,51200,0.15152000056372747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,3072,0.01827911039193471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,3072,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,5120,0.031960888041390315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,65536,0.2605431079864502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,1536,0.06662488645977445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,65536,0.06299466556972928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,256,0.0021973333011070886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,16384,0.01965688831276364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,1024,0.17216889063517252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,256,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,3072,0.1278417772716946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,1024,0.018896889355447557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,1536,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,7168,0.03069066670205858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,2048,0.032771554258134626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,51200,0.16312177975972494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,4096,0.04636089007059733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,32,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,1024,0.028452446063359577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,65536,0.07818933327992757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,3072,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,256,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,256,0.04046577877468533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,6144,0.010264889233642155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,5120,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,7168,0.035312000248167254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,2048,0.00702044450574451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,5120,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,2048,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,128,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,16384,0.07480088869730632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,512,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,6144,0.13573600186242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,8192,0.009415999882751042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,12288,0.3276044527689616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,10240,0.026408889227443274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,5120,0.10126666890250312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,32,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,8192,0.014268444644080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,768,0.03163911236657037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,4096,0.01219466659757826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,2560,0.016695110334290397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,65536,0.03468799922201369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,1024,0.025231111380789015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,3584,0.02641333308484819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,2048,0.032405333386527166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,256,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,7168,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,128,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,32,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,128,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,51200,0.046225776274998985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,1536,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,128,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,16384,0.01962577799956004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,4096,0.07493599918153551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,256,0.020430222153663635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,10240,0.7665617730882434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,32,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,3584,0.011348444554540845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,128,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,6144,0.02700533303949568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,3584,0.031091554297341242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,3072,0.020786666207843356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,1536,0.05369155605634054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,12288,0.018954666124449838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,65536,0.05783466498057047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,6144,0.9253155390421549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,4096,0.012055111428101858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,10240,0.5463946660359701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,5120,0.12875733110639784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,7168,0.14707822269863552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,5120,0.011947555674446953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,128,0.053759998745388456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,4096,0.008262222011884054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,2560,0.007052444749408298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,2048,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,2560,0.026716444227430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,64,0.06867111391491361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,10240,0.2327946556939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,12288,0.030135111676322088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,10240,0.10016444656583999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,16384,0.05116622315512764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,8192,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,12288,0.10287644465764363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,7168,0.019259555472267997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,3072,0.12986932860480413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,4096,0.029698666599061754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,768,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,512,0.00258577780591117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,512,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,6144,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,5120,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,5120,0.032551109790802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,768,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,2560,0.021086222595638696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,51200,0.05029244555367363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,1536,0.03494311041302151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,3584,0.05886489152908325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,3584,0.03524177935388353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,32,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,8192,0.034141331911087036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,2560,0.14852800634172228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,2048,0.05832622448603312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,128,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,3584,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,1024,0.011947555674446953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,8192,0.019357333580652874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,512,0.05003822180959913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,256,0.10695466730329727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,1536,0.0073315559162033936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,768,0.01871200071440803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,3584,0.038451555702421404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,3584,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,3072,0.0075866662793689305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,1024,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,768,0.020429333051045735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,51200,0.10748177766799927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,1536,0.03957155678007338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,128,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,64,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,16384,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,8192,0.0803991092575921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,1536,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,256,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,5120,0.023190221852726404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,2560,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,51200,0.6498079829745823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,12288,0.032695998748143516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,128,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,512,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,128,0.024777778320842322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,65536,1.5574817657470703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,10240,0.2921804322136773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,2048,0.010325333310498131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,6144,0.021751110752423603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,2560,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,3072,0.009714666340086196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,3072,0.3412071069081624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,1024,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,3584,0.01610488858487871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,65536,0.18400088946024576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,3584,0.13279644648234049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,6144,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,10240,0.0491920014222463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,256,0.006151999864313338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,8192,3.3098915947808156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,8192,0.034761776526769005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,1536,0.23343822691175672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,256,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,2048,0.12368710835774739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,7168,0.03023911184734768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,51200,0.07244088914659289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,256,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,3584,0.03673511081271701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,256,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,51200,0.11576978365580241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,6144,0.01535555554760827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,12288,0.013403555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,768,0.03921688927544488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,1024,0.02312533391846551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,6144,0.030966222286224365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,768,0.01629244453377194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,1024,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,65536,2.4114676581488714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,2048,0.022223111655977037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,7168,0.5836462444729275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,51200,0.42793067296346027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,512,0.19161600536770293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,512,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,3584,0.028377778000301782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,512,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,2560,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,5120,0.060243553585476346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,64,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,768,0.006365333166387346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,128,0.0053013331360287136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,1024,0.018933332628673978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,3584,0.020419556233617995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,16384,0.025705777936511572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,256,0.033463110526402794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,16384,0.05199822121196323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,3584,0.20217511388990614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,1024,0.005751111027267244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,1024,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,12288,3.6790792677137585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,8192,0.020855110552575853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,2560,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,8192,0.2210622231165568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,512,0.02330933345688714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,8192,0.0741528868675232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,128,0.19274666574266222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,1536,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,10240,0.012119111087587146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,2560,0.38223200374179417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,10240,0.028100444210900202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,7168,0.02237777743074629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,2560,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,512,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,12288,0.0500239994790819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,4096,0.04428799947102865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,64,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,10240,0.013819555441538492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,12288,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,7168,0.004987555659479565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,6144,0.012774222426944308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,768,0.00906044410334693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,5120,0.03211288981967502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,3584,0.039136889908048846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,5120,0.016460445192125108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,2048,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,12288,5.1295522054036455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,2048,0.020562666985723708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,8192,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,512,0.02754577828778161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,32,0.001777777862217691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,7168,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,51200,0.7164693408542209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,512,0.008041777544551426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,768,0.006167111297448476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,1536,0.008024888734022776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,6144,0.10159733560350205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,6144,0.04615822103288439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,32,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,128,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,128,0.01237511138121287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,3072,0.019165333774354722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,256,0.0053280000057485365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,2560,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,2048,0.028516444895002577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,5120,0.3468666606479221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,1024,0.04422488808631897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,768,0.007063999772071838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,4096,0.04996977912055122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,8192,0.09088444709777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,2048,0.048192888498306274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,2560,0.01923377811908722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,51200,0.10872089200549656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,5120,0.09601333406236436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,5120,0.04262222184075249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,4096,0.02667466633849674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,7168,0.7800275484720866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,10240,0.09582310914993286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,2560,0.024655110306209985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,256,0.013392888837390475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,1024,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,2048,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,10240,0.04065333472357856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,2560,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,3072,0.11556177669101292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,51200,0.060216890441046826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,16384,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,51200,0.048840890328089394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,768,0.009463110731707679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,12288,0.01443822185198466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,12288,0.029701332251230877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,1536,0.004223111189073986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,768,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,512,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,128,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,16384,0.05702133311165703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,3584,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,10240,0.06776177883148193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,32,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,10240,0.025042666329277888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,3072,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,1024,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,128,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,51200,0.25599466429816353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,4096,0.4236186610327826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,16384,0.019156444403860304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,16384,0.15367999341752794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,8192,0.20991022057003447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,32,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,3584,0.49039824803670246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,3584,0.19827466540866426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,256,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,3584,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,12288,0.7370462417602539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,512,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,512,0.03752622339460585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,512,0.013956444131003486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,3072,0.01626399987273746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,8192,0.0852924452887641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,51200,0.09706222348743015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,12288,0.18235111236572266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,10240,0.012636444634861417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,6144,0.14565511544545492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,2560,0.5481111208597819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,12288,0.02367199957370758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,3584,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,7168,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,768,0.01920000049802992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,3584,0.06399822235107422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,8192,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,8192,0.01645955608950721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,3584,0.12585155169169107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,4096,0.016418667303191293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,64,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,256,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,12288,1.0797466701931424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,256,0.025042666329277888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,2560,0.052910221947564014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,2048,0.04591111010975308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,768,0.018793778287039865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,1024,0.014191110928853353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,3072,0.34081599447462296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,51200,0.08993066681755914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,12288,0.2364115450117323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,5120,0.019941333267423842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,1024,0.004221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,1536,0.011738667057620155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,2560,0.04169422056939867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,1536,0.02606044378545549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,4096,0.020587555236286588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,128,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,768,0.024111110303137038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,1024,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,512,0.021980444590250652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,2560,0.024843555357721116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,1536,0.015973332855436537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,1536,0.035395556026034884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,2048,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,128,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,51200,0.1705848905775282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,51200,3.761712816026476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,12288,0.024745777249336243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,3584,0.020454222957293194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,32,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,16384,0.08673777845170762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,12288,0.023456000619464453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,51200,0.136698669857449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,12288,0.48499642478095156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,7168,0.05307911170853508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,10240,0.011808888779746162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,64,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,8192,0.031128889984554712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,8192,0.02842666705449422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,8192,0.07661510838402642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,8192,0.07179555628034803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,32,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,12288,0.06202666627036201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,768,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,3072,0.010135999984211391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,8192,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,768,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,6144,0.3144177860683865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,3584,0.010777778095669217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,7168,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,8192,0.016692444682121277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,10240,0.010443555812040964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,512,0.01086400035354826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,51200,3.992198944091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,64,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,2048,0.02146133283774058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,2048,0.07373066743214925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,768,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,16384,0.24262576633029512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,2048,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,3584,0.024451555477248296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,7168,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,2048,0.050019555621676974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,10240,0.034421334664026894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,4096,0.026778666509522334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,7168,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,12288,0.03988800115055508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,2048,0.021737777524524268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,51200,0.04758755697144402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,1024,0.009309333231714036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,128,0.006050666587220297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,5120,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,51200,0.06555200285381742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,2048,0.07374044259389241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,8192,0.26069598727756077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,12288,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,10240,2.6267884572347007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,10240,0.1933057838016086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,128,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,3584,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,1024,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,512,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,512,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,64,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,768,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,1024,0.3200826644897461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,128,0.010334221853150262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,65536,0.6369893285963271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,65536,0.9767110612657336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,65536,0.06248178084691366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,6144,0.01796355512407091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,16384,0.021475555168257818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,5120,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,1536,0.5530408753289117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,1536,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,12288,0.01882133384545644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,1024,0.3022106753455268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,5120,0.11332889397939046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,3584,0.09759555260340373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,10240,0.009948444035318162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,6144,0.017417778571446735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,8192,0.02495199938615163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,6144,0.01711466742886437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,51200,0.05779466364118788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,2560,0.03817777832349142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,1024,0.018446221947669983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,8192,0.01369066701995002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,8192,0.01594577729701996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,51200,0.10298933585484822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,12288,0.03797244363360935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,65536,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,64,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,512,0.030123554997973975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,5120,0.06432177623112996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,1536,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,65536,0.0783546699417962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,51200,0.9743475384182401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,65536,0.09984444247351752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,12288,0.08008978101942274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,51200,0.22957955466376412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,2560,0.06969066460927327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,6144,0.03590933481852213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,8192,0.052200890249676175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,7168,0.0123511114054256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,1536,0.07711111174689399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,8192,0.04251911242802938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,12288,0.05720000134574043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,7168,0.02277244461907281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,10240,0.1834657722049289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,6144,0.009940444595283931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,6144,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,16384,0.027444443768925134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,65536,0.24637333552042642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,128,0.015466666883892484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,64,0.005196444276306364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,2048,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,10240,0.0491831103960673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,7168,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,2560,0.0834862192471822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,12288,0.017091555727852713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,8192,0.10018489095899795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,768,0.018184888694021437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,7168,0.021672889590263367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,768,0.036385777923795916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,8192,0.025106666816605464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,3584,0.0096142225795322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,1536,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,16384,0.05839911434385511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,128,0.015631111131774057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,768,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,256,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,51200,0.07361955775154962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,8192,0.07686044772466023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,512,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,3584,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,2560,0.031791110833485924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,3072,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,1024,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,2048,0.006686222222116258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,16384,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,128,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,1024,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,768,0.0244195560614268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,5120,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,65536,0.07506133450402154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,51200,0.3176639874776204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,8192,3.5443678961859812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,8192,0.14070133368174234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,12288,0.04256177941958109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,65536,0.3894195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,6144,0.021704889006084863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,4096,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,512,0.1039431095123291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,12288,0.03074666526582506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,12288,0.03334310981962416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,1536,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,2560,0.011559999651379056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,51200,9.690988328721788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,3072,0.024680889315075342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,3584,0.0064293332397937775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,10240,0.01627111103799608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,32,0.06210133102205065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,512,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,128,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,3584,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,5120,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,2048,0.033612443341149226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,512,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,32,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,128,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,32,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,256,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,6144,0.004598222259018156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,12288,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,12288,0.02241866621706221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,16384,0.02780533168050978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,128,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,51200,0.09521244631873237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,2560,0.008930666579140557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,768,0.03933777742915683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,8192,0.05805689096450806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,65536,0.1833653317557441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,12288,0.14621155791812473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,10240,0.10346221923828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,10240,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,128,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,8192,0.05765333440568712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,1024,0.005326222214433882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,768,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,64,0.006122666514582104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,3584,0.008795555267069075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,768,0.03424355718824599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,1024,0.021168889270888433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,2048,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,6144,0.3285200066036648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,5120,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,2048,0.02128622266981337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,10240,0.04628355635537041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,2048,0.019259555472267997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,2048,0.04246044490072462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,6144,0.8326240115695529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,8192,0.027248000105222065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,8192,0.13179022736019558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,2560,0.022077333596017625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,16384,0.21373866664038765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,3072,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,32,0.013119999733236102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,16384,0.26324444346957737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,5120,0.0222080002228419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,51200,0.01329955541425281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,1024,0.011104889214038849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,3584,0.021682666407691106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,51200,0.0492888887723287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,3584,0.0820337798860338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,6144,0.24296887715657553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,6144,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,1024,0.020488889680968392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,12288,0.735819551679823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,512,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,1024,0.03665688965055678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,512,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,2560,0.01573777695496877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,512,0.00811288919713762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,51200,2.093900468614366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,1024,0.036554667684766985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,32,0.0655919975704617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,65536,0.03638133406639099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,8192,0.05253244439760844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,7168,0.026175111532211304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,3072,0.03198577629195319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,128,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,65536,0.14970489343007407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,4096,0.05836977561314901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,8192,0.010016889207892949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,2048,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,7168,0.015318221516079374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,128,0.028798222541809082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,51200,0.22713422775268555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,128,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,1536,0.004359111189842224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,256,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,8192,0.0602933300866021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,51200,0.19784888956281874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,10240,0.3018737898932563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,10240,0.01518311103185018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,5120,0.009112888740168678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,768,0.01128799964984258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,16384,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,1536,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,768,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,3072,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,2048,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,1536,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,7168,0.023476445012622412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,2560,0.07591288619571261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,32,0.004855111241340637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,65536,0.06961955626805623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,32,0.02379377848572201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,10240,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,65536,0.04727466570006477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,10240,0.1447653373082479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,64,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,1536,0.014440889159838358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,51200,0.060177776548597545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,51200,0.5260142220391167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,64,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,7168,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,2048,0.006816000160243776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,2560,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,2048,0.029119110769695703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,6144,0.14220710595448813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,1536,0.022994667291641235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,1536,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,5120,0.10643466975953843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,128,0.007917332980367873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,12288,0.2141280041800605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,12288,0.07668888568878174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,12288,0.03138400117556254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,6144,0.10762044456270005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,8192,0.044109334548314415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,65536,0.08541244268417358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,2048,0.27130667368570965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,51200,124.3880886501736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,2048,0.04087555408477783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,64,0.002410666603181097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,512,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,12288,0.01608622239695655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,3072,0.253295103708903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,2560,0.03742488887574937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,3072,0.0701191094186571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,256,0.005740444279379315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,7168,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,2048,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,3072,0.06456355253855388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,51200,0.10360711150699192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,4096,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,51200,4.757294125027127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,2560,0.04472000069088406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,6144,0.08317422204547457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,768,0.016435555285877652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,512,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,6144,0.034966223769717746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,32,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,768,0.01754844519827101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,128,0.04209333327081469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,2048,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,32,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,12288,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,1024,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,128,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,12288,0.027627555860413447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,2560,0.05065244436264038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,5120,0.02607288956642151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,512,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,51200,0.05189955565664503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,2048,0.014667555689811707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,768,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,256,0.03864355550871955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,32,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,256,0.02533333334657881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,2560,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,12288,0.013615999784734515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,65536,0.25227999687194824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,12288,1.559067514207628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,7168,0.011075555450386472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,3584,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,7168,0.0288702216413286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,1024,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,4096,0.04764088988304138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,1536,0.019151111443837483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,65536,0.21982667181226942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,3072,0.031209776798884075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,6144,0.20310666826036242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,6144,0.031102223528756037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,8192,0.10023644235399033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,10240,0.03706666827201843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,2560,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,1536,0.014873777826627096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,128,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,6144,0.023048889305856492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,1536,0.021000888612535264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,10240,0.04824622140990364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,256,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,2560,0.0332915551132626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,4096,0.04484355449676514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,256,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,512,0.0159582214223014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,1536,0.4267733361985948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,51200,3.6430943806966147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,8192,0.013720889058378009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,10240,0.031939556201299034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,10240,0.1591715547773573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,1536,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,12288,0.034450666771994695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,4096,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,2048,0.011516444385051727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,8192,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,5120,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,4096,0.0424542228380839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,6144,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,10240,0.03660533494419522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,3072,0.04702755477693346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,2048,0.03573866685231527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,256,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,2560,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,51200,0.212135116259257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,51200,0.06274577644136217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,2560,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,2048,0.008296888735559251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,512,0.015708444847000968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,64,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,1024,0.006701333241330252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,64,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,5120,0.013119999733236102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,1024,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,5120,0.01649777756796943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,768,0.14195377296871609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,5120,0.010899555351999072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,256,0.028275556034511987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,5120,0.041718221373028226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,10240,0.009045333498054082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,51200,0.03963377740648057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,128,0.00685511115524504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,65536,0.058713780509101025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,32,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,1536,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,768,0.0038488888078265717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,3584,0.026127111580636766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,16384,0.021575111481878493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,3584,0.06895022259818183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,65536,0.5100284682379829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,51200,0.5280417866177028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,2560,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,7168,0.10460266802046035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,8192,0.056792888376447886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,16384,0.03622755408287048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,12288,0.04898488852712843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,2560,0.06348888741599189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,51200,0.1256035566329956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,768,0.022683555881182354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,128,0.008466666771305932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,512,0.05841244591606987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,12288,0.013819555441538492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,2560,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,1024,0.0037377778854635027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,3584,0.03070844544304742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,51200,0.10079644123713176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,10240,0.027484445108307734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,5120,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,65536,0.0793715582953559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,3072,0.23804354667663574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,512,0.007090667055712805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,32,0.005790222022268508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,2048,0.0125608891248703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,2560,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,768,0.008776889079146916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,16384,0.08302933639950223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,2048,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,32,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,1024,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,12288,0.14930310514238146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,1536,0.005032888717121548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,64,0.00535733335547977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,12288,0.23903022872077095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,6144,0.04014399978849623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,128,0.26459821065266925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,16384,0.07441688908471002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,6144,0.01202666676706738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,4096,0.012819555898507437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,51200,0.10996355613072713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,7168,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,8192,0.1679768827226427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,12288,0.3460640112559001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,6144,0.06750755839877658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,3072,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,2048,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,1536,0.005192000004980299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,64,0.009939555492666032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,512,0.021881777379247878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,7168,0.04124266571468777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,16384,0.11869866318172878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,256,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,768,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,128,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,128,0.028635554843478735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,3072,0.02215733296341366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,5120,0.008776889079146916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,5120,0.049673779143227466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,1536,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,1536,0.05910755528344048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,6144,0.009029332962301042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,768,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,1536,0.015996444556448195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,768,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,3584,0.026724444495307073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,128,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,6144,1.1720079845852323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,2048,0.05082222157054477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,3072,0.013193777865833707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,1536,0.17447022596995035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,51200,0.14234844843546549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,512,0.008902221918106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,5120,0.029066665305031672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,64,0.006994666324721442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,65536,0.04789066645834181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,7168,4.315188513861762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,1024,0.012350222302807702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,1024,0.0090506664580769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,5120,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,128,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,2560,0.06458666589525011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,3072,0.13584000534481472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,2048,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,12288,0.02131822208563487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,16384,0.05344444513320923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,128,0.32696710692511666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,8192,0.018629334039158292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,51200,0.012074666718641916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,65536,0.1456044382519192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,2560,0.009384888741705153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,65536,19.411855061848957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,3584,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,128,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,64,0.008541333178679148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,16384,0.024479111035664875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,65536,0.15368444389767116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,5120,0.01867733399073283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,7168,0.012210666305489011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,1536,0.011516444385051727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,2048,0.034308443466822304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,256,0.01573511130279965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,128,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,768,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,768,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,256,0.005487111293607288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,64,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,512,0.02365422248840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,7168,0.27417243851555717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,8192,0.12171822124057347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,256,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,1536,0.01016266644001007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,8192,0.0491484436723921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,128,0.02295200030008952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,5120,0.01663644446267022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,3072,0.030415998564826116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,32,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,2048,0.07365422116385566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,4096,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,4096,0.010223111344708337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,64,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,1536,0.006823110911581252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,16384,0.03503200080659654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,3072,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,2048,0.3098497655656603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,32,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,128,0.002303999952144093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,12288,0.03279377685652839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,12288,0.030794666873084173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,7168,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,6144,0.042302221059799194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,128,0.008405333591832055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,3584,0.08397244744830662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,256,0.006849777781301075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,2048,0.024085332949956257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,128,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,4096,0.007830222447713217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,7168,0.04338577720854017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,2048,0.07260711325539483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,3584,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,51200,14.669981214735243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,2048,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,4096,0.1587537791993883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,51200,0.09861866633097331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,16384,0.3780213461981879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,3584,0.013711999687883588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,10240,0.2595457765791151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,7168,0.3760080072614882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,8192,0.029206222958034937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,65536,0.22007466687096489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,8192,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,1536,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,2048,0.025037333369255066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,768,0.01754844519827101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,7168,0.04029777646064758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,3072,0.007871111234029135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,128,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,3584,0.0943137804667155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,6144,0.10152177678214179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,16384,0.3343368901146783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,10240,0.04268177681499057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,1024,0.0031946665710873077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,10240,0.027963555521435205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,7168,0.06506666872236463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,5120,0.09893244504928589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,3072,0.01202488856183158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,2048,0.007341333561473423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,6144,0.00923555592695872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,128,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,3072,0.029189334975348577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,7168,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,512,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,32,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,768,0.09921422269609238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,2048,0.015595555305480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,4096,0.034400890270868935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,6144,0.031011554929945204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,8192,0.02494844463136461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,7168,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,10240,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,4096,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,768,0.019762666689025033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,2048,0.0678008861011929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,3072,0.06209333075417412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,1536,0.05692355500327217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,2560,0.02646044393380483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,3584,0.02323644525474972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,51200,1.422578705681695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,12288,0.015786666009161208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,3584,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,768,0.021503999829292297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,64,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,2048,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,256,0.01700177788734436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,7168,0.022678222921159532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,2560,0.008256889051861232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,1024,0.032540443870756365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,12288,0.0551528897550371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,1536,0.06864000029034086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,128,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,6144,0.06505244308047824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,32,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,2048,0.011337777806652917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,65536,0.06903644402821858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,12288,0.013535999589496188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,512,0.015312888556056552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,6144,1.095253308614095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,128,0.02158044444190131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,51200,0.19329333305358887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,2048,0.00889244427283605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,3072,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,16384,0.04353422257635328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,16384,0.767756462097168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,3072,0.03250044584274292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,10240,24.886588202582463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,16384,0.0559519992934333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,10240,0.9321902592976888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,2048,0.024951110283533733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,512,0.011710222396585675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,1536,0.020243555307388306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,2560,0.02035466664367252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,256,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,51200,0.055938667721218534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,256,0.027064889669418335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,768,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,3072,0.009569777382744683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,256,0.22543554835849336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,4096,0.5682106547885472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,12288,0.04907733201980591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,2048,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,1024,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,1024,0.010065777434243096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,10240,0.039509332842297025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,256,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,3072,0.014527999692493014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,128,0.027620444695154827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,7168,0.010263111028406356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,2048,0.04378755556212532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,51200,0.31365688641866046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,51200,0.7502871089511447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,10240,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,2048,0.5194097624884711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,512,0.019976889093716938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,768,0.03719200028313531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,7168,0.7422204547458224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,1536,0.04951999915970696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,256,0.018631999691327412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,8192,0.09533777501848008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,51200,0.2819173336029053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,1536,0.010565333068370819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,16384,0.02115999990039402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,512,0.004006222304370669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,512,0.016527111331621807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,128,0.015963556038008798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,32,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,16384,0.2707102298736572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,768,0.016026667422718473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,2560,0.016269332832760282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,65536,0.06400622261895074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,7168,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,51200,0.060941331916385226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,12288,1.2262747022840712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,6144,0.12336977322896321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,65536,0.09196444352467854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,768,0.024031110935741003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,2560,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,65536,0.25036178694831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,8192,0.02862488892343309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,5120,0.029151111841201782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,768,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,256,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,3584,0.01735555628935496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,2560,0.030431999100579157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,2048,0.03152622116936578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,256,0.015634665886561077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,6144,0.12585333983103433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,5120,0.018773333893881906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,3584,0.01498311095767551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,16384,0.02944444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,7168,0.055826663970947266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,6144,0.4016257656945123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,12288,0.15362932946946886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,2560,0.1231351163652208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,8192,0.06955022282070585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,1536,0.08476533492406209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,4096,0.04907733201980591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,3072,0.07168710894054837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,1024,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,128,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,8192,0.016557332542207506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,1024,0.014197332991494073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,512,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,1024,0.02899466620551215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,8192,0.010430222584141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,10240,0.01219911128282547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,2048,0.004893333133724001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,6144,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,2560,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,256,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,1024,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,3584,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,65536,0.2501937813229031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,5120,0.02237777743074629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,4096,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,65536,0.05788355403476291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,65536,0.32835200097825795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,3072,0.030974222554100886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,128,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,1024,0.06244533591800266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,5120,0.032532443602879844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,5120,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,12288,0.03493244449297587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,7168,0.03327555457750956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,32,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,7168,0.009566222627957663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,3072,0.027253333065244887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,4096,0.05759288867314657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,7168,0.05763022104899088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,1536,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,12288,0.04353955719206068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,2560,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,65536,0.17062311702304414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,10240,0.05176266696718004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,10240,0.31548889478047687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,16384,0.425176011191474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,128,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,51200,0.12350755267673069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,3584,0.08259644773271349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,3072,0.0074275558193524676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,32,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,3584,0.011424000064531961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,512,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,4096,0.012915555801656512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,6144,0.02160533269246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,6144,0.00925333301226298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,65536,0.019468444916937087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,6144,0.042304890023337476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,5120,0.011346666349305047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,2048,0.03055466545952691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,768,0.1506942245695326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,16384,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,768,0.008722667064931657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,16384,0.022124444444974262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,4096,0.0038319999972979226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,16384,0.048419555028279625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,10240,0.03704444567362467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,1024,0.0038693332009845306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,768,0.008376888930797577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,3072,0.02492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,51200,0.05376266770892673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,16384,0.0583217806286282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,5120,0.008721777962313758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,6144,0.01145066652033064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,2560,0.018966222802797954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,256,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,10240,0.02608888844648997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,51200,0.05303377906481425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,128,0.02095199955834283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,1536,0.0651715530289544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,3072,0.007418666448858049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,1536,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,4096,0.0251884443892373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,5120,0.00905777762333552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,10240,0.013083555632167392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,51200,0.4918524424235026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,32,0.007054222126801808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,256,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,8192,0.2555777761671278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,1536,0.004511111312442356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,10240,0.036509331729676985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,512,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,512,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,2560,0.05987289216783312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,32,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,2048,0.02804533309406704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,51200,0.4161626762813992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,128,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,7168,0.024354666471481323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,256,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,8192,0.12661955091688368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,256,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,16384,0.6628906461927626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,3584,0.04530044396718343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,16384,0.026036444637510512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,768,0.11171021726396348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,10240,0.006506666541099548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,65536,0.9683769014146594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,65536,0.07416622506247626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,3584,0.028872887293497723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,3072,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,1536,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,1536,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,7168,0.22994934187995064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,2048,0.008025777836640676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,1536,0.029736889733208552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,5120,2.387879901462131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,12288,0.06460622284147474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,10240,0.4170791043175592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,12288,0.014913777510325113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,16384,0.14865956041547987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,7168,0.4559013578626845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,65536,0.5366177558898926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,2560,0.07573599947823419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,51200,0.014761777387724983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,12288,0.18443644046783447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,2048,0.008563555777072906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,2048,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,4096,0.3279342121548123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,1024,0.017519111434618633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,10240,0.08678311109542847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,10240,0.16059822506374782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,51200,0.07038488652971056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,7168,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,1536,0.008444444172912175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,16384,0.02046133412255181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,2048,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,12288,0.03425777620739407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,7168,0.029376887612872656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,65536,0.10167911317613389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,2048,0.0063946665161185795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,1024,0.02132355504565769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,5120,0.022795556320084467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,6144,0.013043555948469373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,10240,0.016521778371598985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,256,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,32,0.0023111111174027124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,51200,0.05859555800755819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,16384,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,512,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,128,0.018239999810854595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,7168,0.028647998968760174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,5120,0.009471110999584198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,2560,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,10240,0.04901688959863451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,4096,0.012972444295883179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,65536,0.10432266526752049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,1536,0.04201511210865445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,3584,0.03756711218092177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,51200,0.05799111392762926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,1024,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,2048,0.00869066682126787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,65536,1.8990675608317058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,256,0.015788444214397006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,768,0.030324445830451116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,4096,0.10264533095889622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,4096,0.015600000818570455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,256,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,512,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,7168,0.011695110963450538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,4096,0.22173421912723115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,2048,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,4096,0.07058222426308526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,7168,0.7599129146999783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,5120,0.0123511114054256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,51200,0.7301671240064832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,10240,0.3756835460662842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,51200,0.2789662149217394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,51200,0.4321635564168294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,8192,0.048022223843468555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,3584,0.10288533237245347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,3584,0.3119262324439155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,3584,0.009400000174840292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,65536,0.06364177995257907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,1024,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,1536,0.022229333718617756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,512,0.04227022329966227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,2560,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,65536,0.2897128793928358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,6144,0.0295413335164388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,768,0.049636443456014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,5120,0.018883556127548218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,512,0.015218666858143277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,16384,0.02609688871436649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,10240,0.10465421941545273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,2048,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,4096,0.09562044673495823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,10240,0.017581333716710407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,3584,0.017528888252046373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,768,0.010450666977299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,12288,0.016548444827397663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,16384,0.05282310975922478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,128,0.0254711111386617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,51200,0.115100449985928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,4096,0.13232266902923584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,10240,0.02088711162408193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,16384,0.1687715583377414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,5120,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,12288,0.013254221942689685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,8192,0.04007911019855075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,10240,0.03126577867401971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,3584,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,4096,0.02767733401722378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,6144,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,768,0.016025778320100572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,1024,0.006528888725572162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,2048,0.005306666509972678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,512,0.025607999828126695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,64,0.0035955554081334006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,12288,0.037478221787346735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,2560,0.012732444538010491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,3072,0.032124443186653986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,5120,0.023324444890022278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,2560,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,3072,0.1431386603249444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,2048,0.024088000257809956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,768,0.017525333497259352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,12288,0.014371555712487964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,768,0.025014221668243408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,1024,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,8192,0.010059555371602377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,7168,0.030574222405751545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,64,0.020230222079488967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,6144,0.01811288959450192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,10240,0.4847271177503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,51200,1.5801813337537978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,256,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,256,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,8192,0.028021332290437486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,65536,0.08123022317886353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,32,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,1536,0.006698666761318843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,2560,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,64,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,256,0.016304888659053378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,51200,0.16426400343577066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,128,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,256,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,64,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,2048,0.1461697816848755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,3072,0.08534222178988987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,6144,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,10240,0.07241688834296332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,768,0.015452444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,32,0.014470222923490735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,2560,0.04574399855401781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,51200,2.5030222998725042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,768,0.03128355410363939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,3584,0.024903999434577093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,512,0.01904000010755327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,6144,0.024687111377716064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,7168,2.7180887858072915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,6144,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,5120,0.2968568801879883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,2048,0.4225422276390924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,1024,0.011179555621412067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,512,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,4096,0.015300444430775113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,7168,0.014750222365061441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,51200,0.11384444766574436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,1024,0.0038328886859946777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,2560,0.03287911083963182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,51200,0.03938044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,1024,0.008397333323955536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,2560,0.006656000183688269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,64,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,128,0.005306666509972678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,512,0.03964444332652622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,2560,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,12288,0.08971822261810303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,3072,0.036743998527526855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,5120,0.010405333505736457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,768,0.013277333643701343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,12288,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,6144,0.06417600313822429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,128,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,256,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,512,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,6144,0.05088177654478285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,6144,2.5889627668592667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,2048,0.090993775261773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,12288,2.407298617892795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,10240,0.0573671129014757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,16384,0.373074664009942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,1024,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,2048,0.019215111931165058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,512,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,65536,0.1408417754703098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,8192,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,8192,0.07689688603083293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,256,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,10240,0.025957332717047796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,1536,0.0068400001360310455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,16384,0.09435200028949314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,1536,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,12288,0.0576711098353068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,8192,0.014325333966149224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,3584,0.011315555208259158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,1536,0.006516444600290722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,128,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,1536,0.04740000102255079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,1536,0.003531555748648114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,7168,0.033505777517954506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,3072,0.01255555533700519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,5120,0.040595554643207125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,128,0.0018133332745896445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,1536,0.021781333618693884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,12288,0.03542666633923849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,7168,0.1110888918240865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,16384,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,8192,0.05650844176610311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,3584,0.018560000591807894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,4096,0.016368000043763053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,5120,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,3584,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,51200,15.080710517035591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,2048,0.11254310607910156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,7168,0.010568000376224518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,128,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,256,0.01753866672515869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,128,0.023781332704755995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,3072,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,2048,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,10240,0.11029422283172607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,512,0.017816889617178176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,3072,0.16704711649152967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,65536,0.13336977693769667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,51200,0.053977777560551964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,1536,0.40695820914374453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,128,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,65536,0.07629777987798055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,51200,1.4350577460394964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,16384,0.11873688962724473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,8192,1.723398208618164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,3072,0.030071109533309937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,7168,0.028147555059856836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,10240,0.3191502094268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,1536,0.0063475556671619415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,2048,0.0136817776494556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,7168,0.12222933769226074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,128,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,7168,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,51200,0.11627822452121311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,128,0.01181688904762268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,7168,0.00978488889005449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,65536,0.08725511365466648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,12288,0.051227556334601514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,4096,0.011110222174061669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,8192,0.04249688982963562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,65536,0.016756445169448853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,16384,0.03465777635574341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,64,0.03912533322970072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,65536,1.1012747022840712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,4096,0.03947910997602675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,10240,1.526786698235406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,7168,0.07644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,8192,0.03016177813212077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,12288,0.044066667556762695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,8192,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,768,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,512,0.012688889271683164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,65536,0.032456000645955406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,128,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,51200,0.812354670630561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,51200,0.12383377552032471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,7168,0.010289777484205034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,768,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,3584,0.01074577785200543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,8192,0.26815822389390737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,64,0.01329955541425281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,51200,0.04961422085762024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,64,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,32,0.028791109720865887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,1024,0.003354666754603386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,2560,0.007530666887760162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,128,0.01625333395269182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,3072,0.05559555689493815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,16384,0.01642933322323693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,1024,0.013824889229403602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,7168,0.12178311083051895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,128,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,1024,0.09030489126841228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,6144,0.0185253338681327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,1024,0.018631999691327412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,32,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,8192,0.011232888533009423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,256,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,16384,1.556311077541775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,10240,0.027076444692081873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,8192,0.019401777121755812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,128,0.01628977722591824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,512,0.003565333369705412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,12288,0.08050577508078681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,5120,0.09368355406655206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,51200,0.06365422407786052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,3072,0.03464622298876444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,128,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,8192,0.024382222029897902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,2048,0.0355040000544654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,4096,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,1024,0.004267555558019214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,4096,0.09750044345855713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,3584,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,768,0.01994311147265964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,6144,0.028063999281989202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,16384,0.15389245086246067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,65536,0.16463288995954725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,5120,0.05776266919242012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,65536,0.6367911232842339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,16384,0.09708799918492635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,6144,0.04933955603175693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,2560,0.011307555768224927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,128,0.0046533333758513136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,12288,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,2048,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,10240,0.2958622243669298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,8192,0.03329688972897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,10240,0.4687155617607965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,7168,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,3584,1.1385244793362088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,8192,0.016858667135238647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,12288,0.022660444180170696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,32,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,3072,0.11230933666229248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,32,0.01277599980433782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,4096,0.20513333214653862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,64,0.026944888962639704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,2048,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,1536,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,65536,0.4751146634419759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,3072,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,768,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,10240,0.4203226566314697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,65536,0.19120800495147705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,32,0.0028719999310043124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,4096,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,2560,0.047296000851525195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,8192,0.06560533576541476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,768,0.005670222143332164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,7168,0.04184800055291918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,10240,0.03353600038422479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,2560,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,1024,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,4096,0.05067288875579834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,1024,0.0035582222044467926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,512,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,2560,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,256,0.002700444517864121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,2048,0.011191999746693505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,6144,0.012524444195959302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,1024,0.24956443574693468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,1536,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,65536,0.1340657737520006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,16384,0.013497778111033969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,7168,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,10240,0.16315200593736437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,8192,0.01862844493654039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,5120,0.16682399643792045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,1024,0.0063164445261160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,12288,0.004929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,2560,0.07363822062810262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,10240,0.5783155759175619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,128,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,5120,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,768,0.020598222812016804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,2560,0.037672887245814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,128,0.006874666859706243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,3072,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,12288,0.034640000926123724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,128,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,51200,0.04723644587728712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,10240,5.103556315104167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,128,0.01869244376818339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,4096,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,1536,0.012200888660218982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,1536,0.0063768890168931745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,3584,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,2560,0.017789334058761597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,5120,0.0064399999876817065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,5120,0.5386337704128689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,512,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,768,0.016368000043763053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,3584,0.029409776131312054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,3584,0.09965688652462429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,3072,0.11451200644175212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,2048,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,10240,0.06625599993599786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,8192,0.010414222048388587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,6144,0.02571022179391649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,2560,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,2048,0.004597333156400257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,12288,0.12744266457027859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,2560,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,16384,0.9884577857123481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,7168,0.05317866802215576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,8192,0.0562737782796224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,10240,0.026727999250094097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,1024,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,1536,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,1536,0.01167555567291048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,3584,0.02421155571937561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,51200,0.2477653291490343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,3584,0.025607110725508794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,65536,0.6917200088500977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,1536,0.01850844422976176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,768,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,10240,0.013354666531085968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,12288,0.16048088338640001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,5120,0.17922400103674996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,1536,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,1024,0.03495733274353875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,4096,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,8192,0.12351022826300727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,3584,0.16707555452982584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,6144,0.02476355599032508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,3584,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,256,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,10240,0.03691733214590285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,512,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,10240,0.015180443723996481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,768,0.06018666426340739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,1536,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,2560,0.22401422924465605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,5120,0.05267644590801663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,128,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,768,0.026177777184380427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,6144,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,64,0.024688889582951862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,6144,0.018464888135592144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,12288,0.03453244434462653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,51200,0.42415467898050946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,768,0.01793333391348521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,6144,0.01460444430510203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,3584,0.021596444977654353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,7168,0.6733039749993218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,12288,0.018001778258217704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,10240,0.01752622259987725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,65536,0.06525866852866279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,256,0.024494222468800012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,65536,0.2062631050745646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,3584,0.03134844369358487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,1024,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,7168,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,10240,1.5945599873860676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,16384,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,3072,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,7168,0.04277333286073473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,1024,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,1024,0.004223111189073986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,3584,0.02873777680926853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,64,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,3072,0.02583911187118954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,7168,0.08080089092254639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,64,0.0031848889258172777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,1024,0.4697013431125217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,1536,0.00674933319290479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,5120,0.03826933436923557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,256,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,8192,0.06883200009663899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,256,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,128,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,10240,0.09956889020072089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,6144,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,1024,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,6144,0.363954676522149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,8192,0.03457066747877333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,2560,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,256,0.021743999587164983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,8192,0.011398221883508893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,8192,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,5120,0.058542225095960826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,3072,0.015474667151769003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,768,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,6144,0.12868532869550917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,768,0.006671999891599019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,1536,0.008188444707128737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,2048,0.006385777973466449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,768,0.08399822314580281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,12288,0.09321066406038071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,3584,0.024785778588718835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,512,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,1024,0.028088887532552082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,7168,0.016433777080641854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,128,0.006694222076071634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,8192,0.010987555815113915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,8192,0.05151199963357714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,1024,0.010455110834704505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,51200,0.4849431249830458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,1536,0.007675555845101674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,768,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,3584,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,6144,0.009740444521109263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,12288,0.06482755475574069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,1024,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,2560,0.02784444557295905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,16384,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,65536,2.5609573788113065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,1536,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,2048,0.007378666765160031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,768,0.03045155604680379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,64,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,128,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,16384,0.02055199940999349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,5120,0.020791111720932853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,16384,0.033024887243906655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,3072,0.010601777997281818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,128,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,12288,0.12951821751064724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,256,0.01680355601840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,3072,0.4072515434688992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,6144,0.01035199976629681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,7168,0.22970755894978842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,4096,0.03703822361098395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,5120,0.009558222360081142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,2048,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,1024,0.030973331795798406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,64,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,768,0.03838133480813768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,64,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,10240,0.06409155660205416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,16384,0.5744675530327691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,8192,0.011541333463456897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,256,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,12288,0.03471733464135064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,8192,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,1024,0.009930666950013902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,1024,0.0063546668324205614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,6144,0.05811022387610542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,51200,0.868500444624159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,5120,0.02481599979930454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,1536,0.006049777898523543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,2048,0.025009777810838487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,32,0.01942755613062117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,1024,0.026032888227038916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,128,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,2048,0.046331554651260376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,7168,0.027855998939938013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,512,0.026664889521068994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,65536,0.1634675529268053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,512,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,2560,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,6144,0.013376000026861826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,51200,0.05406399899058872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,7168,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,512,0.016809778081046212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,3072,0.07100977500279744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,5120,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,8192,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,512,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,12288,0.039792001247406006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,3584,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,768,0.02952444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,8192,0.027216888136333887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,6144,0.012372444073359171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,51200,0.017683555682500202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,512,0.02421066661675771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,512,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,4096,0.06782755586836073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,2048,0.020238222347365487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,5120,0.016935999194780987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,2048,0.004518222063779831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,7168,0.15386133723788792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,4096,0.07392533620198567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,7168,0.04060533311631944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,64,0.0018764444523387486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,16384,0.01605688863330417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,4096,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,4096,0.008604444563388824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,128,0.030508445368872747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,1024,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,12288,0.054139554500579834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,16384,0.2314293384552002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,10240,0.04396888944837782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,2560,0.012480888929631976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,512,0.00564177789621883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,3072,0.037792887952592634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,12288,0.21827288468678793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,10240,0.07026577658123441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,51200,0.052608887354532875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,5120,0.31572622723049587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,256,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,10240,0.049345778094397656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,16384,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,3072,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,64,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,32,0.017879999346203275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,256,0.13025332821740046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,6144,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,51200,2.926140467325846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,3072,0.02404266595840454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,6144,0.15916089216868082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,3584,0.017888888716697693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,2048,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,512,0.02534666657447815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,8192,0.09026400248209636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,16384,0.03473511007097032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,12288,0.03415111038419936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,32,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,5120,0.09113422367307876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,6144,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,1536,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,32,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,2560,0.015004444453451367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,5120,0.03253688746028476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,12288,0.38884978824191624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,1536,0.0042133331298828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,512,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,4096,0.014909333652920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,3584,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,768,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,51200,0.11286932892269558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,512,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,1024,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,768,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,12288,0.03989244500796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,3072,0.07952621910307142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,65536,0.19945955276489258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,4096,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,3584,0.05712622404098511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,12288,0.013474666410022311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,2048,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,3584,0.01407999959256914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,7168,0.016745777593718637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,2560,0.4418328868018256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,65536,0.09038844373491074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,2048,0.018957333432303537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,8192,0.0185244447655148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,7168,0.037267555793126426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,65536,0.12940088907877603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,5120,0.04381688767009311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,768,0.0033964444365766314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,6144,0.00923555592695872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,64,0.01163200040658315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,8192,2.020696004231771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,2048,0.02787289023399353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,65536,2.0474355485704208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,16384,0.17999288770887586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,12288,0.02540622154871623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,10240,0.05203910999827915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,5120,0.015113777584499784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,6144,0.1166951126522488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,8192,0.012445333103338877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,128,0.012033777932325998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,3584,0.012306667036480375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,128,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,10240,0.09063911437988281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,51200,0.05751022365358141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,32,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,256,0.007050666544172499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,4096,0.0558604465590583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,8192,0.06811644633611043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,512,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,1536,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,1024,0.011431111229790581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,3072,9.86676279703776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,65536,0.01425244410832723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,12288,0.032518221272362605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,768,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,4096,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,6144,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,10240,0.02537600033813053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,4096,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,2048,0.020287111401557922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,1024,0.014464888307783337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,128,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,7168,0.09052622318267822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,64,0.01756444407833947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,7168,0.013051555388503604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,65536,0.08550577693515354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,2560,0.054514666398366295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,7168,0.7917457686530219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,65536,0.0544417765405443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,6144,0.009714666340086196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,16384,0.048624889718161694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,10240,0.03826577795876397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,128,0.004587555511130227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,16384,0.3015555540720622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,768,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,2560,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,768,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,65536,0.027626666757795546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,512,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,12288,0.04229777720239428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,12288,0.01571111049917009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,32,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,32,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,7168,0.04022666811943054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,51200,0.05439644389682346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,51200,0.08149333132637872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,2560,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,8192,0.07212089167700873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,512,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,1024,0.003863999827040566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,10240,0.0592213339275784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,512,0.01276177747382058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,5120,0.03051111102104187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,1024,0.06423111094368829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,256,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,64,0.0017911110901170308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,5120,0.021828444467650518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,16384,0.029450668228997126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,3584,0.013574221895800697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,7168,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,3072,0.6463564236958822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,128,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,7168,0.029578665892283123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,5120,0.02141422198878394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,12288,0.026826666461096868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,128,0.028034667174021404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,1536,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,8192,0.023676445086797077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,1536,0.054367111788855664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,16384,0.2774888939327664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,2048,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,2048,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,32,0.010434666441546546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,3072,0.026113778352737427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,1024,0.031535999642478094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,3072,0.08233422040939331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,4096,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,2048,0.008037333687146505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,1536,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,64,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,256,0.014154665999942355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,7168,0.039334221018685236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,3072,0.033987555238935686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,65536,3.515030119154188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,51200,0.12480889426337348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,32,0.021550221575631037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,512,0.08511733346515232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,5120,0.06522755490409003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,512,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,512,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,12288,0.030524445904625788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,3584,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,6144,0.004600888739029567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,6144,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,2560,0.256839116414388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,6144,0.05668444765938652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,5120,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,64,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,7168,0.05556178092956543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,5120,0.01295377810796102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,12288,0.017308443784713745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,51200,1.9981733957926433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,65536,0.06484622425503202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,768,0.006855999843941794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,12288,0.07271022266811795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,1536,0.021128889587190416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,3072,0.04512888855404324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,5120,0.028567999601364136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,10240,0.5686062177022299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,5120,0.05298666821585762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,768,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,3584,0.015440000428093804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,3584,0.02954044606950548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,64,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,768,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,32,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,2048,0.041913777589797974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,2048,0.06721244255701701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,4096,0.039255999856524997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,64,0.06631466415193346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,3584,0.011869333684444427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,768,0.011711999773979187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,6144,0.10315910975138347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,768,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,8192,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,16384,0.32479821311102974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,6144,0.022467555271254644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,5120,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,256,0.010879110958841113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,16384,0.020586666133668687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,5120,0.029501332177056208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,65536,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,128,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,51200,0.49725691477457684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,7168,0.014019555515713163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,1536,0.011201777391963534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,12288,0.03994311226738824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,512,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,256,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,32,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,2560,0.01700266698996226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,6144,0.15755555364820692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,8192,0.03073599934577942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,10240,0.08099644713931613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,4096,0.025846223036448162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,7168,0.027983112467659846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,512,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,8192,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,4096,0.034394668208228216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,5120,0.07607199748357137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,512,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,7168,0.0358515547381507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,32,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,5120,0.02234222160445319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,65536,0.25280621316697865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,5120,0.2122488949033949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,4096,0.030509332815806072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,51200,0.04924355612860786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,51200,0.0473288893699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,65536,0.38812534014383954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,3584,0.03555288910865784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,1536,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,16384,0.2633226712544759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,6144,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,3584,0.018159111340840656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,4096,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,5120,0.029025776518715754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,512,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,2048,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,1024,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,2048,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,2560,0.07818133301205106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,1024,0.04168088899718391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,128,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,64,0.0017911110901170308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,10240,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,2560,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,65536,0.07499288850360446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,1536,0.03923111160596212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,7168,0.03428355521625943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,12288,0.03754399882422553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,6144,0.08923288848665024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,4096,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,12288,0.07715111308627658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,10240,2.4445955488416886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,16384,0.03889777925279405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,12288,2.7090293036566844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,16384,0.23261777559916177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,2048,0.028241776757770117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,16384,0.200799994998508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,2048,0.010784000158309937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,3072,0.03922222057978312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,1536,0.013549333645237817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,1024,0.015055111712879606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,1536,0.016784888174798753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,16384,0.008692444198661381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,3072,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,1536,0.16524622175428602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,8192,0.18411466810438368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,65536,12.237350463867188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,12288,0.22147733635372588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,3584,0.04881599876615736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,10240,0.021651556094487507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,8192,0.024478221933046978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,51200,0.020776889390415616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,12288,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,256,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,512,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,6144,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,2560,0.07284266418880887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,5120,0.18887466854519316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,512,0.0037946667936113146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,16384,0.30637065569559735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,16384,3.1366009182400174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,2048,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,32,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,512,0.03571022219128079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,10240,0.1987182166841295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,2048,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,3072,0.8696479797363281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,1024,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,2048,0.018564444449212816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,2560,0.024940444363488093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,5120,0.07816355758243136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,65536,0.14768977959950766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,1024,0.01789155602455139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,51200,0.45078399446275497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,1536,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,128,0.018613333503405254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,12288,0.12591289149390325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,7168,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,6144,0.05505599909358554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,7168,0.018792889184421964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,6144,0.02443733314673106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,64,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,8192,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,7168,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,1536,0.024944000773959692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,3072,0.38722578684488934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,2048,0.02769333289729224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,4096,0.04804889029926724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,8192,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,2560,0.036383999718560114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,10240,0.04522222280502319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,128,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,3072,0.03385333220163981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,2560,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,6144,0.009529777699046666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,3584,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,10240,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,12288,0.1724639998541938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,128,0.019894222418467205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,8192,0.0440631111462911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,12288,0.03242666522661845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,65536,0.11555111408233643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,5120,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,1024,0.015572445260153877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,4096,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,32,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,768,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,65536,0.3799288802676731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,5120,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,3584,0.024695111645592585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,6144,0.33818401230706113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,128,0.016902221573723685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,256,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,8192,0.09989955690171982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,6144,0.042211555772357516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,128,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,256,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,5120,0.01920622256067064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,3072,0.02334755493534936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,8192,0.030926220946841772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,3072,0.09148355325063069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,16384,2.823452419704861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,8192,0.0731031099955241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,128,0.016908443636364408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,65536,14.691650390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,8192,0.014115555418862237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,2048,0.029000888268152874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,2048,0.014743111199802823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,10240,0.01607111096382141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,6144,0.1475048859914144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,6144,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,1536,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,4096,0.7505982187059191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,1024,0.012191111014948951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,65536,0.32886489232381183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,16384,1.9031946394178603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,3072,0.13994488451215956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,3584,0.10394489102893406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,4096,0.013279111021094851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,8192,0.04617955618434482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,6144,0.025119110941886902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,2560,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,4096,0.025916443930731878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,8192,0.02661422226164076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,1536,0.04541688826349047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,3584,0.02055999967787001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,65536,1.1969440248277452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,32,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,3584,0.016775111357371014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,4096,0.03822844558291965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,4096,0.01590577761332194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,65536,0.5973982281155056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,16384,0.020612445142534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,256,0.003175111073586676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,1536,0.013192888763215808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,12288,0.03247911069128249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,1536,0.019762666689025033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,5120,0.02360800074206458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,2048,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,1024,0.06419733497831556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,4096,0.011811555259757571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,7168,0.17692977852291533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,16384,0.03327733278274536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,2048,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,10240,0.012033777932325998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,64,0.0035111109415690103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,8192,0.08950044711430867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,2560,0.34082844522264266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,1536,0.008751110898123847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,7168,0.04675999946064419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,10240,0.4367857774098714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,8192,0.019481778144836426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,6144,0.021295999487241108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,12288,0.01867911054028405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,65536,0.14766132831573486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,6144,0.15485066837734646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,4096,0.13576532734764948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,51200,0.0135057775510682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,12288,0.028599109914567735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,768,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,5120,0.04587822159131368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,1024,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,51200,0.24561601214938691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,65536,0.27813956472608775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,2048,0.016057777735922072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,512,0.005723555468850666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,65536,0.1587511168585883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,512,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,7168,0.12682578298780653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,51200,0.05056888858477274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,4096,0.032497776879204646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,3584,0.02497422198454539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,12288,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,64,0.006801777829726537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,64,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,10240,0.08459644185172187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,51200,0.10282755560345119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,4096,0.02808622188038296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,2048,0.02540622154871623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,12288,0.030942221482594807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,32,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,5120,0.12298133638170029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,51200,0.21856533156500924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,12288,0.018239110708236694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,65536,0.22526221805148652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,51200,0.19080621666378447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,7168,0.13135643800099692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,10240,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,8192,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,16384,0.37760178248087567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,3072,0.04764444298214383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,2048,0.007787555456161499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,768,0.0092622223827574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,3072,0.05284088850021362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,8192,0.02365511159102122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,12288,2.5868356492784286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,5120,2.480342229207357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,3072,0.01846577723821004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,65536,0.19988089137607148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,512,0.007583111524581909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,3072,0.04279377725389269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,7168,0.07463910844590929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,32,0.026707554856936138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,3072,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,6144,0.02648266653219859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,5120,0.02977333466211955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,6144,1.8460648854573567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,4096,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,8192,0.17406311300065783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,5120,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,12288,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,128,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,65536,0.5407884385850694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,10240,0.056167112456427686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,8192,0.08014044496748182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,64,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,8192,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,5120,0.027948445743984644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,768,0.012562666502263812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,4096,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,2560,0.01148977792925305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,65536,0.13275378280215794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,5120,0.5177813106113011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,5120,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,1536,0.03906133439805772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,16384,0.10404088762071396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,3584,0.0428435570663876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,6144,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,4096,0.03865155577659607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,512,0.015263999501864115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,16384,0.011955555942323474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,6144,0.047091556919945605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,768,0.01350044459104538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,1024,0.012344888514942594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,51200,0.05672978030310737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,3584,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,512,0.009747555686367882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,256,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,16384,0.04840355449252658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,12288,0.04343555702103508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,7168,0.058083554108937584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,2560,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,256,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,32,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,6144,0.5551057921515571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,4096,0.011174221833546957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,3584,0.08283999893400404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,1024,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,4096,0.013363555901580386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,2560,0.05851111147138807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,7168,0.03215288784768846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,1536,0.170815110206604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,256,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,16384,0.5560906728108724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,8192,0.11933599578009711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,65536,2.3614603678385415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,7168,0.06869066423839994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,5120,0.01657511128319634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,512,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,128,0.020012444920010038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,65536,0.271871116426256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,65536,0.027644443843099806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,1024,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,2048,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,512,0.020015110572179157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,256,0.021967111362351313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,4096,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,5120,0.23253154754638672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,64,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,64,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,1536,0.020048000746303134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,7168,0.0633208884133233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,10240,1.059591081407335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,3584,0.021192888418833416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,256,0.03752533263630337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,65536,1.1190417607625325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,5120,0.027241778042581346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,10240,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,768,0.06891555256313749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,256,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,32,0.030930668115615845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,8192,0.18414666917588976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,2048,0.025605332520272996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,3584,0.01794488893614875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,65536,0.08085866769154866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,5120,0.03077510992685954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,1024,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,2048,0.011404444773991903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,6144,0.02417244513829549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,12288,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,10240,0.016456888781653512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,256,0.006794666664467917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,5120,0.038570665650897555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,12288,0.028142223755518597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,3584,0.00980355590581894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,32,0.02222488820552826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,12288,0.09806755516264175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,12288,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,128,0.08970044718848334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,768,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,10240,2.0582461886935763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,6144,0.04649688800175985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,6144,0.01017155581050449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,256,0.02915289004643758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,3584,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,16384,0.03757066528002421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,6144,0.009775111244784461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,65536,0.13888622654808894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,51200,0.06833333439297147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,7168,0.04378933376736111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,256,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,32,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,8192,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,1536,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,10240,0.027330666780471802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,7168,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,8192,0.04138488901986016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,10240,0.01556799974706438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,10240,0.24260176552666557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,2048,0.03376444511943393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,6144,0.04179999894566006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,512,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,1536,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,10240,0.029109332296583388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,128,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,3584,0.5768319765726725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,5120,0.011144888897736868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,256,0.01756088932355245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,2560,0.02088177700837453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,6144,0.03738933470514085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,51200,0.04581955406400892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,51200,0.717573324839274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,12288,0.013415111435784234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,1024,0.013037333057986366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,3584,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,16384,0.0365608897474077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,1536,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,4096,0.029470221863852605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,16384,0.031930665175120033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,512,0.008351999852392409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,512,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,4096,0.026851556367344324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,51200,0.8187084197998047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,10240,0.1883395512898763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,12288,0.062195552719963916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,2560,0.010503110786279043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,8192,0.074526223871443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,3072,0.078311112191942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,256,0.015266666809717814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,4096,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,3072,0.015455999308162265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,128,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,8192,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,2560,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,1536,0.02033866610791948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,8192,0.14930933051639134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,51200,0.2963635656568739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,2560,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,2048,0.021253334151373968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,2048,0.024383111132515803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,3584,0.0075599998235702515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,16384,0.02541422181659275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,8192,0.018848000301255118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,512,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,128,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,1024,0.015082667271296183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,256,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,4096,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,512,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,4096,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,768,0.010111110905806223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,51200,0.038133333126703896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,51200,0.47861512502034503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,6144,0.012118221984969245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,10240,0.2838853465186225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,16384,0.17539200517866346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,7168,0.010418666733635796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,12288,0.041470223002963595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,51200,0.8450524542066785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,2560,0.0123831108212471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,768,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,128,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,3072,0.011352000137170156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,1536,0.06790133317311604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,2048,0.01555377741654714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,512,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,32,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,768,0.016688000824716356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,256,0.01110311100880305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,10240,1.2223208745320637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,16384,0.026260443859630164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,3584,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,4096,1.5835928387112086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,2048,0.30047112041049534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,12288,0.5384089152018229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,4096,0.04261599977811178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,7168,0.09711288743548924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,5120,0.05544711152712504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,1024,0.012718222207493253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,64,0.006684444430801604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,3072,0.04575022061665853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,51200,0.07410844167073567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,5120,0.008954666554927826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,51200,0.06484622425503202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,1536,0.15259110927581787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,2048,0.012450666891203986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,2560,0.023071999351183575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,51200,0.2656017674340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,5120,0.012526222401195101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,6144,0.035263111193974815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,512,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,51200,0.24840709898206922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,128,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,512,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,3584,0.022259554929203455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,10240,0.232335991329617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,7168,0.02296000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,10240,0.03785599933730231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,8192,0.05861866474151611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,12288,0.4076133304172092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,2048,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,5120,0.07453155517578125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,5120,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,12288,0.013635555903116861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,1536,0.007206222249401941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,12288,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,4096,0.048935112026002675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,1024,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,5120,0.011493333511882357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,3072,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,8192,0.05429511268933614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,1536,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,768,0.0038124442928367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,6144,0.02429155508677165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,16384,0.0762231085035536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,51200,0.2148195505142212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,4096,0.056948443253835045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,1536,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,128,0.004882666799757215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,5120,0.010616000327799054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,10240,0.044510222143597074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,128,0.01517777807182736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,768,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,2048,0.004284444368547863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,768,0.005330666485759947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,1024,0.04905777838495043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,2048,0.03607111175855001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,65536,0.16465777821011013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,51200,0.4280417760213216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,8192,1.4314951366848414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,3072,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,768,0.031222220924165513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,51200,0.1927493280834622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,2048,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,6144,0.0163102216190762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,32,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,16384,0.14207643932766384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,8192,0.018211555149820115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,3584,0.04723466767205132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,51200,0.9807937410142687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,2560,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,65536,0.09532177448272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,128,0.015199111567603217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,256,0.022346667117542688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,5120,0.02939644455909729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,512,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,65536,0.04714222086800469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,256,0.011811555259757571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,128,0.033610665135913424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,10240,0.015226667126019796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,7168,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,256,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,4096,0.08968089024225871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,2560,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,512,0.011707555916574268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,1024,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,512,0.019335110982259113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,3584,0.007340444458855524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,12288,0.02363555630048116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,2048,0.012772444221708508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,12288,0.0454328887992435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,1024,0.010106667048401302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,256,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,32,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,3072,0.12853866153293186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,768,0.023499555057949487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,768,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,512,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,12288,0.44711022906833225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,16384,3.98508792453342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,512,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,7168,0.01130222198035982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,6144,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,64,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,4096,0.010627555350462595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,1024,0.6027688980102539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,1536,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,7168,0.15443466769324407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,1024,0.0871608919567532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,1024,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,16384,0.04239022069507175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,16384,0.023242667317390442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,3072,0.06911466519037883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,2048,0.1243715551164415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,3584,0.04321155614323086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,512,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,12288,0.04518844352828132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,65536,9.375701904296875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,12288,0.29397066434224445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,2560,0.01314755529165268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,512,0.08703466918733384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,1536,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,3072,0.03599288728502061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,128,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,65536,0.04831822382079231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,8192,0.01143200033240848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,1536,0.003549333247873518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,1536,0.020558221472634208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,3072,0.015629332926538255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,32,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,51200,0.07867111100090875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,3072,0.07412355475955539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,64,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,8192,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,2560,0.007146666447321574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,12288,2.1459325154622397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,1024,0.019668444991111755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,10240,0.014874666929244995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,5120,0.03663377629386054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,51200,15.190544976128471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,51200,1.3934657838609483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,3072,0.10181333621342976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,2048,0.03190844588809543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,5120,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,64,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,2048,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,3584,0.02365511159102122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,7168,0.04668622215588888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,10240,0.011706666813956367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,5120,0.07218222485648261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,128,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,64,0.008363555702898238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,7168,0.020938666330443487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,2560,0.016395555602179635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,65536,0.08043644163343641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,2048,0.026983110441101923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,16384,2.716976801554362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,8192,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,1024,0.045695109499825366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,32,0.026198221577538386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,12288,0.03212889035542806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,1024,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,1024,0.0181040002240075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,16384,0.11163289017147487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,12288,0.09626489215426975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,1024,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,16384,0.05515733361244202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,51200,0.18890399403042266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,512,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,16384,0.13494222693973118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,10240,0.02652799917591943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,6144,0.04098666707674662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,16384,0.08961688809924656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,3584,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,6144,0.42644532521565753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,2048,0.0749173363049825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,51200,0.1708293358484904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,768,0.020634666085243225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,768,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,1536,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,768,0.02591200007332696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,1024,0.006692444284756978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,768,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,6144,0.023935111032591924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,6144,0.004252444538805219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,12288,0.01552977826860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,3584,0.07127022080951266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,51200,0.02256177696916792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,128,0.005385777602593104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,12288,0.25581155882941353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,6144,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,128,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,3072,0.0533902214633094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,2048,0.030163556337356567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,51200,0.059030221568213574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,64,0.006108444597986009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,4096,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,7168,0.029863112502627905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,8192,0.08306933111614652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,3584,0.013386666774749756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,51200,0.07161333163579305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,256,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,8192,0.620739565955268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,4096,0.02644355595111847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,512,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,3584,0.03563200102912055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,10240,0.03197333216667175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,4096,0.008739555875460306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,512,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,4096,0.05345066719584995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,768,0.003925333420435588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,2560,0.024879111184014216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,6144,0.04905688762664795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,7168,0.013399110900031196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,5120,0.059734225273132324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,2048,0.006708444406588872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,4096,0.037131554550594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,8192,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,4096,0.00816000004609426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,1024,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,1536,0.020988444487253826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,3584,0.024375110864639282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,256,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,3584,0.00775022225247489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,51200,0.2850302325354682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,16384,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,5120,0.011762667033407422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,256,0.01716977854569753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,5120,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,16384,0.3349582354227702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,12288,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,2048,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,768,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,16384,0.1218328873316447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,6144,0.18665599822998047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,1536,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,2048,0.021750221649805706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,5120,0.02666933337847392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,10240,0.018048889107174344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,10240,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,1536,0.006106666806671355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,10240,0.07151466608047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,10240,0.020954666866196528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,2560,0.00960177762640847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,6144,0.027785778045654297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,8192,0.3685866726769342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,10240,0.09258221917682224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,2048,0.009522666533788046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,4096,0.023872888750500146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,7168,0.029709332519107397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,6144,0.6573893229166666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,32,0.15927733315361872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,65536,0.6260426839192709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,10240,0.08267555634180705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,10240,0.10256622234980266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,512,0.06083644760979546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,768,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,256,0.03762222329775492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,1024,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,2048,0.019867555962668527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,2048,0.006737777756320105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,8192,0.032440000110202365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,10240,0.025260445144441392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,7168,0.0778684417406718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,128,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,3072,0.022745778163274128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,6144,0.06663644313812256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,3584,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,32,0.010208889014191097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,6144,0.011594666375054253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,4096,0.010527999864684211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,3584,0.019705777366956074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,768,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,12288,0.05233777893914116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,768,0.018888889087571036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,7168,0.02572622232966953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,10240,0.057838221391042076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,2560,0.03573155403137207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,12288,4.1823272705078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,7168,0.018175111876593698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,7168,0.016773333152135212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,12288,0.01979555520746443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,8192,0.03074044320318434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,10240,0.06996444198820326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,16384,0.03188177943229675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,1024,0.01827022267712487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,768,0.07333244217766656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,6144,0.04908533228768242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,51200,2.837067498101128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,1024,0.06295288933648004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,768,0.007584888901975419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,2048,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,51200,0.054992887708875865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,6144,0.009745777481132084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,3584,0.01778133379088508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,32,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,4096,0.007345777418878343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,3072,0.010159999959998662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,2560,0.16258845064375135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,12288,0.02408711115519206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,2048,0.005482666608360078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,8192,0.13779289192623562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,8192,0.02717777755525377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,12288,0.29996177885267467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,1536,0.0035111109415690103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,51200,0.050091554721196495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,4096,0.01475911173555586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,64,0.0021084443562560612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,6144,0.026359111070632935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,5120,0.016578666037983365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,7168,0.05881066454781426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,6144,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,12288,0.017737777696715463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,3584,0.025058666865030926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,2048,0.02777333392037286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,128,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,1024,0.0028826666788922418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,3584,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,1536,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,512,0.00701155596309238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,51200,0.35650844044155544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,1536,0.0034328889515664843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,256,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,256,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,2560,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,32,0.0057813334796163775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,8192,0.22449866930643717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,1024,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,2560,0.03192977772818671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,4096,0.01611822181277805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,10240,0.00564533347884814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,2048,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,6144,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,10240,0.07879910866419475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,1536,0.03772977656788296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,65536,0.07007822063234118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,3584,0.14788088533613417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,51200,1.3394906785753038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,65536,0.1514186726676093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,1536,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,5120,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,6144,0.00871022211180793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,5120,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,2048,0.14245333936479357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,12288,0.014858666393491956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,51200,0.03805155555407206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,3072,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,1536,0.019031110737058852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,3072,0.07073777914047241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,65536,0.2664711210462782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,128,0.015297777122921415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,5120,0.013016888664828407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,768,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,7168,0.051073776351081006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,6144,0.03132889005872939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,128,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,7168,0.013415111435784234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,1024,0.03161511156294081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,7168,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,5120,0.010746666954623329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,10240,0.05067733261320326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,2560,0.007141333487298753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,32,0.009107555780145857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,5120,0.019401777121755812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,256,0.004226666771703296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,256,0.0032702222880389956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,7168,0.010420444111029306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,2560,0.3579973379770915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,16384,0.023376888699001733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,5120,0.01682400041156345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,5120,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,256,0.0022675556441148124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,3584,0.08210044436984591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,2560,0.045371555619769625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,1536,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,3584,0.21453777949015299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,1536,0.019739554988013375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,1536,0.019336000084877014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,5120,0.09045510821872288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,512,0.01897333396805657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,7168,0.021330666210916307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,3072,0.014410666293568082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,4096,0.025042666329277888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,2048,0.023589332898457844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,4096,0.02711822258101569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,51200,0.07102577553855048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,6144,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,3584,0.024344889654053584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,256,0.024105777343114216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,51200,0.049698667393790356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,128,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,3584,0.19076266553666857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,4096,0.023119111855824787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,2560,0.03551910983191596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,10240,9.94645521375868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,51200,0.15436000294155544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,5120,0.02162133322821723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,256,0.007712889048788283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,7168,0.03794666793611314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,6144,0.025661332739724055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,1024,0.035460445615980364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,7168,0.026517333255873785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,65536,0.06431021955278185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,512,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,32,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,128,0.04438044296370613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,51200,0.05113510953055488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,6144,0.045109334919187755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,7168,0.034640888373057045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,10240,0.04366222355100843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,4096,0.018183110488785636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,65536,0.062011553181542285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,2560,0.014914666612943014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,3072,0.01365244471364551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,10240,0.06764266888300578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,1536,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,16384,10.411070929633246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,7168,0.08284888664881389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,2560,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,768,0.030576888057920668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,16384,0.08898488680521648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,65536,2.4718880123562283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,3072,0.034035556846194796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,2048,0.1539191140068902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,2048,0.08221955431832208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,65536,0.058485335773891874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,512,0.008242666721343994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,4096,0.03974044322967529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,4096,0.15158400270673963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,2560,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,51200,0.05586489041646322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,256,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,2560,0.03291822142071194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,6144,0.08033244477377997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,12288,0.08471644586986965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,2048,0.007032000356250339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,51200,0.5408533414204916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,12288,0.018564444449212816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,1536,0.015535111228624979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,4096,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,3584,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,16384,0.014244443840450711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,2048,0.005668444352017508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,12288,0.041504889726638794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,4096,0.011109333071443768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,16384,0.18899467256334093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,32,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,65536,0.143068446053399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,10240,0.2326728767818875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,16384,0.027465777264700994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,12288,0.7544310887654623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,4096,0.021731555461883545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,256,0.013760888742076026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,7168,0.5937013096279568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,1024,0.02030933399995168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,2560,0.01945688823858897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,512,0.02310577697224087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,2048,0.06770310799280803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,2560,0.07625422212812635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,10240,0.03944355580541823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,64,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,10240,0.01924622224436866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,1024,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,5120,0.03433510992262099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,10240,0.02328799996111128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,3584,0.12655999925401476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,768,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,1024,0.0038782221575578055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,5120,0.067921777566274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,32,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,65536,1.4860106574164496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,3072,0.025450666745503742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,6144,0.028055999014112685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,8192,0.013603555659453074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,6144,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,51200,0.4669777552286784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,512,0.01912088857756721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,12288,0.0558604465590583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,7168,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,8192,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,4096,0.008941333327028487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,2560,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,256,0.008422222402360704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,2048,0.015328889091809591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,768,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,16384,0.2406702306535509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,512,0.17788622114393446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,3072,0.04123555620511373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,51200,0.04874666531880697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,512,0.023228444986873206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,5120,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,768,0.06077510780758328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,2048,0.07134666707780626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,5120,0.017723555366198223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,8192,0.03109333250257704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,12288,0.033928887711630926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,4096,0.03461155626508925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,1536,0.02959377898110284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,256,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,10240,0.030028445853127375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,128,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,51200,0.051035556528303355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,10240,14.453679402669271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,768,0.04795733425352308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,8192,0.01809244520134396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,10240,0.18104888333214653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,2560,0.01459377838505639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,10240,0.42381776703728574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,64,0.022277333670192297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,12288,0.47048356797960067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,16384,0.10243555572297837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,1536,0.02867199977238973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,51200,0.5742204454210069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,768,0.023303111394246418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,65536,0.09849244356155396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,8192,0.02472711106141408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,128,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,5120,0.03934311204486423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,5120,0.016394666499561734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,2048,0.022660444180170696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,1024,0.005971555494599872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,3072,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,32,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,16384,0.17016888989342582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,12288,0.03436622354719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,12288,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,3584,0.01053155544731352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,51200,0.16831555631425646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,2560,0.02420444455411699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,768,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,16384,0.02299555473857456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,2560,0.022126222650210064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,65536,0.24363822407192656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,2048,0.049011554982927114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,12288,0.026537777649031744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,2560,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,2560,0.03095199995570713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,8192,0.03369777732425266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,768,0.04138311081462436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,3584,0.051354666550954185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,7168,0.19540354940626356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,3072,0.017477333545684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,5120,0.028297778632905748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,51200,0.10079110993279351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,4096,0.007704888780911763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,3584,0.592007107204861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,12288,0.09943911102082993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,6144,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,16384,0.015815110670195687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,10240,0.03267466690805223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,1536,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,1024,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,7168,0.02221244408024682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,256,0.016885333591037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,4096,0.020786666207843356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,4096,0.007723555796676212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,1536,0.015555555621782938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,3072,0.016748444901572335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,51200,0.06929866472880046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,128,0.016521778371598985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,3584,0.5326648818122016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,3072,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,768,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,51200,0.06473066409428914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,7168,0.07248266537984212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,2560,0.042103111743927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,2560,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,4096,0.039431111680136785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,2560,0.031762666172451444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,2560,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,128,0.007704888780911763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,4096,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,2048,0.011240888800885944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,65536,9.480836656358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,2560,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,4096,0.03392533461252848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,2560,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,128,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,512,0.0032257777121331957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,256,0.0675271087222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,10240,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,2048,0.024919110867712233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,7168,0.18326044082641602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,768,0.019225777851210702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,2048,0.02624888883696662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,1536,0.03161866797341241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,10240,0.06441244151857164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,16384,0.6790569093492297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,8192,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,512,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,65536,0.10018755329979791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,16384,0.04940711127387153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,512,0.015308444698651632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,7168,0.019676445258988273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,1024,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,7168,0.03321244319279989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,10240,0.04596711198488871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,16384,0.01846222248342302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,16384,0.023974221613672044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,128,0.014839111102951897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,8192,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,7168,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,10240,2.0611485375298395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,8192,0.10321066776911418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,64,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,256,0.03979911075698005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,3584,0.011208888557222156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,8192,0.21869066026475695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,3584,0.016775111357371014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,2048,0.024125332633654278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,1024,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,3072,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,10240,0.0169840008020401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,2048,0.14266933335198298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,768,0.00869688888390859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,16384,0.05968266725540161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,1536,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,12288,0.03927910990185208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,128,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,8192,0.022436444958051045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,32,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,3584,0.015612444943851896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,3072,0.050832887490590416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,10240,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,16384,0.01794666714138455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,32,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,5120,0.21568089061313203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,10240,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,12288,0.038507554266187884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,8192,0.01601777805222405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,1024,0.0061048890153567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,5120,0.012884444660610624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,4096,1.2478097279866536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,10240,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,8192,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,3072,0.06988710827297635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,2048,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,51200,0.6356479856703017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,7168,0.05483022332191467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,8192,0.016286222471131217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,16384,0.06528355677922566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,7168,0.1366951068242391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,16384,0.058470222685072154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,7168,0.01758311192194621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,12288,0.8374640146891276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,2560,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,3584,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,256,0.01921866668595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,3584,0.03301066822475857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,4096,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,128,0.010604444477293225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,65536,0.5308213233947754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,5120,0.010777778095669217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,3584,0.008942222429646386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,3072,0.03159555461671617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,256,0.01740533279048072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,4096,0.047762665483686656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,4096,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,3072,0.06913155317306519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,256,0.016707556115256417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,768,0.016157334049542744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,1024,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,5120,0.08822577529483372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,65536,0.13232355647616917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,1024,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,1024,0.007891555627187094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,2560,0.03290133343802558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,6144,0.014025777578353882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,768,0.0029786665820413162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,3072,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,12288,0.01679022279050615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,2048,0.013893332746293811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,7168,0.35097066561381024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,3072,0.02604088849491543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,4096,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,10240,0.018166222506099276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,4096,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,12288,0.01605422298113505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,8192,0.14984177218543157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,10240,1.7431386311848958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,7168,0.19647022088368735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,1536,0.0051893335249688895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,65536,1.7261483934190538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,8192,0.07852533128526476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,7168,1.3476248847113714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,5120,0.18608533011542427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,1536,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,3072,0.0503111117415958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,32,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,3072,0.07633955611122979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,1536,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,1536,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,7168,0.023152889476882085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,51200,0.06304088566038343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,3072,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,65536,0.14184089501698813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,512,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,64,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,1024,0.0035075553589397003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,1536,0.03291911217901442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,4096,0.0949288871553209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,3584,0.07917866441938612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,16384,0.35128090116712785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,51200,0.14031110869513616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,8192,0.022467555271254644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,256,0.015294222368134392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,51200,0.1272791094250149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,1024,0.056757330894470215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,12288,0.0836942195892334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,6144,0.03456088900566101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,2048,0.009931555224789513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,512,0.019083556201722886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,1024,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,65536,0.053295112318462796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,10240,0.1240968836678399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,65536,1.4516578250461156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,32,0.010393777655230628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,2560,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,3072,0.22337155871921113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,512,0.0451200008392334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,4096,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,51200,0.016920889417330425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,1024,0.4324773417578803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,3072,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,51200,0.007804444266690149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,128,0.011142222417725457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,4096,0.050921777884165444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,128,0.012170666621790992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,16384,0.038316445218192205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,768,0.46116087171766496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,768,0.03759377863672044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,7168,0.06367022461361356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,2048,0.023395554886923894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,64,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,64,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,51200,0.19487467077043322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,256,0.13555644618140325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,1536,0.008955555657545725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,1536,0.22496978441874185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,6144,0.01350488927629259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,16384,0.5173813501993815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,128,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,6144,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,2560,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,1536,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,10240,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,64,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,1024,0.021623111433453027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,6144,0.03289422061708238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,2560,0.24641243616739908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,3072,0.013745778136783175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,512,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,12288,0.11922044224209255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,8192,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,2048,0.05063377817471822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,16384,0.06571377648247613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,256,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,5120,1.6738755967881944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,1536,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,10240,0.060793777306874595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,2560,0.03569777806599935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,128,0.0026311110705137253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,7168,0.03234666585922241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,3584,0.16216799947950575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,16384,0.019271110494931538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,1536,0.05782488981882731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,10240,0.060193777084350586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,16384,0.04054044352637397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,16384,0.03763555486996969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,2560,0.017338666650984023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,32,0.01275022245115704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,5120,0.038711110750834145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,4096,0.02028888960679372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,6144,0.025624889466497634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,6144,0.11237155066596137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,1536,0.016878222425778706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,512,0.012828444441159567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,3072,0.050945778687795006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,65536,0.15761244297027588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,512,0.014308444327778287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,8192,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,1536,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,10240,0.016548444827397663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,4096,0.038134220573637224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,3584,0.03909777932696872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,512,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,12288,0.014141332772043018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,512,0.008348444269763099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,3072,0.03165600034925673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,256,0.0024497778051429326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,10240,0.01181688904762268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,16384,1.3493688371446397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,10240,0.05415022373199463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,7168,0.021929777330822412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,1536,0.009139555195967356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,2560,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,5120,0.016344888342751395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,128,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,12288,0.023040889037979975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,7168,0.01901333365175459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,8192,1.7822969224717882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,1536,0.07196888658735487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,512,0.010101333260536194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,4096,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,10240,0.18124800258212617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,3072,0.00928177767329746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,12288,0.0233271105421914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,7168,0.039466665850745306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,3072,0.03415022293726603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,1536,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,6144,1.284089830186632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,12288,0.05229599939452278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,1024,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,3072,0.027442667219373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,4096,0.013510222236315409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,768,0.007156444092591603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,256,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,7168,0.029091556866963703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,3584,0.023596444063716467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,7168,0.4420008924272325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,512,0.07974400122960408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,2048,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,2048,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,256,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,128,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,768,0.020583111378881667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,65536,0.3173146777682834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,64,0.00518844442235099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,6144,0.031601776679356895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,4096,0.02071644365787506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,1024,0.09950488805770874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,2560,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,1536,0.06891110870573255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,6144,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,6144,0.04470488760206434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,12288,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,2048,0.025047999289300706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,1536,0.06104355388217502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,512,0.02570666703912947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,8192,0.08710755242241754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,4096,0.09815022018220688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,16384,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,3072,0.02882666720284356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,8192,0.07449511024687025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,3072,0.02774222195148468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,2560,0.0867991116311815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,2048,0.0064533332155810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,5120,0.020453333854675293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,3584,0.021957332889238994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,3072,0.5025377803378636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,8192,0.028769777880774602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,128,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,4096,0.036848889456854925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,6144,0.02214844524860382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,32,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,32,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,2048,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,768,0.08092266983456083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,2560,0.008737777670224508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,256,0.01533600025706821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,768,0.0351457761393653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,3584,0.378478209177653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,10240,1.0515955819023979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,768,0.02540622154871623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,16384,0.13748266961839464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,1536,0.026424888107511733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,768,0.017648888958825003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,12288,0.03302399979697333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,1024,0.10077866580751206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,32,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,768,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,3072,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,512,0.022533333963818018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,2048,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,7168,0.03431466552946303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,65536,0.08016355832417806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,5120,0.030943112240897283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,16384,0.060200889905293785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,3072,0.030167112747828167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,2048,0.022974222898483276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,6144,0.027930667002995808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,10240,0.49867290920681423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,4096,0.01199288914601008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,512,0.013842666314707862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,7168,0.8117084503173828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,512,0.02436977790461646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,256,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,3072,0.023006222314304773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,3072,0.016371554798550077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,256,0.0021973333011070886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,16384,0.40096354484558105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,256,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,256,0.028484443823496502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,3072,0.020618667205174763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,5120,0.027355555031034682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,3584,0.0101742222905159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,128,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,10240,0.030508445368872747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,7168,0.01828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,256,0.022050667140218947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,16384,0.03597866826587253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,1536,0.011781333221329583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,3072,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,65536,0.23211465941535103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,128,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,512,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,16384,0.01757688820362091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,2048,0.009832888841629028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,128,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,4096,0.01993600030740102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,51200,0.27764977349175346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,3584,0.009400000174840292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,3584,0.04499199986457825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,1024,0.020720889170964558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,3584,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,512,0.004254222330119875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,16384,0.02402488887310028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,2560,3.0988648732503257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,4096,0.018290667070282828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,2048,0.03492355677816603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,51200,0.14186755816141763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,4096,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,3584,0.024326221810446844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,256,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,2560,0.0595031115743849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,7168,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,12288,0.06934310992558797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,768,0.012528888881206512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,8192,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,1536,0.37670310338338214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,128,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,3072,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,8192,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,2048,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,64,0.03345600101682875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,4096,0.03000977635383606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,2048,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,3584,0.04193066557248434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,6144,0.032023111979166664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,12288,0.13839377297295466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,512,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,128,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,256,0.09214311175876194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,5120,0.019976889093716938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,10240,0.02885422110557556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,51200,1.0052826139662001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,4096,0.1648577716615465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,6144,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,12288,0.24757777320014107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,5120,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,1536,0.012676444318559436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,1024,0.12540622552235922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,1536,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,51200,0.9927839703030057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,32,0.028118222951889038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,768,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,3072,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,128,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,51200,1.1198861863878038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,3072,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,512,0.008574222524960836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,6144,0.017324444320466783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,64,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,4096,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,5120,0.054745776785744556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,128,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,51200,0.11406044165293376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,768,0.01536266671286689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,64,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,51200,0.3398133383856879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,12288,0.04361244373851352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,7168,0.044569777117835156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,256,0.032744000355402626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,6144,0.011996444728639392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,3072,0.029044446018007066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,16384,0.437084436416626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,512,0.01994311147265964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,5120,0.025528889563348558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,6144,0.05297777718967862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,1536,0.007753777835104201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,64,0.10347288846969604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,32,0.0829004446665446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,10240,0.02940622303220961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,12288,0.03424177898301019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,1024,0.025797333982255723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,16384,0.24919822480943468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,3584,0.0336079994837443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,12288,0.24140000343322754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,256,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,2560,0.01127466642194324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,128,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,1024,0.0035537779331207275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,256,0.007123555574152205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,2048,0.02713777787155575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,1536,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,2560,0.018760888112915885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,2560,0.016527111331621807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,65536,0.13001510832044813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,3584,0.019371555911170114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,2560,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,16384,0.11834043926662868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,8192,0.0809297760327657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,3584,0.10639555586708917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,64,0.00906133320596483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,16384,0.025416889124446448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,5120,0.04068533248371548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,10240,0.10842311382293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,5120,0.02370044423474206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,8192,0.5653022130330404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,1024,0.02973866793844435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,1536,0.06840533018112183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,65536,0.36944089995490176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,128,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,3584,0.01606755620903439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,2560,0.07678400145636664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,5120,0.017260443833139207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,8192,0.011324444578753578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,4096,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,512,0.07107110818227132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,32,0.009082666701740688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,10240,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,6144,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,32,0.0053048887186580235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,4096,0.45177422629462344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,2560,0.23064266310797799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,2560,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,4096,0.07104533248477511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,128,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,8192,0.013804444836245643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,32,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,5120,0.05440533492300245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,2560,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,12288,0.01725955473052131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,5120,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,3584,0.03540888759824965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,5120,0.010981332924630908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,3584,0.46861955854627824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,128,0.004181333300140169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,51200,0.17981155713399252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,65536,0.15758489237891302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,4096,0.9314720365736219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,32,0.0779271125793457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,12288,1.076097806294759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,65536,0.03805955582194858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,256,0.002975111206372579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,1024,0.25319377581278485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,4096,0.017658667431937326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,4096,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,2560,0.006992888947327931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,6144,0.03426222337616815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,256,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,6144,0.010416888528399997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,3072,0.07344888978534274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,6144,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,16384,0.02372444503837162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,65536,0.06565244330300225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,10240,0.033963554435306124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,32,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,10240,0.02014044423898061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,7168,0.03716177741686503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,1536,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,5120,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,256,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,32,0.0023662222342358693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,5120,0.01643111142847273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,3584,0.4341306686401367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,3584,0.6088142395019531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,65536,0.38941865497165257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,8192,1.0566950903998482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,1024,0.0346995559003618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,5120,0.023399111297395494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,6144,0.02752799954679277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,2560,0.014204444156752693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,3072,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,10240,0.022069333328141108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,7168,0.013638222383128272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,5120,0.012279111478063794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,512,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,3584,0.08970489104588826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,16384,1.6846675872802734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,4096,0.03808444407251146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,512,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,10240,0.02689244515366024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,5120,0.08691733413272434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,16384,0.1677937772538927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,10240,0.016728000508414376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,12288,0.020440888073709276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,1024,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,51200,0.12039555443657769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,8192,0.024328889118300542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,1536,0.026492445005310908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,3584,0.024354666471481323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,10240,0.036637332704332136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,51200,0.06900888681411743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,3072,0.898614247639974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,1024,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,256,0.015554666519165039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,128,0.01919911139541202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,7168,0.24339466624789766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,128,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,4096,0.052447110414505005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,12288,0.039834668238957725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,2048,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,10240,0.16196621788872614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,512,0.05407911207940844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,5120,0.020607110526826646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,1024,0.023102222217453852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,128,0.022690667046440974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,12288,0.03347999850908915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,128,0.015486222174432544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,10240,0.07836000124613444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,256,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,5120,0.06222666634453667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,64,0.07565066549513075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,1536,0.09488622347513835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,128,0.0058266665372583605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,65536,0.16046844588385686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,4096,0.05139999919467502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,256,0.09720977809694077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,256,0.005396444350481033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,3584,0.02233422299226125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,32,0.010091555615266165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,512,0.01680799987581041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,16384,2.0264515346950955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,8192,1.27202574412028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,128,0.010330666270520952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,10240,0.0181404451529185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,51200,0.2252497673034668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,1024,0.14845155345069036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,2048,0.03153422143724229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,6144,0.02474311159716712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,2560,0.24224532975090873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,65536,0.030681778987248737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,128,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,3584,0.0797297755877177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,16384,0.019768888751665752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,5120,0.020478222105238173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,6144,0.028140445550282795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,8192,0.027110222313139174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,6144,0.011869333684444427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,7168,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,12288,0.19043644269307455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,512,0.014610666367742749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,8192,0.03196444445186191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,8192,1.745340453253852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,12288,0.19518310493893096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,64,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,3584,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,12288,0.06933422221077813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,12288,0.041836443874571055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,256,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,1536,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,3072,0.06192355685763889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,3072,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,768,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,512,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,65536,0.07319911321004231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,768,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,8192,0.1709280014038086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,3584,0.05191911260286967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,51200,0.17861421902974448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,10240,0.0365555551317003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,2560,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,2048,0.042797333664364286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,6144,0.09568177991443211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,2560,0.021716444028748408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,7168,0.022687110635969374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,768,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,5120,0.009461333354314169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,3072,0.04426577687263489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,16384,0.32113244798448354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,51200,5.0313699510362415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,5120,0.05499200026194254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,128,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,8192,0.0196426659822464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,1536,0.015402666396564908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,51200,0.1882044474283854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,4096,0.4987795617845323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,256,0.020643555455737643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,2048,0.49475378460354275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,1024,0.03718400001525879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,3072,0.03671822283003066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,8192,0.022289777795473736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,768,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,7168,0.03943999939494663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,768,0.024421332610978022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,768,0.019223110543357003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,12288,0.1070551143752204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,6144,0.02216888964176178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,65536,0.26017332077026367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,8192,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,51200,0.044666667779286705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,2560,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,3584,0.04543200135231018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,32,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,1024,0.21258933014339873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,2560,0.015474667151769003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,7168,2.0548587375217013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,8192,0.02083555526203579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,65536,0.170415997505188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,32,0.003318222239613533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,2560,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,5120,0.049901333120134145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,128,0.00499199993080563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,1536,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,12288,0.09646933608584934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,3584,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,8192,0.05631555451287163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,256,0.006170666466156642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,51200,0.05230489042070177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,7168,0.07642133368386163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,5120,0.031236443254682753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,3072,0.024478221933046978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,8192,0.07600800196329753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,1024,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,1024,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,51200,0.41371467378404403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,1536,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,256,0.12944889068603516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,1024,0.021917333205540974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,3584,0.012824888858530255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,3072,0.007671111159854465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,2048,0.01789866718981001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,4096,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,6144,0.21071822113460967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,256,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,65536,0.06729511419932048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,768,0.014924445086055331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,6144,0.03525955478350321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,512,0.009149333669079674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,3072,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,6144,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,4096,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,7168,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,16384,0.2729964521196154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,7168,0.010915555887752108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,8192,0.026929777529504564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,4096,0.15676000383165148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,10240,0.01408622165520986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,6144,0.022264000442292955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,10240,0.011905777785513135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,7168,0.09803822305467393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,512,0.04691733254326714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,1024,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,3584,0.007520889242490132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,768,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,1024,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,1536,0.015442666080262927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,7168,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,512,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,4096,0.020035554965337116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,16384,1.127655135260688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,51200,2.7240657806396484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,512,0.023710222707854375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,12288,0.03051466743151347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,512,0.044529779089821704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,4096,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,768,0.03687466515435113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,7168,0.015239111251301236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,2560,0.032638221979141235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,5120,0.026385777526431616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,5120,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,8192,0.02251200046804216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,7168,0.13175288836161295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,2048,0.014027555783589682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,7168,0.04621866676542494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,10240,0.012116444607575735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,3584,0.007709333466158972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,1536,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,768,0.03037244412634108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,4096,0.026605332891146343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,7168,0.2594133218129476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,3584,0.041908442974090576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,128,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,51200,0.04957688848177592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,65536,0.08210844463772245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,10240,0.07725510994593303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,512,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,32,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,5120,0.026123555170165166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,8192,0.025437333517604407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,7168,0.011415999796655444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,3072,0.021013332737816706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,65536,0.06646222538418241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,1024,1.9019040001763239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,5120,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,1536,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,3072,0.023572444915771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,2560,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,4096,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,128,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,12288,0.039507554637061224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,10240,0.033434665865368314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,128,0.007768889268239339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,1024,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,5120,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,8192,0.011223999990357293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,12288,0.013247999880048962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,64,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,128,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,6144,0.030658665630552504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,7168,0.021381333470344543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,1536,0.012727999852763282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,1024,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,51200,0.5710577964782715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,768,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,128,0.01537333263291253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,2048,0.021525333325068157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,51200,2.570277320014106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,2560,0.14256621731652155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,2560,0.8930035697089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,4096,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,7168,0.06418133444256253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,4096,0.09332888656192356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,5120,0.13913244671291775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,3584,0.011678222152921887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,4096,0.012058667010731168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,1536,0.014441778262456259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,1536,0.02956533432006836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,8192,0.3840480115678575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,4096,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,5120,0.12475732962290446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,12288,0.04162133402294583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,1024,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,768,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,1024,0.04239822096294827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,12288,0.04279555545912849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,3072,0.03227555420663621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,2048,0.017616889543003507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,51200,0.5155493418375651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,2048,0.009294222626421187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,5120,0.0695528917842441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,10240,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,1536,0.012548444171746572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,4096,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,12288,0.06128088633219401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,512,0.00254488881263468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,128,0.01162577751610014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,1536,0.0617191129260593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,7168,0.017865777015686035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,10240,0.032836443848080106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,16384,0.21027910709381104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,4096,0.107479108704461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,2048,0.02089511189195845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,2048,0.013487111363146039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,32,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,1536,0.019896888070636325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,51200,0.07385155227449205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,64,0.01255555533700519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,3072,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,1024,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,3072,0.012153777811262341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,65536,3.1927405463324656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,3072,0.6672382354736328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,12288,0.29746845033433705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,12288,0.025830222500695124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,3584,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,2048,0.008461332983440822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,768,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,256,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,512,0.047232889466815524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,256,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,768,0.00300799993177255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,256,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,12288,0.042040000359217324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,256,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,1024,0.06320266591178046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,7168,0.028515554136700098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,5120,0.050276445017920605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,128,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,10240,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,51200,0.10053422053654988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,512,0.011431111229790581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,512,0.04692622356944614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,2048,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,65536,0.21219910515679252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,7168,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,2560,0.014553777045673795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,256,0.005771555420425203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,6144,0.05015288790067037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,768,0.006797333144479328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,4096,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,12288,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,12288,0.019865777757432725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,4096,0.017108443710539076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,3584,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,7168,0.01920000049802992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,12288,0.015640000502268474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,16384,0.04416444566514757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,1536,0.28760266304016113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,65536,0.2307137648264567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,12288,0.43982577323913574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,7168,0.011797332929240333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,1536,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,16384,0.024085332949956257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,8192,0.0107342220014996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,10240,0.0451564457681444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,1024,0.004407111141416761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,32,0.03251555562019348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,256,0.06491111384497748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,2560,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,1024,0.019343111250135634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,65536,0.4629333284166124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,12288,0.03622400098376804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,32,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,1536,0.004299555387761858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,6144,0.02300888962215847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,32,0.015849777393870883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,3072,0.0236435549126731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,128,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,4096,0.09958044687906902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,256,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,512,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,256,0.013095999757448832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,6144,0.03360266817940606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,3072,0.03212089008755154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,4096,0.15346399943033853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,65536,0.8240799903869629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,2560,0.02715644405947791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,2048,0.03152444296412998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,7168,0.2528897656334771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,8192,0.12186311350928412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,32,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,3584,0.009099555512269339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,7168,0.01309333327743742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,768,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,10240,0.05911822451485527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,128,0.027423111928833857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,1536,0.038866665628221296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,8192,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,6144,0.027798222170935735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,32,0.004666666603750653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,64,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,1024,0.06325778033998278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,6144,0.009861333502663506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,2560,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,8192,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,2048,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,6144,0.019550222489568923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,3072,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,32,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,3584,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,512,0.004654222064548069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,1536,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,1536,0.0057946667075157166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,10240,0.022670222653283015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,4096,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,64,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,5120,0.04578488734033373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,16384,2.2031983269585504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,768,0.06460977925194634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,8192,0.023338667220539514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,768,0.0041351111398802865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,2560,0.0790862242380778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,12288,0.015154666370815702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,7168,0.041788445578681103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,32,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,3072,0.06487733125686646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,7168,0.04554666744338142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,5120,0.10896266831292047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,2048,0.02053955528471205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,32,0.003576888806290097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,2048,0.023782221807373896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,7168,0.04114755656984117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,8192,0.0574586656358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,4096,0.007732444339328342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,64,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,4096,0.045351111226611666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,65536,0.16692711247338188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,4096,0.4234488805135091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,8192,0.024719999896155462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,65536,0.16587732897864446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,128,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,512,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,256,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,5120,0.08636533551745945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,51200,0.6291919814215766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,51200,0.11355111334058975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,7168,0.018043556147151522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,5120,0.3650835620032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,12288,0.05319822165701124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,8192,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,10240,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,7168,0.01961599952644772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,2560,0.036335110664367676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,2048,6.497647179497613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,32,0.010129777921570672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,3072,0.04460533459981283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,65536,0.46344354417588973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,128,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,12288,0.2408275604248047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,4096,0.01664355562792884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,2048,0.01292711082432005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,10240,0.016164445214801364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,6144,0.04293422235382927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,10240,0.044236444764667086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,1536,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,65536,0.06260177824232313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,16384,0.02017422186003791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,7168,0.03808533483081394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,16384,0.02382133404413859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,65536,0.059940444098578564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,12288,0.013471110827393003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,51200,0.06952355305353801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,7168,0.044397334257761635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,16384,0.18675733937157524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,8192,2.089005364312066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,2560,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,256,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,3072,0.10302578078375922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,8192,0.08703821897506714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,128,0.01457333399189843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,12288,0.15272888872358534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,6144,0.017659554878870647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,128,0.008365333080291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,8192,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,32,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,128,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,5120,0.16324355867173937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,1536,0.019761777586407132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,3584,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,16384,0.0436097780863444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,7168,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,256,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,1024,0.01972888906796773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,3584,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,64,0.0024560000747442245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,16384,0.044716444280412465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,5120,0.09379911422729492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,6144,0.004593777573770947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,2048,0.03517333335346646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,768,0.017663111289342243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,768,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,128,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,768,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,8192,0.04977066649330986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,1536,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,2560,0.027280888623661462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,1536,0.006350222147173352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,3584,0.008548444343937768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,256,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,6144,0.024082667297787134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,3072,0.14884177843729654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,768,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,2048,0.007496000164084964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,5120,0.15483200550079346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,8192,0.021648888786633808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,4096,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,256,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,2560,0.03289866778585646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,5120,0.3875102202097575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,3072,0.01440533333354526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,7168,0.07101511292987399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,1024,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,16384,0.28619199328952366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,2560,0.07935822010040283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,128,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,10240,0.5520097944471571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,5120,0.04808800088034736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,2560,0.01905955539809333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,7168,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,4096,0.009054222040706212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,1024,0.03275022241804335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,8192,0.014783999986118741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,8192,0.02695111102528042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,768,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,2560,0.007335111498832703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,5120,0.046556442975997925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,5120,0.024939555260870192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,512,0.012251555919647217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,2560,0.01920000049802992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,1536,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,10240,0.03931377662552728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,6144,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,12288,0.04711822337574429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,12288,0.026518222358491685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,3072,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,5120,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,16384,0.025799999634424847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,64,0.11489689350128174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,256,0.019088000059127808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,128,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,51200,0.04973511232270134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,5120,0.08751999669604832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,51200,0.0494026674164666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,2560,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,128,0.03493955400254991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,768,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,51200,0.029909332593282063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,6144,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,256,0.03887288769086202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,8192,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,512,0.05676088730494181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,2048,0.025069332785076563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,7168,0.021721776988771226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,128,0.010253333383136326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,16384,0.023930667175187006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,65536,7.760270012749566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,1024,0.003903999924659729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,1536,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,512,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,1536,0.008609777523411645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,12288,0.6553013589647081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,6144,0.014071111877759298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,7168,0.039691554175482854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,3072,0.024107555548350017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,4096,0.024680889315075342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,64,0.005356444252861871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,4096,0.3733946747250027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,7168,0.011333333121405708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,3584,0.08753510978486802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,6144,0.1115457746717665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,2048,0.033047109842300415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,128,0.02164711058139801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,1536,0.06395555867089166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,2560,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,128,0.019857777489556205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,128,0.053344001372655235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,7168,0.15041066540612116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,8192,0.5579333305358887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,8192,0.053893334335751004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,5120,0.00904977818330129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,512,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,128,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,65536,0.7567315631442599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,51200,0.0270115551021364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,3584,0.013878222141000958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,16384,0.0444106658299764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,3584,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,2048,0.126129772928026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,512,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,3072,0.01108977778090371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,5120,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,2048,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,51200,0.11393333805931939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,256,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,256,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,6144,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,128,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,768,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,2048,0.006802666518423293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,512,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,1536,0.02106222179200914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,2560,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,3584,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,6144,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,12288,2.218947516547309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,6144,0.8779013421800402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,1024,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,1536,0.03224444389343262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,3072,0.07797155777613322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,12288,0.2635057767232259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,1536,0.003960888832807541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,16384,0.026735999517970618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,65536,0.07578311363855998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,3584,0.09366399712032741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,8192,0.025751999682850305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,1536,0.3426524533165826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,128,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,65536,0.017765333255132038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,16384,0.07212711042828031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,2560,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,12288,0.812921788957384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,51200,0.06414044565624662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,256,0.02126933303144243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,768,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,6144,0.17567200130886504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,768,0.021067554752031963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,2048,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,7168,0.42630932066175675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,7168,0.697936905754937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,128,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,5120,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,16384,0.01814311080508762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,65536,0.07697511381573148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,128,0.0057911111248864075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,2560,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,12288,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,8192,0.021710221966107685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,3072,0.039820445908440485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,5120,0.03162755568822225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,1536,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,12288,0.07326666514078777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,3072,0.00962933318482505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,768,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,7168,0.10859644412994385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,2560,0.04603644543223911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,7168,0.05390666590796577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,3584,0.019680000013775296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,1024,0.01850044396188524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,128,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,32,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,2048,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,51200,0.07587644126680163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,256,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,8192,0.005362666729423735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,8192,1.2869137658013237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,4096,0.004259555704063839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,10240,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,3072,0.029271999994913738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,256,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,1024,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,8192,0.026937777797381084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,3072,0.03705155518319871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,51200,0.2889457808600532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,3584,0.0090524446633127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,64,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,3584,0.054056889481014676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,2560,0.013156444662147097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,32,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,12288,0.03654488921165466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,3072,0.037725332710478045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,256,0.01924888955222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,5120,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,6144,0.0717137787077162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,3584,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,51200,0.08644977543089126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,65536,0.398693323135376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,2048,0.012863111164834766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,12288,0.03769866625467936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,1536,0.029149333635965984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,2560,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,32,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,128,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,768,0.00675999994079272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,3072,0.013064000341627332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,4096,0.03410311208830939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,16384,0.1257128847969903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,16384,0.015797333584891427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,2048,0.09717066420449151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,3072,0.02623555560906728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,1536,0.03546133306291368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,2560,0.03630222214592828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,2560,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,256,0.019069333871205647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,1024,0.03407199846373664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,4096,0.3478693432278103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,1024,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,2048,0.030048890246285334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,3584,0.020003555549515616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,2048,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,64,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,2048,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,12288,0.05882933404710558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,1536,0.03841422332657708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,512,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,7168,0.012776888906955719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,128,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,3072,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,5120,0.010062221851613786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,16384,0.021225778592957392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,2048,0.018883556127548218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,32,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,3584,0.013844444519943662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,128,0.02755822241306305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,16384,2.5425521002875433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,4096,0.015060444672902426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,768,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,3584,0.013342222405804528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,4096,0.010452444354693094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,4096,0.018221333622932434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,5120,0.021571555071406897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,2048,0.01626755628320906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,512,0.010494222243626913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,3584,0.025072889195548162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,3584,0.0063786668082078295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,7168,0.06498044729232788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,3584,0.01349422252840466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,4096,0.14778399467468262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,10240,0.05299377772543165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,7168,0.03531288769510057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,51200,0.15484178066253662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,12288,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,3584,0.018808888064490426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,8192,0.05680977635913425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,7168,0.04331377810902066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,10240,0.6785742441813151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,16384,0.04675288995107015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,4096,0.0717173351181878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,2048,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,1536,0.05647644731733534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,768,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,3584,0.04223733478122287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,2048,0.01749955614407857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,64,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,3584,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,4096,0.010696888797812991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,3072,0.010397333237859938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,2560,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,256,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,7168,0.019367999500698514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,4096,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,1536,0.011640888949235281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,12288,0.04319733381271362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,8192,0.024100444383091394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,51200,0.0518871115313636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,512,0.01198755535814497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,2560,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,8192,0.023173333870040044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,7168,0.049343112442228526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,65536,0.24630488289727104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,3072,0.36610666910807294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,3584,0.007753777835104201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,2048,0.014236445228258768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,10240,0.04309689005215963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,1536,0.07844444115956624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,64,0.01309866706530253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,1024,0.012268444730175866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,2048,0.028416001134448584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,1024,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,3584,0.16378045082092285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,3584,0.8158515294392904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,3072,1.1364764107598198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,2048,0.047395557165145874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,5120,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,1536,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,3072,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,10240,0.02757955590883891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,51200,1.5007350709703233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,8192,0.8569386800130209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,32,0.008021333151393467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,512,0.01787911189926995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,16384,0.2026684416664971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,4096,0.031101332770453557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,3072,0.011178666518794166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,10240,0.07175288597742717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,10240,0.04722311099370321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,2560,0.02983733349376255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,1536,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,5120,0.04383111000061035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,512,0.02936799989806281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,8192,0.1768266624874539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,5120,1.3995333777533636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,32,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,8192,0.17775732941097686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,65536,0.47107553482055664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,6144,0.21532532903883192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,768,0.005069333232111401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,10240,0.6211093266805013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,64,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,10240,0.01611911091539595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,51200,0.0530008872350057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,512,0.011759999725553723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,5120,0.02204088866710663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,12288,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,65536,6.485223134358724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,4096,0.03941155473391215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,65536,0.08182133568657769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,1536,0.007833778030342525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,4096,0.009176889227496253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,768,0.030280888080596924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,65536,0.11939999792310928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,3072,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,10240,0.031493332650926374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,512,0.022290666898091633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,64,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,7168,0.036045332749684654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,1536,0.02443555494149526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,32,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,3584,0.13110400570763484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,4096,0.5390426847669814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,2560,0.020956445071432326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,1536,0.05515822105937534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,8192,0.038783109850353666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,2560,0.02741333345572154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,16384,0.03752177622583177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,1536,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,3072,0.013348444468445249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,64,0.002198222196764416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,1024,0.01924533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,3584,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,768,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,4096,0.06302311023076375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,51200,4.866357167561849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,12288,0.04332355658213297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,51200,1.358775986565484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,8192,0.057320886188083224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,12288,0.004943999979231092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,768,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,3584,0.04611022273699442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,1024,0.0035048888789282907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,16384,0.09637422031826443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,7168,0.04989599850442675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,51200,0.07036533620622423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,3072,0.011951111257076263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,256,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,1024,0.04186577929390801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,768,0.021887999441888597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,2048,0.006809777683681912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,768,0.023388443721665278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,256,0.005412444472312927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,2560,0.04500888784726461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,12288,0.012234666281276278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,4096,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,12288,0.039641777674357094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,4096,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,12288,0.02139555580086178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,51200,0.08256889051861234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,65536,0.9443955951266818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,7168,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,16384,0.09339555766847397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,1536,0.0038693332009845306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,6144,0.1149937841627333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,128,0.005382222019963794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,512,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,5120,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,2560,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,2048,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,2560,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,768,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,7168,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,16384,0.047953777843051486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,3584,0.4425484339396159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,8192,0.1209075583351983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,6144,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,8192,0.06698489189147949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,51200,0.07800889015197754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,1024,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,128,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,128,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,10240,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,1024,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,1536,0.4526693556043837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,256,0.009768889182143742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,6144,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,1536,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,65536,0.2176462279425727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,8192,0.023197333017985027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,3072,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,512,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,3584,0.03355200091997782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,512,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,256,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,10240,0.013057777451144325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,128,0.007440889047251807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,3072,0.009752000371615091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,16384,0.029148442877663508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,2048,0.331608878241645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,3072,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,512,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,8192,0.022651554809676275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,10240,2.6221591101752386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,256,0.026721777187453374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,128,0.006875555548402999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,512,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,1536,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,128,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,2048,0.049636443456014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,3584,0.026416889495319788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,2048,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,256,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,16384,0.08830044666926067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,3072,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,7168,0.041464000940322876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,65536,0.08675999773873223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,1024,0.08637866708967422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,2560,0.013062222136391534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,16384,0.07813422547446357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,2560,0.038690666357676186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,51200,0.06730755832460192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,3072,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,7168,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,12288,0.029134223858515423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,10240,0.03103111187616984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,768,0.10149155722724067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,768,0.016709332664807636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,1024,0.024830222129821777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,2560,0.048933333820766874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,7168,0.016858667135238647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,16384,0.08261333571539985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,2560,0.04567199945449829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,1536,0.04879200127389696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,51200,0.8416604465908475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,2048,0.25415288077460396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,1536,0.010968888799349466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,768,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,6144,0.11196711328294541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,8192,0.016297777493794758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,65536,0.14614221784803602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,3072,0.0236817780468199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,3072,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,1536,0.010950222611427307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,1536,0.06341511011123657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,10240,0.17979910638597277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,4096,0.020455999506844413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,1024,0.018004443910386827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,65536,0.20748444398244223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,8192,0.019302222463819716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,7168,0.08039022154278226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,51200,0.3427591058943007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,128,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,512,0.015211555692884656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,51200,0.4339582125345866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,1536,0.01556622154182858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,1024,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,7168,0.13413955105675593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,7168,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,768,0.08485511276457046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,512,0.02200711104604933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,512,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,6144,0.03353955679469638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,8192,0.02535466684235467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,16384,0.017878222796652053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,65536,0.4312933286031087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,51200,0.05951999955707126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,3584,0.021494223011864558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,51200,1.7441626654730902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,768,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,256,0.003181333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,768,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,3584,0.036421332094404436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,65536,0.5499999788072374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,16384,0.02602488961484697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,16384,0.042691555288102895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,1024,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,5120,0.018589332699775696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,32,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,5120,0.12990310457017687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,768,0.03616355525122748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,51200,0.4021751085917155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,51200,0.0835040012995402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,256,0.015927111109097798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,51200,3.7853058709038625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,3584,0.4070133368174235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,65536,2.593576855129666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,7168,0.012551111479600271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,65536,0.17595911026000977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,64,0.025133333272404138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,512,0.028774221738179524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,65536,0.0536480016178555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,768,0.011295999917719098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,3584,0.01833511061138577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,256,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,3072,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,65536,0.08133777644899157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,4096,0.027165333429972332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,32,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,10240,0.011936888926559024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,16384,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,4096,0.07113866673575507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,5120,0.04012533360057407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,128,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,6144,0.1275955571068658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,768,0.013841778039932251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,6144,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,51200,0.05344000127580431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,51200,0.2829759915669759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,65536,0.13871288299560547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,128,0.023001778456899855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,51200,0.05820088916354709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,2048,0.01588000026014116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,2048,0.023786667320463393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,16384,0.07212177912394206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,6144,0.013336000343163809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,256,0.0028968888024489083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,7168,0.018043556147151522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,16384,2.592573377821181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,65536,0.08527555730607773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,128,0.028813332319259644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,51200,0.1015591091579861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,12288,0.01346577786737018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,7168,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,3584,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,3584,0.00981422182586458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,256,0.010767999622556897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,1024,0.016858667135238647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,256,0.01386844449573093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,256,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,256,0.010637332995732626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,8192,0.010781333678298526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,3584,0.01202222208182017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,2560,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,7168,0.22048266728719076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,10240,0.012895999683274163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,1536,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,768,0.0041582224269707995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,2048,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,12288,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,6144,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,64,0.03237777617242601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,3584,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,4096,0.107515553633372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,8192,0.028181334336598713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,512,0.015306666493415833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,12288,0.21853066815270317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,2560,0.004963555683692296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,768,0.0203146669599745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,8192,0.03356711069742838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,5120,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,1536,0.04901777704556783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,8192,0.032186667124430336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,1536,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,1024,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,32,0.005468444277842839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,3072,0.004249777644872665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,5120,0.00871377769443724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,51200,0.7698382271660699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,65536,0.06393688917160034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,3584,0.009391999906963771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,1024,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,1536,0.020263999700546265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,5120,0.010789333118332757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,2560,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,12288,0.023239110906918842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,512,0.7712595727708605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,6144,0.011704000333944956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,256,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,3072,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,64,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,3584,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,51200,0.05897866355048286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,7168,0.06912177801132202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,1536,0.020364445116784837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,1024,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,6144,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,65536,0.07236355543136597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,1024,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,10240,0.08926933341556126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,5120,0.47194576263427734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,2048,0.020239111449983384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,3584,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,8192,0.03991288940111796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,512,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,16384,0.1789217789967855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,768,0.0036524443162812125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,2560,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,12288,0.18215289380815294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,128,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,32,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,51200,0.07383377684487237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,64,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,768,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,6144,0.03125955661137899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,10240,0.028598222467634413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,65536,0.3018248875935872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,6144,0.02863822380701701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,8192,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,51200,0.12394489182366265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,16384,0.7910657988654243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,65536,0.084297776222229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,4096,0.018904000520706177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,7168,0.08425244357850815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,512,0.04163377814822727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,512,0.055854221185048424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,1024,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,128,0.018616888258192275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,4096,0.018560000591807894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,1536,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,3584,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,6144,0.09894311428070068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,51200,0.0539537767569224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,7168,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,3584,0.020586666133668687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,3072,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,64,0.0018826667219400406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,128,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,16384,0.02202933364444309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,3072,0.008157333566082848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,10240,0.1158906618754069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,2560,0.05981689029269748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,8192,0.0343155562877655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,5120,0.05015110969543457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,1536,0.005374222166008419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,51200,0.2886008951399061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,4096,0.029459555943806965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,3584,0.011439999772442712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,32,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,3584,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,2560,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,16384,0.3644844426049127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,1536,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,10240,0.21953954961564806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,12288,0.40685333145989316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,1536,0.013023999830087027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,65536,0.12594844235314265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,4096,0.007824000385072496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,512,0.16352355480194092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,3072,0.019080888893869188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,3072,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,10240,0.03195910983615451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,3072,0.016363556186358135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,768,0.01386844449573093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,65536,1.1054400338066948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,3072,0.02401155564520094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,1536,0.0034862222770849862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,1024,0.0032195556494924757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,1536,0.015511110424995422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,6144,0.027816888358857896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,16384,0.25155289967854816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,2560,0.0946719977590773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,256,0.012373333175977072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,256,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,2560,0.004879111217127906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,4096,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,65536,0.21227910783555773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,6144,0.014533332652515836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,3072,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,6144,0.04092622134420607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,512,0.015632000234391954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,6144,0.04762844575775994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,51200,0.040214220682779946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,3072,0.0073928890956772705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,5120,0.024728889266649883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,16384,0.057470222314198814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,4096,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,10240,0.05393333236376444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,512,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,1536,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,512,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,1536,0.039094222916497126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,32,0.022389334109094407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,1536,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,3584,0.00816533300611708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,12288,0.025379555092917547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,512,0.015784889459609985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,8192,0.0038444445365005066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,768,0.18004888958401152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,64,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,128,0.010660444696744284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,768,0.01421688828203413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,4096,0.02202577723397149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,2560,0.388758209016588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,8192,0.019175110591782462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,256,0.005535999933878581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,3584,0.05276266733805338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,8192,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,51200,0.06191644403669569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,7168,0.03336355421278212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,2048,0.03685777717166477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,3072,0.015063110325071545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,256,0.011247110863526663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,512,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,6144,0.05692444245020548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,3072,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,16384,0.35193244616190594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,3072,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,12288,0.23291378551059297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,3072,0.01961511042382982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,8192,0.07033244768778484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,12288,0.06765777534908719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,2560,0.0277493331167433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,2560,0.0646595557530721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,128,0.018216000662909616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,7168,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,3072,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,2048,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,2560,0.049810667832692467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,512,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,128,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,51200,0.10444533162646824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,2560,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,3072,0.024461333950360615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,12288,0.025049777494536504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,64,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,3072,0.03629600008328756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,2048,0.08083733585145739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,1024,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,4096,0.15309955014122859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,8192,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,32,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,10240,0.02637333340115017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,1536,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,512,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,8192,0.017156443662113614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,256,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,128,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,16384,0.17282133632236055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,51200,0.06196622053782145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,256,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,64,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,768,0.035466667678621076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,4096,0.04112799962361654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,12288,0.013401777380042605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,7168,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,65536,0.02529688841766781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,1024,0.024989333417680528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,3072,0.07481155792872111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,768,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,7168,0.04006133476893107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,512,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,51200,0.0432204438580407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,256,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,1536,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,1536,0.0035253332720862497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,6144,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,1024,0.022434666752815247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,3072,0.023895111348893907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,256,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,128,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,6144,0.01092888911565145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,10240,0.025046222739749487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,2048,0.0196115556690428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,65536,0.05315377646022373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,16384,0.1400124496883816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,2048,0.011481777661376528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,8192,0.014849777022997538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,3584,0.5333377520243326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,3072,0.023709333605236475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,65536,3.1206372578938804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,6144,0.0761022236612108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,1024,0.008266666697131263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,8192,0.4109644360012478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,2560,0.008746667040718926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,12288,0.23866134219699434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,8192,0.01277511070171992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,512,0.010993777877754636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,3584,0.010090666512648264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,65536,0.07871733109156291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,512,0.0251217782497406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,65536,0.06259022156397502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,128,0.0487333337465922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,3584,0.013479111095269522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,3072,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,5120,0.12459111213684082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,4096,0.009725333087974125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,4096,0.015856888559129503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,10240,0.022227555513381958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,6144,0.03415200114250183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,128,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,2560,0.01925244430700938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,3072,0.05704177750481499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,12288,0.019912888606389362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,1024,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,10240,0.04455111092991299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,10240,0.019696000549528334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,3584,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,1536,0.004232888834344016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,6144,0.04000000158945719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,256,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,8192,0.022869333624839783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,16384,0.07639022005928887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,51200,1.213529798719618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,4096,0.009852444132169088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,1024,0.008398222426573435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,65536,0.07461155785454644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,4096,0.030547555949952867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,6144,0.0355679988861084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,128,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,5120,0.009621333744790819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,3072,0.05698133177227444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,65536,0.44777599970499676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,256,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,8192,0.17827199565039742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,6144,0.031098667118284438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,1536,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,3072,0.02011999984582265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,512,0.027518222729365032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,4096,0.021613332960340712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,65536,0.09404266542858547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,12288,0.052110221650865346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,5120,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,2048,0.12932533688015407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,512,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,65536,0.33386845058865017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,12288,0.01406044430202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,256,0.016570667425791424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,8192,0.02271288964483473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,4096,0.4436444441477458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,768,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,51200,0.11657155884636773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,128,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,2048,0.06492977672153048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,4096,0.021715554926130507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,2560,0.07671199904547797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,5120,0.07500622007581922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,12288,0.10282755560345119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,16384,0.020248888267411124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,64,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,256,0.005731555736727185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,64,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,12288,0.034167998366885714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,8192,0.17211822668711343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,16384,0.20937689145406088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,3584,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,2560,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,768,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,51200,0.04998133248753018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,512,0.022221333450741235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,65536,1.333712895711263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,3072,0.12776177459292942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,3072,0.012673777838548025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,10240,0.018445332845052082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,256,0.015099555253982544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,65536,0.11424356036716038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,10240,0.101383109887441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,10240,0.03345866666899787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,7168,0.036107556687461004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,16384,0.10868889093399048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,10240,0.15105066034528944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,256,0.06428088744481404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,512,0.01237511138121287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,2048,0.026310222016440496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,6144,0.1993324491712782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,51200,2.4837413363986545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,2560,0.783240000406901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,6144,0.032279110617107816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,32,0.18369600507948133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,2048,0.010991111397743225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,5120,0.01994044416480594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,10240,0.04867644442452324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,3072,0.18354577488369414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,8192,0.19290222062004936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,1536,0.02251111136542426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,3584,0.008092444803979661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,2048,0.014543111125628153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,768,0.0158906661801868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,128,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,768,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,65536,0.06260533465279473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,1536,0.014167111780908374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,512,0.022294221652878657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,32,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,10240,0.02645955483118693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,16384,0.044439110491010875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,1536,0.0047306666771570844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,65536,0.19664622677697075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,4096,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,256,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,51200,0.015973332855436537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,1024,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,3584,0.08121511009004381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,1024,0.022447110878096685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,1024,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,3072,0.03221599923239814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,6144,0.010800888968838586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,16384,0.386919101079305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,1536,0.06326399909125434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,8192,0.6076693534851074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,128,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,8192,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,10240,0.07109689050250582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,12288,0.0337137778600057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,6144,0.23060356246100533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,128,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,10240,0.1074755522939894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,768,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,4096,0.040629333919949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,64,0.0020844443804687923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,12288,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,64,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,7168,0.026597332623269822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,16384,0.2755040062798394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,3584,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,768,0.01678400072786543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,2048,0.01924088928434584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,3072,0.026168000366952684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,256,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,3072,0.2657039960225423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,3072,0.013680000272062091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,768,0.03663555449909634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,768,0.013748444616794586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,4096,0.00924622184700436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,512,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,4096,0.025672889418072168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,256,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,2048,0.012125333150227865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,12288,0.03644888930850559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,2560,0.4658355712890625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,1536,0.004865777575307422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,51200,0.5960124333699545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,2048,0.027931554449929133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,512,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,12288,0.018051554759343464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,32,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,512,0.0027928888383838865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,3072,0.02420799930890401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,5120,0.010440000229411654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,256,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,65536,0.06330488787757026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,51200,0.07034666670693292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,5120,0.022563555174403723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,10240,0.10781511333253647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,8192,0.054451555013656616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,6144,0.1310080024931166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,4096,0.00960799968904919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,64,0.005413333161009683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,512,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,1024,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,256,0.015743111570676167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,2560,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,12288,0.06669066349665324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,4096,0.03569777806599935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,10240,0.057519998815324574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,10240,0.06470844480726454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,51200,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,128,0.020589333441522386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,8192,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,1024,0.13789955774943033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,2048,0.010031110710567897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,12288,0.0244186669588089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,4096,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,65536,0.3762328889634874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,12288,0.046747555335362755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,65536,0.15013866954379612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,1536,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,512,0.006823110911581252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,6144,0.014195554786258273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,8192,1.177793820699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,4096,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,2048,0.008254222571849823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,512,0.007229333122571309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,2048,0.05684266487757365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,1024,0.039601776334974505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,2048,0.012321777641773224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,51200,0.05819288889567057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,7168,0.010406222608354356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,65536,10.102559407552084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,3584,0.01738133364253574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,768,0.021643555826610986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,8192,0.04376711116896736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,7168,0.07808355490366618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,64,0.007738666401969061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,12288,0.04639110962549845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,1536,0.026385777526431616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,3584,0.039192888471815325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,1024,0.07495111227035522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,3584,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,768,0.017280000779363845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,51200,0.14012444019317627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,8192,0.023071999351183575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,1024,0.0032133333798911837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,3072,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,512,0.09274755583869086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,51200,0.4444755448235406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,512,0.02327377763059404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,2560,0.04586311181386312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,65536,0.07478844457202487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,64,0.0042088888585567474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,5120,0.1134106715520223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,3584,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,51200,0.12317600515153672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,12288,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,1024,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,3072,0.030863109562132094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,768,0.06475733386145698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,64,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,512,0.0025457777082920074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,3072,0.0225582222143809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,7168,0.03468266791767544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,51200,0.2387608951992459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,128,0.02254844374126858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,1024,0.010452444354693094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,64,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,5120,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,51200,0.2720968988206652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,3072,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,128,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,8192,0.15225688616434732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,6144,0.09244266483518813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,4096,0.015230221880806817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,64,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,10240,0.04594844579696655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,1536,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,1024,0.010227555202113258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,768,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,7168,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,2560,0.045689778195487134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,6144,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,3072,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,6144,0.02424888809521993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,1024,0.021018667353524104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,32,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,10240,0.29107022285461426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,3584,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,7168,0.01276533305644989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,6144,0.06563555532031588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,128,0.02679555614789327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,1536,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,3072,0.05259911219278971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,768,0.01365066650840971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,64,0.008564444051848518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,512,0.012849777936935425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,1536,0.008781332936551835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,2560,0.02234755622016059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,1024,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,1536,0.06040177742640177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,256,0.005405333307054307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,10240,0.08790755271911621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,4096,0.08003466659122042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,4096,0.9680328369140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,6144,0.010272000398900773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,16384,0.0381022228135003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,7168,0.023966223001480103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,8192,0.03025955624050564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,2560,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,3584,0.026721777187453374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,64,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,6144,0.21774044301774767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,4096,0.008269333177142674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,7168,0.06012355619006687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,512,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,65536,0.06753155257966784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,2560,0.062352001667022705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,5120,0.023168888356950548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,10240,0.03177333209249709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,3072,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,3072,0.020113777783181932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,2560,0.019708444674809773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,256,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,256,0.005068444543414646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,3072,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,128,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,1024,0.12617244985368517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,3584,0.028240889310836792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,2048,0.01924800044960446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,512,0.0058062221441004015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,1536,0.010415111151006486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,7168,0.06309866905212402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,8192,0.01959911154376136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,3072,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,768,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,1024,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,10240,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,2560,0.015360888507631091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,12288,0.16878133349948457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,51200,0.059727112452189125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,10240,0.1888728936513265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,64,0.01315822203954061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,6144,0.04862311151292589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,4096,0.08642133076985677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,512,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,512,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,8192,0.031046221653620403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,3072,0.012635555532243518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,4096,0.08761244350009495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,7168,0.06260622209972806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,3584,0.005650666852792104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,128,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,3072,0.013783999615245394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,512,0.011062222222487131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,6144,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,2560,0.015154666370815702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,6144,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,5120,0.14982755978902182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,2048,0.13783110512627497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,128,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,8192,0.020767110917303298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,1536,0.04653244548373752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,12288,0.013411555853154926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,768,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,10240,0.013854222165213691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,8192,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,51200,0.048707554737726845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,4096,0.026320889592170715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,3072,0.022966222630606756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,128,0.006351999938488007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,64,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,1536,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,65536,0.457840018802219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,2560,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,4096,0.024009777439965144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,65536,0.6535893546210395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,6144,0.012533333566453723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,7168,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,4096,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,512,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,5120,0.058697779973347984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,51200,0.024884444144037034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,64,0.0035573335157500375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,1536,0.017539555827776592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,4096,0.009777777724795872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,5120,0.011393778026103973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,3584,0.020340444313155282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,256,0.002198222196764416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,1536,0.022791999909612868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,3072,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,256,0.006656000183688269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,5120,0.011686222420798408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,10240,0.007367110914654202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,16384,0.043624000416861645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,8192,0.0243520008193122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,6144,0.031108445591396753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,6144,0.12416799863179524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,12288,0.2785671022203234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,5120,0.16803199715084502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,2560,0.008799110849698385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,8192,0.3024675581190321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,10240,0.007147555549939473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,1536,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,768,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,512,0.010664000279373594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,10240,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,3584,0.017044444878896076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,1536,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,7168,0.02165599995189243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,8192,0.02055911057525211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,32,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,51200,0.7329653104146322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,3584,0.07988444301817152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,10240,0.01800355480776893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,5120,0.028957333829667833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,51200,0.1396204498079088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,5120,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,51200,0.107132441467709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,16384,0.034627556800842285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,6144,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,51200,0.05103111267089844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,2560,0.01183644433816274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,8192,0.0369742214679718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,128,0.021223111285103693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,10240,0.1297359996371799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,5120,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,5120,0.00868888861603207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,16384,0.054921779367658824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,51200,0.05788444148169624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,2560,0.013616888887352414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,1536,0.04814044303364224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,64,0.032111111614439226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,8192,0.059557331932915576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,2048,0.024311110377311707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,6144,0.1760951148139106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,3584,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,7168,0.01774222155412038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,16384,0.019486222002241347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,3072,0.024012444747818842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,128,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,8192,0.0705839991569519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,1536,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,5120,0.08967022101084392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,1024,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,51200,0.3079555564456516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,512,0.014236445228258768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,768,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,65536,1.2513440450032551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,256,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,32,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,10240,0.06835644774966769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,65536,0.5840693579779731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,3072,0.02442399991883172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,64,0.007744889292452071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,64,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,6144,0.13270666864183214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,65536,1.6114426718817816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,12288,0.018837332725524902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,32,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,512,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,1536,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,7168,0.011828444070286222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,8192,0.02713511056370205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,2560,0.004929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,65536,3.1758560604519315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,10240,0.07800355884763929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,1536,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,512,0.013173333472675748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,16384,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,16384,0.022679999470710754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,7168,0.027753776974148218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,3072,0.04016888803905911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,8192,0.014515555567211576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,3584,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,768,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,12288,0.2624933454725477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,1024,0.02514933380815718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,2560,0.009720000127951304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,256,0.022264000442292955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,16384,0.0503039989206526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,6144,0.01960088809331258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,5120,0.09620977772606744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,1536,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,6144,0.049878223074807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,8192,0.21687377823723686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,128,0.014495111174053617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,8192,0.07164177629682753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,6144,0.19551644060346815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,2560,0.01221777747074763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,32,0.00332266671790017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,7168,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,65536,0.06674577792485555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,5120,0.03728444377581278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,6144,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,5120,0.019647111495335896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,65536,0.22083289093441436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,2048,0.016040000650617812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,3072,0.04473511046833462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,16384,0.5064666536119249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,128,0.023532445232073467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,1024,0.0029315555261241067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,3584,0.007812444534566667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,5120,0.07760622104008992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,6144,0.4678933355543349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,512,0.008034666379292807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,16384,0.02879377868440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,4096,0.013021333350075616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,65536,1.1849102444118924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,16384,0.7473155657450358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,6144,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,2560,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,256,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,512,0.0027102221631341507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,7168,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,8192,0.07301421960194905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,2048,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,12288,0.013336000343163809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,3584,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,3584,0.7048462231953939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,512,0.014767999450365702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,2048,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,65536,4.516741434733073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,2560,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,12288,0.5625137752956814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,1024,0.03837777839766608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,4096,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,10240,0.2206773360570272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,768,0.04010488920741611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,51200,2.719840791490343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,4096,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,65536,0.13930755191379124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,512,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,64,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,64,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,3072,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,3584,0.037054220835367836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,8192,0.0405982236067454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,128,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,10240,0.058881779511769615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,2048,0.01070577816830741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,5120,0.015649777319696214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,2048,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,12288,0.049695998430252075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,4096,0.011925333076053195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,4096,0.03860355416933695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,3584,0.022008889251285132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,768,0.013030221892727746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,2560,0.030656887425316706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,1024,0.0222746663623386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,1024,0.023242667317390442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,256,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,512,0.03957688808441162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,1536,0.008563555777072906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,12288,0.09522755278481378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,8192,0.01623644431432088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,768,0.01941422290272183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,7168,0.005007110950019625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,1536,0.012799999780125089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,8192,0.037195556693606906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,7168,0.019527999891175162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,768,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,1536,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,2048,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,2048,0.008432000047630733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,51200,2.2044383154975042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,256,0.003810666501522064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,5120,0.04864622155825297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,256,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,256,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,128,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,16384,0.761623117658827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,1024,0.008590222232871586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,2048,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,1024,0.013162666724787818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,65536,0.18766044245825875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,256,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,256,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,3072,0.022075555390781824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,128,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,51200,0.04990577697753906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,4096,0.012818666795889536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,3072,0.020596444606781006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,512,0.01867466668287913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,65536,0.06301866637335883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,12288,0.02879377868440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,2560,0.1518124474419488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,2560,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,5120,0.016903999778959487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,6144,0.03574488891495599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,4096,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,64,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,128,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,3584,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,1024,0.017330666383107502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,10240,0.03427199853791131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,4096,0.028042667441897925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,7168,0.5177493095397949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,51200,0.5621182123819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,512,0.20831733279758027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,5120,0.021573333276642695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,51200,0.04310844341913859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,128,0.016578666037983365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,7168,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,768,0.14902044667137995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,1024,0.24217245313856336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,32,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,3072,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,768,0.020566221740510728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,1024,0.014115555418862237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,256,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,65536,0.8680177794562446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,768,0.03940800163480971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,512,0.013464888764752282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,51200,0.7442720201280383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,1024,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,8192,0.05492977632416619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,3584,0.3370782269371881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,64,0.014612444572978549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,10240,0.35412179099188906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,256,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,512,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,5120,0.009129777550697327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,1024,0.24849867820739746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,512,0.012834666503800286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,512,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,3072,0.044750223557154335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,16384,0.38716710938347715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,12288,0.04851999878883362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,1024,0.01686044368478987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,4096,0.029270221789677937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,128,0.03667733404371474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,65536,0.16586400402916804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,1024,0.0063662222690052455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,768,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,16384,0.06793066528108385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,7168,0.04654933346642388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,1024,0.01975555552376641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,512,0.02574399941497379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,128,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,7168,0.12981066438886854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,16384,7.399481879340278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,12288,0.06050666835572985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,1024,0.020568889048364427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,6144,0.062032891644371875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,3072,0.019229332605997723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,8192,0.015288889408111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,65536,1.8823617299397786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,7168,0.013050666285885705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,2048,0.07109422153896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,7168,0.06867111391491361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,2048,0.004811555561092165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,1024,0.10167733165952896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,8192,0.05756800042258369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,128,0.013068444199032254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,5120,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,10240,0.050535112619400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,10240,0.05776533153322008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,16384,0.4038568867577447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,16384,0.06370044416851468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,64,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,8192,0.06901333067152235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,7168,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,3072,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,8192,0.04669688807593452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,768,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,7168,0.09782133499781291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,6144,0.26498577329847545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,512,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,16384,0.5688586764865452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,7168,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,3072,0.011685333318180509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,5120,0.016213332613309223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,1024,0.007839110990365347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,1024,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,768,0.009188444250159794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,128,0.02090311050415039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,12288,0.019680889116393197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,5120,0.3728640079498291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,512,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,12288,0.3916586769951715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,1024,0.028299556838141546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,65536,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,256,0.018752000398106046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,128,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,5120,0.012882666455374824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,2048,0.042282667424943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,1536,0.020947555700937908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,32,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,256,0.017017778423097398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,10240,0.19726844628651938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,512,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,3072,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,4096,0.606553766462538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,16384,0.024551999237802293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,10240,0.012786666552225748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,64,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,2048,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,65536,0.008044444852405125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,1024,0.16237866878509521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,1024,0.030406223403082952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,2048,0.06652622090445624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,1536,0.010503999888896942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,32,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,2560,0.22603999243842232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,10240,0.37813599904378253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,3584,0.030971556901931763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,3072,0.010632889138327705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,768,0.05199733376502991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,12288,0.03973955578274197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,1024,0.0797297755877177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,16384,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,2560,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,128,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,1536,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,512,0.003863999827040566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,1536,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,1024,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,16384,0.013295110729005603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,256,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,6144,0.624767091539171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,65536,0.1265608867009481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,16384,0.053207112683190234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,512,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,768,0.021039111746682063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,3584,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,3584,0.035247110658221774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,256,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,7168,0.021858667333920796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,32,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,1536,0.042656888564427696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,1024,0.006139555739031897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,3584,0.033394667837354876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,768,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,65536,2.7267867194281687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,51200,0.20799288484785292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,3584,0.02588444451491038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,128,0.09607644213570489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,8192,0.2099075582292345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,7168,0.029535998900731403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,1024,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,128,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,3072,0.00886577781703737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,16384,0.056286222404903836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,256,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,128,0.0022728888110982045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,256,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,65536,0.1452044381035699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,5120,0.2270533243815104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,3584,0.007888000044557784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,16384,0.2025768889321221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,64,0.0018088887963030075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,4096,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,5120,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,3072,0.015822221835454304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,65536,0.5481208695305718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,65536,0.09895822074678208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,4096,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,2048,0.05243911014662849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,5120,0.12853422429826525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,16384,0.024868445263968572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,4096,0.008012444608741337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,768,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,4096,0.010836444795131683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,4096,0.05104533169004652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,2560,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,65536,1.2639101876152885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,1536,0.012264889147546558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,7168,0.0899839997291565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,3072,0.01202133297920227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,3072,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,4096,0.16418933868408203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,16384,3.7890294392903647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,2048,0.02998488810327318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,51200,0.06373866399129231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,128,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,2560,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,128,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,3584,0.039868444204330444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,2560,0.004893333133724001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,8192,0.3081208864847819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,4096,0.008737777670224508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,1536,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,4096,0.03798044390148587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,10240,0.028098666005664404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,2560,0.0069057775868309875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,2048,0.028502222564485338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,5120,0.027335999740494624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,4096,0.01756444407833947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,3072,0.1419235600365533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,1024,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,256,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,4096,0.02197155521975623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,3072,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,2560,0.026681777503755357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,2048,0.06506488720575969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,64,0.004232888834344016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,768,0.01664533383316464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,128,0.02823822365866767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,16384,0.031253334548738264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,51200,0.4617466926574707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,512,0.013628444737858243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,12288,0.04047733214166429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,512,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,512,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,8192,0.04593422346644931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,7168,0.07392177979151408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,51200,0.07441600163777669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,768,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,3072,0.016914667354689706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,3072,0.051002668009863965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,512,0.005527110977305307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,768,0.024718221690919664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,1536,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,2048,0.06396355231602986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,64,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,512,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,65536,0.9738106197781033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,3072,0.0954844421810574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,10240,0.024707555770874023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,3584,0.011101333631409539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,3072,0.021592888567182753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,2560,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,2560,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,51200,0.35508889622158474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,1536,0.012263110942310758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,2560,0.028672890530692205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,65536,0.5612177848815918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,2560,0.21145244439442953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,256,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,256,0.018184888694021437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,16384,2.075138727823893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,1536,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,128,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,3072,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,8192,0.04776444368892246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,6144,0.041173332267337374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,6144,0.01909244391653273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,128,0.021535999245113794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,2560,0.02296888828277588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,256,0.006179555422729916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,7168,0.04256177941958109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,8192,0.09556621975368923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,10240,0.05460000038146973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,2560,0.2573004563649495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,1536,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,5120,0.008789333204428354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,2560,0.08130844434102376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,3072,0.0042444442709287005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,12288,0.37636176745096844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,6144,0.0281333327293396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,10240,0.029203553994496662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,2048,0.6321217748853896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,8192,0.07902933491600884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,10240,0.018903111418088276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,3584,0.013095999757448832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,128,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,8192,0.02013777858681149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,768,0.012167111039161682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,4096,0.03331733412212796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,4096,0.0684808890024821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,16384,0.06371733215120104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,5120,0.02053511142730713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,2560,0.014977777997652689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,7168,0.010476444330480365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,16384,0.018546667363908555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,64,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,6144,0.030799110730489094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,32,0.013397333522637686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,2048,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,1536,0.014156444205178155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,65536,0.07751911216311984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,3072,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,4096,0.05335022343529595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,1536,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,51200,0.39534844292534727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,10240,0.018904889623324077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,16384,0.32658934593200684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,3072,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,5120,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,256,0.15543466144137913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,768,0.05501955416467455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,512,0.0031439999325407874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,16384,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,3072,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,32,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,51200,0.22463999854193795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,1024,0.025544888443417017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,10240,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,1024,0.019104000594880845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,1024,0.022875555687480505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,1536,0.017478222648302715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,128,0.003867555409669876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,3072,0.15338399675157335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,8192,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,8192,0.05831111139721341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,512,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,4096,0.03986400034692552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,4096,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,16384,0.022368000613318548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,256,0.016851555969980028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,768,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,1536,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,512,0.06734933455785115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,12288,0.035269333256615534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,3584,0.03955644369125366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,6144,0.2561519940694173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,65536,0.25483910242716473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,5120,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,2560,0.027707555227809485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,128,0.023316444622145757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,2048,0.02001422146956126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,64,0.003491555651028951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,2560,0.015955555770132277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,2560,0.014261333478821648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,512,0.020604444874657523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,7168,0.1457502179675632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,8192,0.23577332496643066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,2560,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,4096,0.025363556212849084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,512,0.029755555921130713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,12288,0.017895999881956313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,8192,0.009747555686367882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,8192,0.016827555166350473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,3072,0.13671289549933538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,3584,0.18330044216579863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,1024,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,2560,0.01385244478782018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,16384,0.17436356014675566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,51200,0.037873778078291155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,2560,0.021540444758203294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,3584,0.023183999790085688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,6144,0.04943111207750109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,1536,0.07385066482755873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,51200,26.378270467122395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,1536,0.020631111330456205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,8192,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,6144,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,512,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,6144,0.03048000070783827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,1536,0.22630932595994738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,3584,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,4096,0.010454222559928894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,256,0.005442666510740916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,768,0.009541333549552495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,1024,0.036855111519495644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,51200,0.10908889108233982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,1536,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,4096,0.018047110901938546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,64,0.008556444611814287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,8192,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,8192,0.011414222419261932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,51200,0.06765955686569214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,8192,0.03827733463711209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,256,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,3072,0.013555555707878537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,16384,0.036567999256981745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,5120,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,64,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,1024,0.00738666703303655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,3584,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,2048,0.00702311098575592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,3072,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,256,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,128,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,128,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,6144,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,12288,0.02075199948416816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,3584,0.02945599953333537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,65536,1.2370488908555772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,7168,0.05056711037953695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,1024,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,8192,1.5507538053724501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,65536,1.2266151640150282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,3072,0.03598311212327745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,512,0.007380444142553542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,12288,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,32,0.05772888660430908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,256,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,4096,0.03938222262594435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,12288,0.013088000317414602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,6144,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,10240,0.023991111252042983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,51200,0.13918577300177679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,6144,1.0892488691541884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,3072,0.012718222207493253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,7168,0.1606746647093031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,8192,0.15240355332692465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,7168,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,3584,0.043728000587887235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,3072,0.059663110309176974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,16384,1.1568319532606337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,768,0.03624799847602844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,6144,0.058992001745435924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,16384,0.03853955533769395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,6144,0.06237688991758558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,7168,0.12146400080786811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,512,0.0036346668170558084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,2048,0.029888000753190782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,16384,0.11122666464911567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,65536,0.12466222710079616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,5120,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,10240,0.06965688864390056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,6144,0.018039999736679923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,256,0.009460444251696268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,128,0.0019333333604865605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,512,0.0459191103776296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,768,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,6144,0.02661422226164076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,5120,0.013753777576817406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,2048,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,10240,3.001919216579861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,32,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,16384,0.6970168749491373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,1024,0.008269333177142674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,2048,0.037905779149797224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,51200,0.04896355668703715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,768,0.2491839991675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,32,0.018558222386572096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,3072,0.007789333661397298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,768,0.015238222148683337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,65536,1.1309786902533638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,1024,0.07329422235488892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,51200,0.12015377150641547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,768,0.09505333503087361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,128,0.10409955845938788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,512,0.010919111470381418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,10240,0.10117244720458984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,3072,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,64,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,1536,0.022729777627521094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,51200,0.662959098815918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,65536,0.29663289917839897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,512,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,4096,0.027298667364650305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,128,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,512,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,128,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,7168,0.017888888716697693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,5120,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,10240,0.06505599949094984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,4096,0.020242666204770405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,2048,0.0876657764116923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,7168,0.3961626688639323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,5120,0.04399111204677158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,16384,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,3072,0.244795560836792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,6144,0.17844711409674752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,51200,0.047107554144329496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,256,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,512,0.06742488675647311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,2048,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,5120,0.009194666312800514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,2560,0.04463111029730903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,2560,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,8192,0.11456355783674453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,512,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,128,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,32,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,5120,0.03155555658870273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,10240,0.026735110415352717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,256,0.11550845040215386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,256,0.015044444137149386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,6144,0.006549333532651265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,51200,6.617833879258897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,8192,0.05879555808173286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,2560,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,1024,0.03751733236842685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,16384,0.09098400010002984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,8192,0.039805332819620766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,5120,0.1315279934141371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,6144,0.030489779180950586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,2048,0.026038222842746313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,16384,4.778722551133898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,1024,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,10240,0.025753777888086107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,7168,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,512,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,7168,0.2615591155158149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,10240,0.015977778368526034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,3072,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,12288,0.0110871113008923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,51200,0.047818667358822294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,5120,0.017136888371573556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,3072,0.024069334069887798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,2560,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,12288,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,128,0.04952177736494276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,3584,0.007717333734035492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,6144,0.011703111231327057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,7168,0.07199377483791776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,128,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,2560,0.00852711084816191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,128,0.012519111235936483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,7168,0.12670666641659206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,3072,0.022321777211295232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,512,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,7168,0.9311804241604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,16384,0.02238044473859999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,2048,0.0042204442951414315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,2048,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,8192,0.08300622304280598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,65536,0.07623555262883504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,7168,0.06910489002863567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,128,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,10240,0.012464000119103326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,32,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,16384,0.05630489190419515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,3584,0.03645244571897719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,3584,0.04103199972046746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,7168,0.02288088864750332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,128,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,3072,0.012455999851226807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,65536,0.0766320029894511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,64,0.007735999921957652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,128,0.02698755595419142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,51200,0.9687386618720161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,16384,0.025681777132882014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,6144,0.08865244521035089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,4096,0.07037955522537231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,3584,0.009128000173303816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,5120,0.012880889077981314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,51200,0.34095554881625706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,3584,0.03080000148879157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,128,0.0064151109092765385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,10240,0.0366960002316369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,16384,0.0407680008146498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,1536,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,512,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,2048,0.03103644318050808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,768,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,1536,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,10240,0.25028355916341144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,1536,0.09807378053665161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,1024,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,12288,0.017078222499953378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,51200,0.48453866110907656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,1536,0.016936888297398884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,256,0.01238666640387641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,65536,0.11260355843438043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,10240,0.11410044299231635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,51200,0.9871991475423177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,4096,0.07159911261664496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,768,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,64,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,3072,0.0551511115498013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,2048,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,768,0.0038862224254343244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,256,0.09931021928787231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,64,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,12288,0.02788089050187005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,5120,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,51200,0.17124711142645943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,256,0.015665777855449252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,1024,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,1024,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,4096,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,1536,0.01796711153454251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,12288,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,5120,0.01568800045384301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,3072,0.013583111266295115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,7168,0.020575111111005146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,65536,0.4691653251647949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,8192,0.09571999973720974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,1024,0.020040000478426617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,512,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,10240,0.03562577896647983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,32,0.09878577788670857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,1024,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,3072,0.008208889100286696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,10240,0.03457955519358317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,4096,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,12288,0.0402453343073527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,12288,0.3710008992089166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,32,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,64,0.0018151110659042995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,256,0.005774222314357758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,12288,0.2140808900197347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,51200,0.3284986548953586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,10240,0.17072178257836235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,1536,0.004039999925427967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,128,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,6144,0.02647999922434489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,256,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,7168,0.18111111058129203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,1024,0.007833778030342525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,12288,0.03251911203066508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,1536,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,2560,0.0064088888466358185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,2048,0.0063635553750726916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,3584,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,12288,0.06799555487102933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,2560,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,6144,0.18650933106740317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,2560,0.04639110962549845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,16384,0.06414933337105645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,3072,0.027472888429959614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,3072,0.09867733054690891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,1024,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,16384,0.0741528868675232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,8192,0.012990222209029727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,1024,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,512,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,12288,0.029775109555986192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,3584,0.01682577696111467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,10240,0.1481226682662964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,1536,0.01649777756796943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,1536,0.04874222146140205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,16384,0.022289777795473736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,3584,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,10240,0.1782399945788913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,51200,0.06398133436838786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,1024,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,768,0.1606853273179796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,3072,0.018918222851223417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,512,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,4096,0.09649688667721218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,12288,0.03197688857714335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,2560,0.04707555638419258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,2048,0.00886933339966668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,256,0.09888355599509345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,4096,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,2048,0.01887911061445872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,65536,0.11553955078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,10240,0.013731555806265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,6144,0.3784444332122803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,5120,0.013256000147925483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,7168,0.06409777535332574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,512,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,128,0.0018942221585247253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,8192,0.08922311332490708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,256,0.004970666848950916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,2560,0.01538222200340695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,256,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,3584,0.011297778122954898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,2048,0.16381866402096218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,8192,0.026545777916908264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,1536,0.013835555149449242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,6144,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,768,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,65536,0.15563910537295872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,12288,0.013576000101036496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,32,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,768,0.033175110816955566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,3584,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,10240,0.9849155214097766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,1536,0.050902220937940806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,7168,0.015085332923465304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,16384,0.19155822859870064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,10240,0.029493331909179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,3072,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,8192,0.019293333093325298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,5120,0.01553422212600708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,512,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,51200,0.22539912329779732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,65536,1.0793893602159288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,12288,0.0387386679649353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,10240,0.023150222169028387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,12288,0.028183999988767836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,256,0.021611554755104914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,32,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,51200,0.08308621909883288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,64,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,6144,0.011863110793961419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,2560,0.004955555415815777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,2560,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,2048,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,1536,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,2048,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,32,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,4096,0.00871733327706655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,5120,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,5120,0.016340444485346477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,2560,0.04517066809866163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,6144,0.11196888817681207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,768,0.017504000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,6144,0.010075555907355415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,32,0.00619555554456181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,512,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,2048,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,3584,0.010148444109492833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,12288,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,1024,0.005987555616431766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,2048,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,5120,0.08907822105619644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,5120,0.018732445107565988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,1536,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,3072,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,3072,0.0523173345459832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,1024,0.010055111514197456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,512,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,12288,0.1687777837117513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,1536,0.06030933062235514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,8192,0.04383733206325107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,768,0.019312888383865356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,16384,1.7045680152045355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,65536,0.3609484566582574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,6144,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,512,0.01165422217713462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,10240,0.01886577738655938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,51200,0.1727599965201484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,32,0.026593777868482802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,2048,0.005379555539952383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,16384,0.5165297720167372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,256,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,512,0.14936178260379368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,1536,0.028130667077170476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,768,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,128,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,32,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,8192,0.39096087879604763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,2048,0.00423911131090588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,16384,0.02542488939232296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,768,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,1536,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,1024,0.033082667324278094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,6144,0.06490221950742933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,3584,0.021599110629823472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,32,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,256,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,6144,0.06896177927652995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,256,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,10240,0.029911110798517864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,3584,0.010277333358923594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,4096,0.02532711128393809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,65536,0.0935608877076043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,3584,0.01145333300034205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,6144,0.23010489675733778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,1536,0.02836266822285122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,768,0.01702133317788442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,65536,0.13898577955034044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,128,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,10240,0.18129332860310873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,16384,0.011411555939250521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,1536,0.010149333212110732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,1536,0.010266666611035665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,2560,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,8192,0.03604888916015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,1536,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,3584,0.007970666719807519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,7168,0.02427022159099579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,5120,0.10612710979249741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,128,0.038372443781958684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,6144,0.01216533366176817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,12288,0.18582933478885225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,65536,0.45685243606567383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,768,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,16384,0.03759377863672044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,5120,0.008352888955010308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,1024,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,10240,0.015243555108706156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,3072,0.02548888822396596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,5120,0.011047999891969891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,12288,0.07684799697664049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,2048,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,128,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,2048,0.00979377743270662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,128,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,768,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,1024,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,768,0.007450666692521837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,10240,0.34592178132798934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,6144,0.031565331750445895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,65536,0.3709546724955241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,256,0.0029804443733559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,51200,0.10212977727254231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,1024,0.15966755814022487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,2560,0.02603911028967963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,7168,0.027315555347336665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,768,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,128,0.024152888192070857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,3584,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,4096,0.017461333009931777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,32,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,3584,0.034675555096732244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,10240,0.017124444246292114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,8192,0.0184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,256,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,3584,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,512,0.0031946665710873077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,16384,0.0364586677816179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,1024,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,1536,0.029412445094850328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,4096,0.030249777767393324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,2048,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,65536,0.10382222466998631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,1536,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,6144,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,8192,0.11821688546074761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,12288,0.14752799934811062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,768,0.003928889003064898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,4096,0.6210533248053657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,6144,0.08674489127265082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,1024,0.013560000393125745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,65536,0.1421040031645033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,6144,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,256,0.035094221433003746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,1024,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,1536,0.013835555149449242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,8192,0.19175644715627035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,10240,0.014902222487661572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,6144,0.032305777072906494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,5120,0.019944889677895438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,3584,0.009937778115272522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,12288,0.01495733360449473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,512,0.02741777731312646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,65536,0.2083057827419705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,256,0.052842666705449425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,2048,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,1024,0.01625422139962514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,51200,0.0998248921500312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,12288,2.7388676537407766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,32,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,10240,0.04211911227968004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,10240,0.10119999779595269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,768,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,1024,0.008203555312421586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,12288,0.17534044053819445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,2560,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,7168,1.097679985894097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,2560,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,3072,0.016559999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,256,0.008159110943476358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,3072,0.5069022178649902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,512,0.007910221815109253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,3584,0.08237333430184259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,1536,0.010608889162540436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,5120,0.010304888917340172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,2048,0.026758222116364375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,2048,0.010991111397743225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,256,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,7168,0.01052266690466139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,3072,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,6144,0.013391999734772576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,1024,0.0270142224099901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,256,0.005208888815508948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,3584,0.036850667662090726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,3584,0.023801777097913954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,8192,0.027913779020309448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,16384,0.17348710695902506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,2560,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,5120,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,65536,0.0915715561972724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,768,0.01789955629242791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,65536,0.07698311408360799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,4096,0.024988444315062627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,3072,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,7168,0.0925262239244249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,10240,0.19393332799275717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,10240,0.1193288829591539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,1536,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,7168,0.010988444089889526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,4096,0.10038933489057754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,2048,0.17274933391147188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,1024,0.018035555879275005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,3584,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,3584,0.021287999219364587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,768,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,8192,0.058351112736595996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,51200,0.6021217770046657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,256,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,51200,0.09923022323184544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,51200,0.05419377817047966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,10240,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,4096,0.03100888927777608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,256,0.02009600069787767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,10240,0.012252444194422828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,2560,0.13264800442589655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,12288,0.017808000246683758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,256,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,64,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,4096,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,1536,0.03728355632887946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,10240,0.008751110898123847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,4096,0.018008889423476327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,8192,0.04624533322122362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,4096,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,5120,0.02068533334467146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,3584,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,3584,0.015463110473420886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,10240,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,64,0.3410080009036594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,8192,0.6134604348076714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,2048,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,32,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,16384,0.022170666191313002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,7168,0.1889244450463189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,2048,0.023988443944189284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,2048,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,2560,0.0233324451578988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,3072,0.019176888797018263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,6144,0.5525217586093479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,2048,0.022405332989162866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,32,0.0274311105410258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,12288,0.08277333445019193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,2048,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,16384,0.39583643277486164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,16384,0.03523200088077121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,2048,0.015018666783968607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,2560,0.02071822186311086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,1536,0.07234666744867961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,256,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,16384,0.13617955313788518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,65536,0.7555928760104709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,16384,1.7261902491251628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,4096,0.025423111187087163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,16384,0.03912977708710565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,3072,0.05922489033805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,8192,0.023777777949968975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,16384,0.03069066670205858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,4096,0.011762667033407422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,1024,0.0028151110228565005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,51200,0.04927555720011393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,16384,0.021475555168257818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,768,0.02438666588730282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,1024,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,2560,0.06208266814549764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,2560,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,5120,0.07357777489556207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,768,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,1024,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,2048,0.020046222541067336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,51200,0.4974942207336426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,4096,0.010316444767846001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,1024,0.018441778090265062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,64,0.002230222233467632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,64,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,512,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,6144,0.021828444467650518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,16384,0.035437332259284124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,10240,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,12288,0.024736000431908503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,7168,0.010460444622569613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,5120,0.03324888812171088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,3584,0.24148800637986925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,8192,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,10240,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,8192,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,64,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,5120,0.24523022439744738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,3584,0.01960533360640208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,128,0.032403555181291364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,12288,0.18675733937157524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,3072,0.03747289048300849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,2560,0.013012444807423485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,3584,0.09842577907774185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,12288,0.12852356168958876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,768,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,4096,0.0162444445821974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,8192,0.009591999981138442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,256,0.005380444642570283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,1024,0.028219554159376357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,3072,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,4096,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,512,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,3584,0.034288889831966825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,5120,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,1536,0.007107555866241455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,16384,0.0455644428730011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,7168,0.20529066191779244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,2048,0.2615662150912815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,768,0.03577866819169786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,2560,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,6144,0.02404444416364034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,1536,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,2048,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,10240,3.9979625278049045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,65536,0.24361512396070692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,10240,0.18792888853285047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,32,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,768,0.01797155539194743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,12288,0.028590222199757893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,65536,0.04045866595374213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,2048,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,256,0.020953777763578627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,1536,0.16771466202206084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,12288,0.2519297864702013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,5120,0.01607733302646213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,65536,0.27081510755750865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,6144,0.006047111004590988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,5120,0.11876266532474095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,1536,0.09555200073454116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,128,0.0028844444702068963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,12288,0.011003555523024665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,12288,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,5120,0.017672888106769986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,64,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,256,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,512,0.008440000315507254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,768,0.16295733716752794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,1536,0.038875556654400296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,1536,0.04285866684383816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,6144,0.016985777351591323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,7168,0.010046222143703038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,10240,0.004997333304749595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,1536,0.0058328890138202244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,256,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,512,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,6144,0.06242844131257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,7168,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,256,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,2048,0.014901333385043673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,8192,0.014560000763999091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,2560,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,2048,0.07696355713738336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,8192,0.016546666622161865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,512,0.010679999987284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,512,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,256,0.03868000043763055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,16384,0.07371377944946289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,7168,0.010983111129866706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,256,0.017581333716710407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,2560,0.015576889117558798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,16384,0.04227911101447212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,16384,0.10435466633902656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,128,0.07024888859854804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,5120,0.13260800308651394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,65536,0.023327999644809302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,2560,0.04430400000678169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,512,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,2560,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,10240,0.018256000346607633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,3072,0.024280889166726008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,32,0.006779555645253923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,51200,0.05099199877844917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,3584,0.30851467450459796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,7168,0.033379554748535156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,5120,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,7168,0.07162311342027453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,256,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,6144,0.07624089055591159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,4096,0.0513573322031233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,7168,0.16067377726236978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,2048,0.01846311158604092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,12288,0.14407200283474392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,2560,0.014746665954589844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,2560,0.027823110421498615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,6144,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,6144,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,1536,0.02417244513829549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,768,0.0038115556041399636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,3584,0.5949253506130642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,7168,0.02167644434505039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,6144,0.012490666574902005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,128,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,3072,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,10240,0.15878578027089438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,3584,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,5120,0.02137600051032172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,768,0.016197333733240765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,3072,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,64,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,3584,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,2048,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,512,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,256,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,65536,0.07184977663887872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,256,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,64,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,12288,0.03408711155255636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,16384,0.7699857817755805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,65536,0.7243759897020129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,16384,0.05003733436266581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,128,0.010422222316265106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,2048,0.01800355480776893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,7168,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,3072,0.025410667061805725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,65536,0.32888534333970815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,128,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,51200,0.5354213184780544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,2048,0.07004177570343018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,768,0.07634222507476807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,2560,0.09018311235639785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,65536,0.061305774582756885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,12288,0.21219110488891602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,10240,0.052780446079042226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,1024,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,65536,0.055471999777687915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,12288,0.041582223441865705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,8192,0.020963556236690942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,65536,0.06261866622500949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,256,0.01389422184891171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,8192,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,3072,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,5120,0.023338667220539514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,4096,0.026779555612140234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,51200,0.016753777861595154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,128,0.001957333336273829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,2048,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,512,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,128,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,6144,0.059691554970211454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,768,0.017982222967677645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,1536,0.011347555451922946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,2048,0.013644444445768992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,12288,0.17628267076280382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,51200,0.013245333400037555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,65536,2.401269276936849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,2048,0.07627822293175592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,64,0.0032844444115956626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,1024,0.007187555233637492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,1536,0.1315982209311591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,256,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,10240,0.13315644529130724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,51200,0.05071555574735006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,10240,0.44444624582926434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,768,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,128,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,2560,0.017270222306251526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,3584,0.008414222134484185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,5120,0.03446133269204034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,51200,14.709560818142362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,4096,0.020612445142534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,1536,0.01458755632241567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,512,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,10240,0.012303999728626676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,5120,0.0506640010409885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,768,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,768,0.04759822289148966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,10240,0.06786844465467665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,2560,0.019831111033757527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,3584,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,8192,0.11672443813747829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,4096,0.009535999761687385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,256,0.01755822201569875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,7168,0.045540445380740695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,1024,0.017535999417304993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,3584,0.04141244292259216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,256,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,12288,0.013491555220550962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,7168,0.056695110268063016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,3072,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,768,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,10240,0.04920533299446106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,7168,0.5972888734605577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,5120,0.017421333326233756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,16384,0.13836799727545843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,3072,0.02517066730393304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,768,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,64,0.0019537777536445195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,6144,0.016751110553741455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,51200,0.11474311351776123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,3584,0.03479911221398248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,2560,0.027002667387326557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,1024,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,128,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,2048,0.28388799561394584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,7168,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,10240,0.2034462160534329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,5120,0.16930045021904838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,12288,0.032413333654403687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,8192,0.058952000406053334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,256,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,256,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,512,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,512,0.0034826666944556763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,3072,0.0277164445983039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,16384,1.0125040478176541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,512,0.06517066558202107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,6144,0.013087111214796701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,12288,0.013267555170589022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,5120,0.03756444321738349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,4096,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,8192,0.030850665436850652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,256,0.00906133320596483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,51200,0.8211750984191895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,51200,0.11601332823435466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,32,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,768,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,4096,0.02749866743882497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,64,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,1024,0.006084444622198741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,1024,0.020231111182106864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,12288,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,1536,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,512,0.10782221953074138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,768,0.07059644328223334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,2560,0.02387111054526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,1536,0.008959111240175035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,1536,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,8192,0.040727999475267194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,8192,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,3584,0.11411555608113606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,128,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,128,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,32,0.0021235555824306277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,4096,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,8192,0.014520888527234396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,3072,0.3875093460083008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,3584,0.054754667811923556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,8192,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,3584,0.016927111479971144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,65536,0.04121066795455085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,768,0.016196444630622864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,51200,0.11361511548360188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,3584,0.08457599745856391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,256,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,16384,0.08070044385062324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,768,0.0038124442928367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,8192,0.005518222020732031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,2048,0.01891999940077464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,512,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,16384,0.030831111801995173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,1024,0.013576000101036496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,1536,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,16384,0.5926737785339355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,8192,0.10075110859341091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,10240,0.20504889223310682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,1024,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,6144,0.01405244403415256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,1536,0.012884444660610624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,256,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,51200,0.22421243455674914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,1024,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,32,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,8192,0.024549333585633173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,51200,0.07082221905390422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,10240,0.052798221508661904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,256,0.012854222622182635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,5120,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,1024,0.038935999075571694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,1536,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,51200,0.8871519830491809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,51200,9.394018385145399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,65536,0.6623146798875597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,3072,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,8192,0.013740444348918067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,8192,0.08617777956856622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,12288,0.34632711940341526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,65536,0.08854222297668457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,2048,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,6144,0.07354844278759427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,768,0.018017777138286166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,6144,0.014755555325084262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,7168,0.1199715534845988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,2048,0.006086222413513396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,51200,0.07697333229912652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,51200,0.5041288799709743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,3584,0.036882665422227644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,10240,1.1601635615030925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,128,0.005649777750174205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,256,0.013825777504179211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,4096,0.009945777555306753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,10240,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,128,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,1536,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,7168,0.12917421923743352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,16384,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,16384,0.0496346652507782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,6144,0.030018667380015056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,256,0.012239110966523489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,12288,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,2048,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,512,0.005408000200986862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,7168,0.03329955538113912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,1024,0.022319111559126113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,128,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,5120,0.01926488843229082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,5120,0.24589512083265516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,12288,0.06763555606206258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,64,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,256,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,768,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,8192,0.12939466370476618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,1536,0.04977155394024319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,16384,0.02093511157565647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,512,0.011927111281288994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,768,0.1054497758547465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,5120,0.0110880004035102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,16384,0.0648933317926195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,5120,0.12826489077674016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,51200,0.05334044496218363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,1536,0.04617599977387322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,32,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,3584,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,3072,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,2048,0.044216887818442456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,65536,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,65536,0.08687377638287014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,2048,0.11097600062688191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,4096,0.03313422203063965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,7168,0.04248088929388258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,512,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,65536,0.15278755293952093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,16384,0.3313395447201199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,10240,0.25928089353773326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,128,0.044871111710866295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,3584,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,768,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,4096,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,6144,0.010212444596820408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,6144,0.011372444530328115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,1024,0.13311644395192465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,7168,0.019391111201710172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,64,0.048671109808815845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,2048,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,3072,0.020287111401557922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,32,0.02550933261712392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,51200,0.4341013431549072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,8192,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,256,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,16384,1.8915884229871962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,10240,0.304695102903578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,4096,0.02602844436963399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,12288,0.07254578007592095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,51200,6.39333258734809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,1536,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,512,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,51200,0.1956648826599121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,12288,0.017098666893111337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,8192,0.014679999815093147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,65536,0.2006231149037679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,128,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,2048,0.0073520003093613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,3072,0.053025778796937734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,768,0.028179556131362915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,128,0.031967110104031034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,256,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,512,0.0029760001020299066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,3072,1.0771520402696397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,7168,0.027109333210521277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,5120,0.05029777685801188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,3584,0.01221333361334271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,1024,0.0199955552816391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,1024,0.015268445014953613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,2048,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,6144,0.019186667270130582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,8192,0.04952444301711189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,2560,0.01236088905069563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,12288,0.024392889605628118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,3072,0.007584888901975419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,4096,0.04262311259905497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,8192,0.1781546672185262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,51200,0.5665004518296983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,12288,0.1338933308919271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,128,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,12288,0.01832355558872223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,512,0.01590844492117564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,64,0.001785777716173066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,32,0.021957332889238994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,7168,0.11790400081210667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,10240,0.021904889080259535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,3584,0.026345777842733595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,64,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,16384,0.02159999973244137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,8192,0.011007111105653973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,128,0.005342222336265776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,1536,0.04053511222203573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,51200,0.24954488542344835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,2048,0.01425866617096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,128,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,5120,0.02122933334774441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,1024,0.06804000006781684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,3072,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,10240,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,256,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,7168,0.023033777872721355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,65536,0.050336887439092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,768,0.012463111016485425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,256,0.0457217792669932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,768,0.017247110605239868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,3584,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,1536,0.05685066514545017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,4096,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,7168,0.035150223308139376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,2560,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,768,0.013750221994188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,3072,0.02054933375782437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,10240,0.2881715562608507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,51200,1.1427378124660916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,12288,0.1967351171705458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,16384,0.2938506603240967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,6144,0.022844443718592327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,16384,0.25456976890563965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,1536,0.019188443819681805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,12288,0.024354666471481323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,6144,0.7842115296257867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,3072,0.014726221561431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,64,0.006691555596060223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,65536,13.454210917154947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,256,0.032023111979166664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,7168,0.02091377807988061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,128,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,256,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,51200,0.09385066562228733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,256,0.011526222030321756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,65536,0.5472773445977105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,8192,0.09556177589628433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,65536,0.1806666718588935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,256,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,3072,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,1024,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,768,0.06793510913848877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,10240,0.029475556479560003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,12288,0.06657510995864868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,768,0.007710222568776872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,5120,0.08744444449742635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,256,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,512,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,2048,0.004232888834344016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,51200,0.6540586683485243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,4096,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,512,0.008430221842394935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,8192,0.025776889589097764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,256,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,7168,0.02334311107794444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,2048,0.03662577933735318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,16384,1.1625395880805123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,7168,0.024393777052561443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,7168,0.019907555646366544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,32,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,2048,0.019539554913838703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,51200,0.3386382261912028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,8192,1.785128911336263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,10240,0.016568889220555622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,16384,0.02873599860403273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,7168,0.3831360075208876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,128,0.0017866666118303935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,128,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,16384,0.21227999528249106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,1536,0.012855999999576144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,16384,0.021735999319288466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,8192,0.026363554928037856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,10240,0.0370124446021186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,768,0.019834667444229126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,16384,0.09938755300309922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,65536,0.1434719959894816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,2048,0.03915822174814012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,1536,0.022184888521830242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,32,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,1024,0.043644444810019604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,8192,0.047750221358405225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,512,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,5120,0.01146311147345437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,1536,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,51200,1.269028451707628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,3584,0.007722666694058313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,1024,0.07519377602471246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,512,0.0026826666047175727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,1536,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,1536,0.2791786723666721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,10240,0.018600889378123812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,2048,0.0203484445810318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,128,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,5120,0.05215466684765286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,12288,0.027756444282001917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,1536,0.007035555938879649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,6144,0.12304533852471246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,6144,0.15070845021141901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,128,0.012830221818553077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,5120,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,16384,0.04553066690762838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,8192,0.011751111182901593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,12288,0.05685422155592176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,128,0.06411200099521212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,128,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,2560,0.026816000541051228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,8192,0.014344000154071383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,256,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,12288,0.18190399805704752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,5120,0.05260266529189216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,16384,0.08698578013314141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,128,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,6144,0.027290667096773785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,5120,0.8878124554951986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,1024,0.0038693332009845306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,51200,0.05983377827538384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,2048,0.030032889710532293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,7168,0.010543111297819348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,6144,0.07847377989027235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,3584,0.02844266759024726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,4096,0.07039022445678711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,3584,0.036601778533723615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,10240,0.1557448837492201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,256,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,65536,0.06492444541719225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,4096,0.03360622127850851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,256,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,12288,0.029770665698581274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,512,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,1536,0.008815999660227034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,65536,0.06425511174731784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,1024,0.09212088584899902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,2560,0.02537511123551263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,512,0.008840888738632202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,65536,1.5609768761528864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,8192,0.09324177768495347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,1536,0.04431822233729892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,3584,0.03569422165552775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,1536,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,3584,0.03659644391801622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,7168,0.013810666898886362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,16384,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,2048,0.005697777701748743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,7168,0.687764432695177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,1536,0.10589422119988336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,3584,0.06427733103434245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,768,0.040807998842663236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,6144,0.046535998582839966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,16384,0.033634665939542986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,3072,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,3072,0.2988479932149251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,2560,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,2048,0.03070488903257582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,1024,0.0392506652408176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,8192,0.10629689031177098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,3584,0.033313777711656355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,3072,0.011678222152921887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,2048,0.012883555557992725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,256,0.018949333164427016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,3072,0.03943555553754171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,7168,0.010111110905806223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,5120,2.1328648461235895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,8192,0.019893333315849304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,1024,0.01686044368478987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,12288,1.15465513865153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,7168,0.014848000473446317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,51200,0.025175111161337957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,65536,0.5471715397304958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,65536,0.16316533088684082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,6144,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,256,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,12288,0.017304889029926725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,2560,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,7168,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,8192,0.13701155450608996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,65536,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,12288,0.25288889143202037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,2048,0.02665155629316966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,7168,0.20961866113874647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,8192,0.02548622257179684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,2048,0.006719111154476802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,3584,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,5120,0.0757768882645501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,768,0.006020444548792309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,16384,0.127966218524509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,3072,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,3584,0.013348444468445249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,5120,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,3072,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,768,0.018423999349276226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,8192,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,65536,0.4992577764723036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,3072,0.0070462218589252895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,6144,0.019813333948453266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,16384,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,2048,0.010776888993051318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,8192,0.037026666932635836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,768,0.006136000156402588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,4096,0.033334222104814314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,51200,0.13228444258371988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,64,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,16384,0.027099554737408955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,6144,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,256,0.015761777758598328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,3072,0.016717332932684157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,6144,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,128,0.04338488976160685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,64,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,4096,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,256,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,6144,0.10809244712193807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,256,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,64,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,1024,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,128,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,2048,0.02010933392577701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,7168,0.01278133359220293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,7168,0.021503999829292297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,4096,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,768,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,8192,0.02106311089462704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,2048,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,1024,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,65536,1.5134710735744898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,8192,0.43748090002271867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,5120,0.0185244447655148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,4096,0.3630399968889024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,768,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,1536,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,128,0.005680000202523337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,16384,0.04986400074428982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,32,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,7168,0.01663466625743442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,12288,0.02323377794689602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,2560,0.014651555154058667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,2048,0.12252888414594863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,4096,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,51200,0.05091999967892965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,7168,0.2174391084247165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,3584,0.08451377683215672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,51200,0.3379884560902913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,2048,0.4139546553293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,768,0.014745778507656522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,4096,0.07449422279993693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,256,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,128,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,65536,0.7195030848185221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,512,0.002542222125662698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,768,0.012125333150227865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,3584,0.17328800095452201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,5120,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,4096,0.015660444895426433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,768,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,128,0.0022106667359670005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,64,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,128,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,6144,0.011188444164064197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,65536,0.06900355550977919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,32,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,3584,0.01756177842617035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,51200,0.06713777780532837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,3584,0.017978666557206046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,2048,0.013000888956917657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,32,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,51200,0.6861866845024956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,6144,0.013526221944226159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,1024,0.019549333386951022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,32,0.07811288701163398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,51200,0.30148622724745006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,2048,0.01926933394538032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,2048,0.03842755489879184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,768,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,32,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,12288,0.1700613366232978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,32,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,16384,0.6457964579264323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,51200,0.37407999568515354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,2048,0.0295448899269104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,6144,0.01849422189924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,1536,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,3584,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,16384,0.020113777783181932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,256,0.009764444496896531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,65536,0.5294622315300835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,51200,0.04714044597413805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,2048,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,512,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,12288,0.03342755635579427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,1024,0.015017777681350708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,1536,0.020988444487253826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,65536,0.08172977632946439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,5120,0.11491200659010146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,3072,0.030569778548346624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,12288,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,8192,0.03784444597032335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,3584,0.02677333354949951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,10240,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,16384,0.14693689346313477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,16384,0.4777973492940267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,6144,0.058232002788119845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,64,0.023337778117921617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,512,0.0053084443012873335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,7168,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,6144,0.10001689195632935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,10240,0.017326222525702584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,512,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,65536,0.07418133152855767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,51200,0.202656004163954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,2560,0.022504889302783545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,4096,0.06180266539255778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,1024,0.02167911165290409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,4096,0.02052977846728431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,512,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,128,0.024528889192475214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,2560,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,10240,0.02870933214823405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,10240,0.21006400055355498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,51200,0.1440088881386651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,2560,0.015096889601813423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,12288,0.03427555494838291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,3072,0.039094222916497126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,2560,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,2560,0.032460444503360324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,51200,0.3760133319430881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,2048,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,1536,0.012310222619109683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,12288,0.2857075532277425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,1536,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,51200,5.533430311414931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,512,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,1024,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,7168,0.012145777543385824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,3072,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,8192,0.020780444145202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,2048,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,5120,0.004775111046102312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,768,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,10240,0.03159911102718777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,2560,0.0075297777851422625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,8192,0.05840177668465508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,256,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,1024,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,1024,0.015804444750150044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,4096,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,2048,4.592147403293185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,128,0.008386666576067606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,3584,0.044715556833479136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,2048,0.07683644692103068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,4096,0.010178666975763109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,5120,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,6144,0.013630222115251752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,32,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,1024,0.027356444133652583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,10240,0.02693333393997616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,2560,0.02712000078625149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,2048,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,2048,1.249492433336046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,3584,0.026076444321208533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,2560,0.029715554581748113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,6144,0.027190221680535212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,2048,0.010263111028406356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,768,0.04122666517893473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,12288,0.2412346733940972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,2560,0.03977510995335049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,2560,0.042315555943383105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,6144,0.031770666440327965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,64,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,32,0.018568888306617737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,5120,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,3072,0.027967111931906805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,6144,0.020598222812016804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,7168,0.03539022141032749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,32,0.004839999808205498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,12288,0.08844355742136638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,65536,0.44451379776000977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,2560,0.010125333236323463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,10240,0.016729777057965595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,4096,0.3590337700313992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,3072,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,512,0.019730667273203533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,1536,0.004034666551484002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,8192,0.01738933391041226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,65536,0.0873288909594218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,128,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,512,0.01201866649919086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,256,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,128,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,256,0.014183110660976835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,768,0.00489688871635331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,2048,0.3983502123090956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,512,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,2048,0.14927111731635198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,4096,0.016297777493794758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,1536,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,4096,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,5120,0.15841778119405112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,8192,0.33947110176086426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,65536,0.4476453463236491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,3072,0.024392889605628118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,768,0.03673600157101949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,7168,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,512,0.0032951111594835916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,3584,0.010658666491508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,12288,2.5407191382514105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,65536,0.17596533563401964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,12288,0.16633244355519614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,6144,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,5120,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,65536,0.09859999683168198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,12288,0.03621600071589152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,768,0.018020444446139865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,65536,0.6471021970113119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,4096,0.008580444587601556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,65536,0.13137155108981663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,8192,0.12740000089009604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,16384,0.05318133367432488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,2560,0.02200800014866723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,7168,0.018607111440764535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,128,0.01484266585773892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,65536,11.992676628960503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,3584,0.008751110898123847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,768,0.02166755497455597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,6144,0.5868657959832085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,7168,0.07938755883110894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,8192,0.0258240004380544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,768,0.024263999528355066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,2560,0.010223111344708337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,512,0.007038222418891058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,2048,0.027791112661361694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,65536,0.1450559960471259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,1024,0.06686133146286011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,5120,0.011112000379297467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,6144,0.010324444207880232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,8192,0.02050488856103685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,64,0.004230222354332606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,4096,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,7168,0.03451822201410929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,4096,0.016787555482652452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,4096,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,10240,0.013722666435771518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,3072,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,3072,0.017113778326246474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,1024,0.016523554921150208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,65536,0.2868879901038276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,2048,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,5120,0.00889066689544254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,5120,0.009778666827413771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,512,0.05095555716090732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,16384,0.033906665113237165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,16384,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,3584,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,2560,0.0410755541589525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,1536,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,512,0.0033377777371141645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,128,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,2048,0.016964443855815463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,10240,0.03166044420666165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,16384,0.044046223163604736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,16384,0.9277626673380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,4096,0.015312888556056552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,128,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,768,0.00489155575633049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,512,0.037335998482174344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,768,0.04426577687263489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,3584,0.016075554821226332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,10240,0.015289778510729471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,2560,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,128,0.07668266693751018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,6144,0.010090666512648264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,51200,0.04918933245870802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,51200,0.28448179033067494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,7168,0.02164711058139801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,7168,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,2048,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,1024,0.02222577730814616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,65536,0.2836951149834527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,6144,0.00812177773978975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,12288,0.15161689122517905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,2048,3.084665722317166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,8192,0.03016622198952569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,51200,0.049560887946022876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,8192,0.17965688970353869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,65536,0.30243643124898273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,768,0.029359999630186293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,3584,0.04471822248564827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,2048,0.03898222247759501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,512,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,10240,0.25452889336480033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,128,0.007342221836249034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,1024,0.016220443778567843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,512,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,128,0.005453333258628845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,12288,0.4416915575663249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,7168,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,768,0.028596444262398615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,10240,0.05784533421198527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,128,0.017808889349301655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,65536,0.49473778406778973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,4096,0.1338906685511271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,512,0.022293332550260756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,16384,0.1752613385518392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,8192,0.00850311087237464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,512,0.020263110597928364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,7168,0.03859555390146043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,1024,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,3584,0.036299556493759155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,2560,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,32,0.0018933332628673976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,5120,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,256,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,65536,0.08454489045672947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,768,0.018546667363908555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,65536,1.7760995229085286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,8192,0.09300533268186782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,1024,0.022037333912319605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,256,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,2048,0.03991733325852288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,256,0.015013333823945789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,3584,0.08733066585328843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,32,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,10240,0.017345777816242643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,128,0.00563733321097162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,5120,0.32391378614637584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,256,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,3072,0.046978665722741016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,256,0.031010667483011883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,12288,0.05097244514359368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,4096,0.008656000097592672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,51200,0.3739893436431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,2560,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,7168,0.10876799954308404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,6144,0.03632622294955783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,51200,0.1568275557623969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,2560,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,6144,0.029912001556820337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,1536,0.014722666806644864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,65536,0.21967021624247232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,8192,0.11327644189198811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,768,0.07762489053938124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,2048,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,2048,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,512,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,7168,0.005504888792832692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,4096,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,12288,0.18143022060394287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,1536,0.01752800080511305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,10240,0.15442132949829102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,512,0.0370782216389974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,3584,0.02247911029391819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,65536,0.26583022541469997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,512,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,4096,0.750274658203125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,6144,0.12278489271799724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,10240,0.018486221631368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,65536,0.06225599845250448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,3584,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,65536,0.0644328859117296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,7168,0.03412266572316488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,64,0.11876000298394097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,3584,0.032313777340783015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,8192,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,3584,0.007838221887747446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,768,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,65536,0.17977600627475313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,768,0.02137600051032172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,256,0.013049778011110095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,8192,0.4990551206800673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,6144,0.027133334014150832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,768,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,32,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,65536,0.1209546724955241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,5120,0.06834044721391466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,65536,0.06517422199249268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,16384,0.032764444748560585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,4096,0.008020444048775567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,256,0.026571555270089045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,51200,0.42566222614712185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,51200,0.13676977157592773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,3584,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,2048,0.00665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,3072,0.07882489098442926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,12288,0.019136000010702345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,16384,0.012694222231705984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,7168,0.026660444007979497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,65536,0.12354311678144668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,32,0.015110222829712762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,16384,0.01591911084122128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,768,0.01202133297920227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,7168,0.028162668148676556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,512,0.03594755464129978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,1024,0.04875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,128,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,16384,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,512,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,65536,0.08303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,2560,0.008464888566070134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,3584,0.012597333225939008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,128,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,2560,0.008200888832410177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,3584,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,65536,0.05898400147755941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,2560,0.01629599928855896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,32,0.002436444370283021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,65536,0.06422133578194512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,10240,0.14341066943274602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,1536,0.04049955474005805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,10240,0.008709333009190029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,8192,0.08026755518383451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,256,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,4096,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,64,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,7168,0.05139999919467502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,256,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,10240,0.07024088833067152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,2560,0.044815109835730664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,10240,0.5431217617458767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,5120,0.03835466504096985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,1024,0.008605333666006723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,51200,2.923494127061632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,12288,0.04532355401251051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,1536,0.05680533250172933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,4096,0.07067822085486518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,1024,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,1536,0.0380711125002967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,1024,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,51200,0.20978311697642008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,10240,0.04199111130502489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,16384,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,256,0.002944888960984018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,5120,0.016706667012638517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,256,0.01240888900227017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,2048,0.03584711088074578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,51200,1.370545811123318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,16384,0.06530133220884535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,8192,0.014096889230940076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,1024,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,16384,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,8192,0.03938044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,3072,0.0221742226017846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,12288,0.08338133494059245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,2560,0.04903733399179247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,2048,0.022210665875011023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,2560,0.012144000165992312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,51200,0.12258399857415093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,51200,0.08755377928415935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,32,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,7168,0.03215377860599094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,32,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,7168,0.17167288727230498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,4096,0.04155644443300035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,6144,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,768,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,8192,0.030606223477257624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,512,0.012127111355463663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,51200,0.05589689148796929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,8192,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,4096,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,6144,0.0454364452097151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,6144,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,6144,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,512,0.011321778098742167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,512,0.005676444619894028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,8192,0.5932364463806152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,1024,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,1536,0.0041546668443414895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,3072,0.0240328891409768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,12288,0.060508443249596484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,8192,0.011878222227096558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,8192,0.6625990867614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,6144,0.10576533608966404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,4096,0.025638222694396973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,768,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,2560,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,2560,0.018745778335465323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,65536,0.09408089187410142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,51200,0.19568888346354166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,51200,0.07399466964933607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,7168,0.07277777459886339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,2560,0.036200887627071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,256,0.011515555282433828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,4096,0.02790044413672553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,65536,0.305218670103285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,128,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,1024,0.014019555515713163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,1536,0.005432888865470886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,64,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,8192,0.020220445262061227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,65536,0.8782462014092339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,3072,0.025632888078689575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,6144,0.01795288920402527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,12288,0.05424888928731283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,1024,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,128,0.05292355683114794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,4096,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,1536,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,6144,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,32,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,6144,0.020238222347365487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,16384,0.014428445034556918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,51200,0.048672000567118325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,10240,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,512,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,5120,0.0706346697277493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,1536,0.013223999904261695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,3072,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,2560,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,65536,0.02342222299840715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,2048,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,51200,0.05968799855973986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,768,0.02852177619934082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,256,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,4096,0.02767555581198798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,1536,0.026134222745895386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,12288,0.019211555520693462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,128,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,4096,0.10508622063530816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,2048,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,65536,0.07436444362004598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,4096,0.009514666265911525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,768,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,512,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,10240,0.01607466737429301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,12288,0.014000889327791003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,65536,0.07320711347791883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,5120,0.028037332826190527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,768,0.0034906665484110513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,2048,0.04947466651598612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,128,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,4096,0.009527111219035255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,128,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,768,0.014838222000333997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,512,0.021538666552967493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,128,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,10240,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,1024,0.014341332846217685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,1024,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,5120,0.05367822117275662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,3072,0.014716444744004143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,1024,0.011090666883521609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,51200,0.09688622421688503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,768,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,51200,0.26539733674791127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,65536,1.5651369094848633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,1536,0.07140355639987521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,2048,0.30454932318793404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,10240,0.014327110515700446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,256,0.09443910916646321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,3072,0.04288799895180596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,256,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,128,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,12288,1.7636391321818035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,768,0.21846932835049102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,3072,0.0336444444126553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,10240,0.3740968969133165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,512,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,3072,0.020340444313155282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,512,0.02777333392037286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,4096,0.008566222257084316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,128,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,768,0.016629333297411602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,64,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,1536,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,51200,0.050359110037485756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,1536,2.4483004675971136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,128,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,3072,0.04236444499757555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,10240,0.08768000205357869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,51200,0.10263911220762466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,32,0.0017928888814316855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,512,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,12288,0.017695999807781644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,4096,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,7168,0.015090665883488126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,10240,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,3584,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,2048,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,51200,0.13938489225175646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,3584,0.04704799917009142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,32,0.05416977736685011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,128,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,768,0.031882666879230075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,768,0.024691555235120986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,51200,0.02735288937886556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,4096,0.015488889482286243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,5120,0.043232887983322144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,10240,0.043291555510626904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,1536,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,10240,0.029871109459135268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,5120,0.12662400139702693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,51200,10.842355516221788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,768,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,1024,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,2048,0.03025955624050564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,12288,0.09179821941587661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,7168,0.055119110478295215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,64,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,10240,0.07094489203559028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,3584,0.07061955663892958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,65536,0.06258755260043673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,256,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,4096,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,6144,0.031035555733574763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,12288,0.0690915584564209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,64,0.00941955546538035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,65536,0.06758044163386027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,3584,0.028135998381508723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,1024,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,1536,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,512,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,1536,0.0038293335172865125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,1536,0.014197332991494073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,5120,0.06626577509774102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,256,0.010764444039927589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,128,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,4096,0.05970222420162625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,6144,0.011376000112957425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,6144,0.049088888698154025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,7168,0.009654222263230218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,3072,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,768,0.01773422294192844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,5120,0.5331964492797852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,12288,1.5940728717380102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,8192,0.21814310550689697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,10240,0.026898667216300964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,2560,0.43769065539042157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,4096,0.019307555423842538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,16384,0.05222488774193657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,12288,0.5499182277255589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,32,0.008113777471913232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,4096,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,12288,0.7383004294501411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,8192,0.024604444702466328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,65536,0.04683466752370199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,2560,0.22921599282158744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,7168,0.26724176936679417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,256,0.10684444506963094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,256,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,5120,0.031939556201299034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,1536,0.05843377775616116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,51200,0.3263351122538249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,12288,0.25031201044718426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,65536,1.2782026926676433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,10240,0.14556177457173666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,3072,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,10240,0.08839822477764553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,5120,0.0469742218653361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,2048,0.005056000004212062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,128,0.0019164443429973389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,768,0.02234488891230689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,128,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,8192,0.05605689022276136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,4096,0.018370666437678866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,32,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,256,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,128,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,7168,0.02238666680124071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,64,0.010321777727868822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,1024,0.020573332905769348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,256,0.03694044550259908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,4096,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,5120,0.042895111772749156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,3584,0.13966755072275797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,10240,0.015414221419228448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,65536,0.35043022367689347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,6144,0.027375110321574744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,16384,0.08031821913189359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,2048,0.0185217774576611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,2048,0.00904799997806549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,3584,0.08845066361957127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,512,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,65536,0.14665955967373318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,2048,0.021874666213989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,2048,0.012440889245933957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,1536,0.04316177633073595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,6144,0.03573955429924859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,16384,0.0362391107612186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,65536,0.5171173413594564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,10240,0.10930222272872925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,1024,0.04587466518084208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,6144,0.14328711562686497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,8192,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,7168,0.08996711174647014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,1024,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,3584,0.029437333345413208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,5120,0.03899644480811225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,65536,0.15472355153825548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,12288,0.21447555224100748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,16384,0.021186666356192693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,32,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,256,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,16384,0.8442151281568738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,4096,0.015102222561836243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,6144,0.2746888796488444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,65536,1.2829280429416232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,64,0.09408444166183472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,7168,0.29574667082892525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,3072,0.011770666473441653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,2048,0.04309155543645223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,4096,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,16384,0.0226239992512597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,16384,0.2469226784176297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,3584,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,2048,0.013518222504191928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,256,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,3072,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,2560,0.03077600068516201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,8192,0.0340266658200158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,2560,0.04836622211668226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,768,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,12288,0.023360000716315374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,2560,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,3072,0.02749866743882497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,1536,0.004969777746333016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,8192,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,65536,0.06172355678346422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,7168,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,256,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,4096,0.014242667290899487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,7168,0.03735377722316318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,5120,0.027464000715149775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,4096,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,2560,0.03315111001332601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,65536,1.8917920854356554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,3072,0.24302755461798775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,16384,0.0661191079351637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,16384,0.022892443670166865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,6144,0.012560000022252401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,32,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,10240,0.06928355826271905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,16384,0.028386665715111628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,3584,0.0869413349363539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,1024,0.019736889335844252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,2048,0.009509333305888707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,1536,0.029343999094433252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,3072,0.5644613371955024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,1536,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,2560,0.02054666644997067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,16384,0.2919786771138509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,8192,0.14207555188073054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,128,0.07226400242911445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,6144,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,7168,0.03735911183887058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,2560,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,3072,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,5120,0.021352888809310064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,256,0.010664000279373594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,256,0.07371911075380114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,256,0.1327608823776245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,3072,0.01978844404220581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,4096,0.021753778060277302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,64,0.034486220942603216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,128,0.011362666885058085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,5120,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,768,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,7168,0.06258844667010836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,5120,0.007423111134105259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,3072,0.02714311083157857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,768,0.02608799934387207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,128,0.002114666625857353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,12288,0.011002666420406766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,3584,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,768,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,1024,0.006542222367392645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,2560,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,3584,0.01699733402993944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,8192,0.030921777089436848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,768,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,2048,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,1024,0.028427554501427546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,4096,0.019223110543357003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,3584,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,4096,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,10240,0.9750613106621636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,10240,0.023688889212078516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,5120,0.12091378370920818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,2560,0.09995910856458877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,16384,0.5436888800726997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,5120,0.01221422188811832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,1536,0.006215111249023014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,10240,0.017444445027245414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,65536,0.4373679955800374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,256,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,32,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,64,0.00572622236278322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,7168,0.017870222528775532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,2048,0.020847999387317236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,8192,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,12288,0.17983733283148873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,1536,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,1536,0.004017777740955353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,51200,0.1095857752694024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,768,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,51200,0.6084897783067491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,512,0.014370666609870063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,10240,0.02405866649415758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,7168,0.01606666710641649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,768,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,2048,0.21817333168453643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,51200,0.9129848480224609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,1536,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,65536,0.145416882303026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,2560,0.032736000087526106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,3072,0.15719732973310682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,2048,0.019075555933846366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,512,0.03164266546567281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,65536,0.41222310066223145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,8192,0.040997332996792264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,51200,0.4245644410451253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,5120,0.13135377566019693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,8192,0.050670223103629224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,8192,0.7944657537672254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,10240,0.027275555663638647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,2048,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,3072,0.02472977836926778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,128,0.02881866693496704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,1024,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,5120,1.5443252987331813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,1024,0.019486222002241347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,10240,0.09453866879145305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,10240,0.028927998410330877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,51200,1.1661538018120658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,2048,0.012476444244384766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,12288,0.04943111207750109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,256,0.016952888833151925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,512,0.006060444646411472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,1024,0.0218240006102456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,2048,0.025787555509143408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,512,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,32,0.04618488748868307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,8192,0.011501332951916588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,2048,0.08658133612738715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,3584,0.1299440066019694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,512,0.007849777738253275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,3584,0.019289778338538278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,2048,0.0042453333735466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,7168,0.013903111219406128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,2560,0.004868444469239977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,3584,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,128,0.03614133265283372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,768,0.012988444831636218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,8192,0.11796977784898545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,6144,0.029556443293889362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,256,0.022356443934970435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,2048,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,7168,0.031311998764673867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,10240,0.07303910785251193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,256,0.020086222224765353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,10240,0.011193777951929303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,16384,0.061327113045586475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,7168,0.034950223233964704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,3584,0.022103110949198406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,3072,0.09137599998050266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,3584,0.0302817788388994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,3072,0.062423110008239746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,256,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,10240,0.1601911120944553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,512,0.009771555662155151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,768,0.015180443723996481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,16384,0.06707288821538289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,512,0.017431111799346078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,16384,0.10468622048695882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,3584,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,2048,0.1306497785780165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,128,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,7168,0.13240444660186768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,10240,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,65536,0.12314400408003066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,10240,0.015933333171738517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,512,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,32,0.020638222495714825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,32,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,128,0.009531555904282464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,3584,0.00888888869020674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,2048,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,3072,0.008935111264387766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,2560,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,128,0.012058667010731168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,7168,0.022376888328128394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,8192,0.10988089111116196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,128,0.014173333843549093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,16384,0.21093244022793242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,12288,0.05039644572469923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,128,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,256,0.009465778039561378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,1024,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,51200,0.8185626665751139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,4096,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,128,0.07199644380145602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,51200,0.06251555681228638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,2048,0.06438044706980388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,10240,0.037011557155185275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,768,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,1024,0.3760275575849745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,512,0.839646233452691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,8192,0.17086399926079643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,2048,0.01572355628013611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,65536,0.08504800001780193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,16384,0.25123198827107746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,512,0.014455111490355598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,7168,0.014512000812424554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,4096,0.05944266584184435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,65536,0.08582933081520928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,4096,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,6144,0.10578755537668864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,6144,0.11416799492306179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,6144,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,8192,0.16453333695729574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,256,0.005408000200986862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,51200,12.692237006293404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,8192,0.03031022349993388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,8192,0.01369955556260215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,3072,0.016121778223249648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,3584,0.01812177730931176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,51200,0.035162667433420815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,12288,0.07903466622034709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,32,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,2048,0.010821333361996545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,8192,0.016972444123691983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,256,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,10240,0.029139555162853662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,128,0.018583110637134977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,64,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,3072,0.02454311152299245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,6144,0.009162666896979014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,512,0.12808799743652344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,8192,0.04810666706826952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,4096,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,7168,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,16384,0.030863109562132094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,65536,0.4428897698720296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,10240,0.1261004474427965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,4096,0.17821155654059517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,512,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,128,0.018567999203999836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,64,0.001681777752108044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,1024,0.007373332977294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,3072,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,65536,1.0045555962456596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,768,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,256,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,1024,0.020462221569485135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,10240,0.031855109665128924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,10240,0.007176000210973952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,2560,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,512,0.04351377818319532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,5120,0.017616889543003507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,4096,0.01148622234662374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,7168,0.1736888885498047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,2560,0.013064888616402945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,6144,0.03070666723781162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,3584,0.00906844437122345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,65536,0.09136088689168294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,768,0.015001777145597668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,6144,0.04458933406405979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,7168,0.09228977892133926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,1536,0.010956444674068026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,768,0.05807555384106106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,3072,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,2048,0.3639350997077094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,7168,0.1792515516281128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,768,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,7168,0.062225778897603355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,768,0.017291555802027386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,8192,0.14458666907416448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,4096,0.05269333389070299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,65536,0.08612266514036392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,7168,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,10240,0.1018115546968248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,7168,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,65536,0.1280684471130371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,12288,0.014348444011476306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,2560,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,7168,0.021324444148275588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,5120,0.01606666710641649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,32,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,2560,0.02028888960679372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,4096,0.017514665921529133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,6144,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,65536,0.030712889300452337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,512,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,4096,0.09443733427259658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,7168,1.2253608703613281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,7168,0.047522667381498546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,3584,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,2048,0.024115555816226538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,1024,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,8192,0.01089866707722346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,128,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,10240,0.03363200028737386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,8192,0.015954666667514376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,3072,0.13124356004926893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,16384,0.7853351169162326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,12288,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,128,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,51200,0.15590755144755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,65536,1.0786729388766818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,51200,0.6394622060987685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,32,0.018606222338146634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,5120,0.02586666742960612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,10240,0.021127111381954614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,2048,0.050817777713139854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,2560,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,65536,1.4935395982530382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,3584,0.007835555407736037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,1536,0.00740533322095871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,16384,0.01866222255759769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,6144,0.01442311041884952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,1024,0.010219555762079027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,65536,0.32412976688808864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,1536,0.011328889264000786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,8192,0.025253333979182776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,3072,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,51200,1.9829368591308594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,6144,0.02681688964366913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,7168,0.12011199527316624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,10240,0.025467554728190105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,3072,0.010450666977299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,1024,0.02533600065443251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,5120,0.0070266665683852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,7168,0.03798755672242906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,8192,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,256,0.015189333094490899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,128,0.008897778060701158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,32,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,2048,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,6144,0.020871111088328894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,2560,0.010480889015727572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,7168,0.03125955661137899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,6144,0.014070222775141398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,768,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,10240,0.25901688469780815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,51200,0.11763644218444824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,128,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,128,0.015178667174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,12288,0.040256887674331665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,3584,0.42699289321899414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,3584,0.0421653323703342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,6144,0.06109866831037733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,6144,0.1558071109983656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,16384,0.039699554443359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,2560,0.014236445228258768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,7168,0.04782399866316053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,1024,0.02659911082850562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,4096,0.17959911293453643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,16384,0.024322667055659827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,5120,0.46897066964043516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,256,0.037500444385740496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,3584,0.018803555104467604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,3072,0.03473688827620612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,1024,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,2560,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,16384,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,6144,0.05829155445098877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,256,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,5120,0.024047111471494038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,128,0.021129777034123737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,128,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,16384,0.03574577636188931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,5120,0.021819555097156104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,1024,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,4096,0.04671377936999003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,3072,0.026556443836953905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,4096,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,3072,0.0391226675775316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,1536,0.18590577443440756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,16384,0.013601777454217276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,10240,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,65536,1.107269287109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,5120,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,128,0.03735288977622986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,512,0.008230222596062554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,768,0.04301155606905619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,1024,0.012319111161761813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,7168,0.02674666709370083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,2560,0.013960889644092984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,12288,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,5120,0.014914666612943014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,7168,0.03866488734881083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,1024,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,6144,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,16384,0.10416889190673828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,12288,0.05853422482808431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,256,0.0057839999596277876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,8192,0.08129066891140409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,3584,0.043863998519049756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,128,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,128,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,12288,0.01772533357143402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,8192,0.013247110777431063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,2048,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,2560,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,768,0.010607110957304636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,4096,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,2048,0.00638133328821924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,2048,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,2560,0.0092257774538464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,10240,0.012060444388124677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,12288,0.01980177727010515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,4096,0.04970577690336439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,8192,0.021754667162895203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,12288,0.16775288846757677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,10240,0.01165244479974111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,128,0.003569777641031477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,2560,0.076164444287618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,4096,0.02742044462098016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,12288,0.10825155840979682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,3072,0.031026668018764917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,12288,0.02846755584081014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,256,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,1024,0.019031110737058852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,512,0.003823111040724648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,2048,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,1536,0.004951999833186467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,51200,0.07076711124844022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,4096,0.037603557109832764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,1536,0.02012622190846337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,3072,0.03266133202446832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,16384,0.19463645087348092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,8192,0.02256800068749322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,32,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,5120,0.0332915551132626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,51200,0.03799644443723891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,3584,1.1361031002468533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,65536,0.198289778497484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,8192,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,1024,0.03380089004834493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,1024,0.017888888716697693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,2048,0.017495110630989075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,51200,2.5520364973280167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,128,0.03500355614556207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,8192,0.05786044730080498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,10240,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,128,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,10240,0.04288799895180596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,2048,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,3072,0.007387555307812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,128,0.009029332962301042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,10240,1.3397031360202367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,10240,0.020590222544140287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,10240,0.11385511027442084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,2048,0.008918222453859117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,1024,0.020990222692489624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,512,0.01923466722170512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,4096,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,65536,6.854374355740017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,51200,0.05864266554514567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,51200,0.008596444295512306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,256,0.09431911177105373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,512,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,1536,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,256,0.021136000752449036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,3072,0.017866666118303936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,51200,1.4938204023573134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,32,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,10240,0.10780888795852661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,4096,0.01146133326821857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,16384,0.041675554381476514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,32,0.13848711384667292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,64,0.013384000294738345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,6144,0.006872889068391588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,64,0.009495110975371467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,5120,0.0480657782819536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,3072,0.009076444639099969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,3584,0.21252177821265328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,32,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,1024,0.014518222875065274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,8192,0.0147724449634552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,512,0.08241511053509183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,1536,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,16384,0.05992533100975884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,10240,0.09071555402543809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,5120,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,6144,0.03611822260750665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,51200,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,768,0.030827555391523574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,3072,0.038455112112893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,3072,0.01813066667980618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,3072,0.022685334086418152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,51200,1.1211040284898546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,2560,0.026515555050637987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,16384,0.1722266674041748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,256,0.005522666705979242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,256,0.006072888771692912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,6144,0.0530373321639167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,128,0.01925155520439148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,65536,0.0318195554945204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,128,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,3072,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,128,0.0023031110564867654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,4096,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,65536,0.19736266136169434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,10240,0.025942222939597234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,1024,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,3072,0.011900444825490316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,7168,0.04071733355522156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,2048,0.01144977741771274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,6144,0.16991378201378715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,3072,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,32,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,10240,0.05737955702675713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,512,0.010646222366227044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,2048,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,8192,0.0188773340649075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,16384,0.04935200015703837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,2560,0.007190222541491191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,128,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,3072,0.06492711438073052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,32,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,3072,0.2746266788906521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,5120,0.023423110445340473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,12288,0.035066667530271746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,51200,1.9566336737738714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,10240,0.033077332708570696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,256,0.026102221674389307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,8192,0.6426240073310004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,4096,0.004948444250557157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,768,0.011930666863918304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,4096,0.05872178077697754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,4096,0.020706666840447318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,5120,0.010097777677906884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,51200,0.08206044303046332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,128,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,128,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,1024,0.01722311145729489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,51200,0.08563733100891113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,10240,0.02215733296341366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,10240,0.16239999400244817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,5120,0.05855111281077067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,64,0.0019422223170598347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,512,0.007243555453088548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,7168,0.04582311047448052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,128,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,512,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,768,0.23011912239922416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,65536,0.3084248966640896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,2048,0.011329777538776398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,3072,0.06804533137215509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,64,0.0018648888087934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,2048,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,5120,0.04703199863433838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,6144,0.03154666556252374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,2048,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,2560,0.14674577448103163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,16384,0.0322364436255561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,4096,0.012950222525331708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,128,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,5120,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,1536,0.5612906879848903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,12288,0.10279289219114517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,3584,0.02909066610866123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,10240,0.03842488924662272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,16384,0.0236835562520557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,10240,0.4609537654452854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,7168,0.012050666742854647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,12288,0.01479111115137736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,12288,0.08221866687138875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,1536,0.052595555782318115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,8192,0.030043555630577937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,6144,0.05601333247290718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,65536,0.056185775332980685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,3072,0.04292622208595276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,2560,0.025037333369255066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,2048,0.020992888344658747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,128,0.01313777764638265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,6144,0.010951999988820819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,12288,0.019641778535313077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,768,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,51200,0.06756622261471219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,768,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,4096,0.020963556236690942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,768,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,1024,0.020582222276263766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,256,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,8192,0.045088890526029796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,512,0.015563555889659457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,51200,0.7498826450771756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,51200,0.05225155419773526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,32,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,4096,0.008208889100286696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,2048,0.019270223047998216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,2560,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,256,0.006903999795516332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,128,0.02462133268515269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,512,0.034381333324644305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,16384,0.39351465966966415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,6144,0.009909333454238044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,3072,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,512,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,2560,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,3584,0.06846755743026733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,65536,0.34520533349778915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,256,0.026886221435334947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,128,0.0185217774576611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,512,0.012893333203262754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,6144,0.0384168889787462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,768,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,10240,0.05345511105325487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,2048,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,7168,0.03595111105177138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,1024,0.014319999350441827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,6144,0.01612622208065457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,5120,0.052617778380711876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,5120,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,4096,0.02641244398223029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,65536,0.8194275432162814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,8192,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,2048,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,7168,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,7168,0.0889342228571574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,256,0.026239110363854304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,1536,0.03738311264250014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,5120,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,256,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,6144,0.0448035564687517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,1024,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,65536,0.1604106691148546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,128,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,1024,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,4096,0.03858310977617899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,1536,0.00778399987353219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,16384,0.026087111896938745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,16384,0.03411022159788344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,10240,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,768,0.016587555408477783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,64,0.0019431112127171622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,256,0.005385777602593104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,3584,0.011015999648306104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,2048,0.061999109056260854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,3072,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,3072,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,3072,0.015250666273964776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,768,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,768,0.03437422381507026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,16384,0.0361768901348114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,256,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,6144,0.08446577522489761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,10240,0.057709336280822754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,6144,0.22098667091793486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,3584,0.04483200112978617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,768,0.014555555250909595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,12288,0.10285244385401408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,768,0.00535111129283905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,64,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,1024,0.2408693366580539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,5120,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,10240,0.04373244444529215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,5120,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,8192,0.05013333426581489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,16384,0.1811973386340671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,3072,0.05064800050523546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,8192,0.23186667760213217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,3072,0.015677332878112793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,8192,0.05324177940686544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,51200,0.1751635604434543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,12288,0.19120444191826713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,16384,0.1359297831853231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,5120,0.07777422004275851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,256,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,3072,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,256,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,3072,0.00721066693464915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,768,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,7168,0.47977956136067706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,2560,0.028412444723976985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,1536,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,10240,0.05882044633229574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,8192,0.04814666509628296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,3584,0.03395288851526048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,10240,0.018639999959203932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,1536,0.0038328886859946777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,51200,0.8156951268513998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,5120,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,2048,0.04756888747215271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,3072,0.035248888863457575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,2560,0.010314666562610203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,512,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,10240,0.025232889586024817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,512,0.011192888849311404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,51200,0.1610435512330797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,10240,0.015096000499195524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,10240,0.17361333635118273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,10240,0.021359999974568684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,7168,0.0248364441924625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,7168,0.15158843994140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,51200,0.2709146605597602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,10240,0.024696888195143804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,1536,0.02557777696185642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,2048,0.033328887489106916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,7168,0.008610666626029545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,32,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,10240,0.02568711174858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,2560,0.009641778137948778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,128,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,1024,0.02165866725974613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,4096,0.011918221910794577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,12288,0.017370666066805523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,51200,0.12239200539059109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,7168,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,8192,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,2560,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,4096,0.1443635490205553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,128,0.024000888069470722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,12288,0.18386133511861166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,8192,0.02798844377199809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,4096,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,3584,0.04626577761438158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,6144,0.3826337655385335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,1536,0.04550844430923462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,16384,0.02146311104297638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,4096,0.031347556246651545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,16384,0.04906222224235535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,256,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,51200,0.18181778324974907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,12288,0.02568088968594869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,10240,0.04501333170466953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,7168,0.03540622194608053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,2560,0.02269866731431749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,4096,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,4096,0.010778666370444827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,10240,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,1024,0.0057546666098965546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,2560,0.010175111393133799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,64,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,7168,0.04622755448023478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,51200,0.04998488889800178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,1024,0.06260977851019965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,6144,0.0718053314420912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,3584,0.009816889133718278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,6144,0.9457769393920898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,3584,0.09353688690397476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,256,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,7168,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,3072,0.22171644369761148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,128,0.018175111876593698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,768,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,12288,0.03677777780426873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,768,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,1024,0.03640710976388719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,2560,0.032046222024493747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,8192,0.03403822249836392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,256,0.015853333804342482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,6144,0.05725510915120443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,8192,0.7453973558213977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,2048,0.02125866711139679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,65536,0.9371555116441516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,7168,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,512,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,3072,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,3584,0.017645332548353408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,6144,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,512,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,16384,0.7902248700459799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,4096,0.12414933575524224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,128,0.008159110943476358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,1536,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,65536,0.06828088892830743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,51200,0.06150844362046984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,8192,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,512,0.051939556996027626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,51200,0.4300826655493842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,5120,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,32,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,4096,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,5120,0.05075733198059929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,2560,0.03993955585691664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,2560,0.05834844377305773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,2560,0.05952178107367622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,512,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,3584,0.023410666320059035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,8192,0.09510133663813274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,6144,0.03675555520587497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,51200,1.4359031253390844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,10240,0.03778310947948032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,10240,0.014502222339312235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,128,0.04307466745376587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,3072,0.013774221969975365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,6144,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,3072,0.03801155421468947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,12288,0.014392000105645923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,3072,0.013191999660597907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,512,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,1536,0.0032257777121331957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,51200,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,256,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,1024,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,10240,0.3081555631425646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,1024,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,6144,0.06979199912812975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,65536,0.17447733879089355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,5120,0.052279111411836415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,512,0.3510453436109755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,12288,0.19883555836147734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,512,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,65536,0.1297244495815701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,2048,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,4096,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,65536,1.0373901791042752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,51200,0.09482400284873115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,256,0.022261333134439256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,256,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,6144,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,1024,0.008434666527642144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,4096,0.015303111738628812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,3072,0.12733244895935059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,10240,0.13375022676255968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,3584,1.0735609266493056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,4096,0.009275555610656738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,10240,0.0584684411684672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,5120,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,1536,0.014025777578353882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,2048,0.012737777498033313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,12288,0.034169776572121516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,5120,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,32,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,64,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,8192,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,4096,0.035102221700880266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,7168,0.027923554182052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,32,0.0028880000528362063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,2560,0.15154755115509033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,65536,0.06368977493709989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,1536,0.0037600000699361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,4096,0.030584888325797185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,51200,0.9844231075710721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,7168,0.02464177707831065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,4096,0.2737884521484375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,32,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,51200,0.034175998634762235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,12288,0.01953866746690538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,256,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,1024,0.04489599996142917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,4096,0.00887288898229599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,6144,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,3584,0.019094222121768527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,3072,0.006010666489601135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,32,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,10240,0.5429946581522623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,3072,0.3101289007398817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,12288,0.3721288839975993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,16384,0.1291688945558336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,16384,0.022946667340066697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,2048,0.020990222692489624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,256,0.014533332652515836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,6144,0.028010666370391846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,51200,0.1254311137729221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,5120,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,3584,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,51200,0.6036266750759548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,256,0.005533333453867171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,6144,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,10240,0.04332888788647122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,2560,0.08994844224717881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,65536,0.0652373300658332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,10240,0.031207111146714952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,5120,0.00462400002612008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,5120,0.09819377793206109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,3072,0.02549955579969618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,4096,0.05021689004368252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,768,0.01372977760103014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,65536,0.1816693279478285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,7168,0.015969778100649517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,51200,0.25803645451863605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,32,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,4096,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,1024,0.1127199994193183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,7168,0.038431111309263445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,1024,0.011869333684444427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,3584,0.02478222217824724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,6144,0.03542222248183356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,10240,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,8192,0.09486933549245198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,2560,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,6144,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,3072,0.010271111296282874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,16384,0.12183733781178792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,64,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,1536,0.08756533596250747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,65536,0.9507706960042318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,7168,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,3072,0.009487999810112847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,256,0.022857778602176245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,6144,0.15903022554185656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,2048,0.0042257776690853965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,8192,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,8192,0.35564178890652126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,256,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,10240,0.02312622302108341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,3584,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,6144,0.03776266508632236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,2048,0.046570665306515165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,768,0.006808888994985157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,64,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,1024,0.018530666828155518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,64,0.07465510898166232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,256,0.019606222709019978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,6144,0.02836444311671787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,768,0.06527822547488742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,2048,0.015063110325071545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,10240,0.24081601036919487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,1024,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,16384,0.6202275488111708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,3072,0.1991164419386122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,2048,0.032701333363850914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,2560,0.07956622044245402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,1536,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,10240,0.15248000621795654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,10240,0.15614755948384604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,51200,1.1682693693372939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,4096,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,4096,0.029676444000667993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,8192,0.006134222365087933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,256,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,128,0.027783112393485174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,2560,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,128,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,512,0.017141333884663053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,2048,0.015403555499182807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,4096,0.01407911048995124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,32,0.005048888838953442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,3072,0.025278222229745653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,3072,0.010995555255148145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,10240,0.6633342107137045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,1024,0.010475555227862464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,3072,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,65536,0.08243378003438313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,65536,0.15339377191331652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,51200,0.05358577768007914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,768,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,6144,0.0244159996509552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,512,0.4456177817450629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,3584,0.0878302190038893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,2048,0.06335910823610094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,3584,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,5120,0.062133332093556724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,4096,0.03953333364592658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,10240,0.03837777839766608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,51200,27.442867702907986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,65536,0.22625244988335502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,8192,0.02105866703722212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,8192,0.03814488980505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,12288,0.07220266924964057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,32,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,768,0.1706248919169108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,1024,0.021511110994550917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,128,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,6144,0.027806222438812256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,65536,0.15944000085194907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,10240,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,64,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,16384,0.4435315661960178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,5120,0.1282071140077379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,32,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,1536,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,2560,0.014408000641398959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,7168,0.03733866744571262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,3072,0.008147555920812819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,51200,0.8345688713921441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,51200,0.07576000028186373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,12288,0.05852533049053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,256,0.10409600204891628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,128,0.08484177456961738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,4096,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,512,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,1024,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,12288,0.021495110458797876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,768,0.008360000120268928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,128,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,65536,0.18925600581698945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,7168,0.062084443039364286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,5120,0.2104648881488376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,4096,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,2560,0.008079110748238033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,4096,0.026226666238572862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,7168,0.043255110581715904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,3584,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,128,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,1536,0.006131555471155379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,10240,0.09886755545934041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,2560,0.011553777588738335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,768,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,1024,0.0181022220187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,8192,0.010871999793582492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,6144,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,12288,0.02792533238728841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,128,0.003560888684458203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,3584,0.04589510957400004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,65536,1.3807377285427516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,2048,0.0314702226055993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,12288,3.008454216851128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,51200,0.19704800181918672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,4096,0.05081511206097073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,2048,0.013633777697881063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,512,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,2048,0.10861155721876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,16384,0.02161600026819441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,6144,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,4096,0.0045422220395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,64,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,8192,0.04862577716509501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,8192,0.37443110677931046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,3584,0.5054346720377604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,65536,0.6334986686706543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,5120,0.033153778976864286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,1536,0.008351999852392409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,8192,0.5109466446770562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,6144,0.07795644468731351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,1024,0.01646933290693495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,3584,0.1124533282385932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,768,0.03294310967127482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,5120,0.06696622239218818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,768,0.13154044416215685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,32,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,10240,0.9114213519626193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,16384,0.29004356596204967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,10240,0.01756622228357527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,128,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,51200,0.3750035497877333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,2048,0.02052533295419481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,2048,0.03410133388307359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,6144,0.03852977686458164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,12288,0.18951111369662813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,6144,0.08283199866612752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,2048,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,1536,0.01790577835506863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,1536,0.02722399930159251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,8192,0.023806222611003455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,51200,0.23976622687445748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,6144,0.29434577624003094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,4096,0.03153866529464722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,4096,0.02176355487770504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,16384,0.0336079994837443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,3072,0.015982222225930955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,5120,0.028107557031843398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,1024,0.03625599874390496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,3072,0.09586222304238214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,2048,0.005103999955786599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,256,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,6144,0.032031112247043185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,51200,0.049887110789616905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,512,0.0057555557125144536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,3584,0.13078399499257407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,2048,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,51200,0.0986444420284695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,5120,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,6144,0.025176000263955858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,1024,0.014576888746685453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,16384,0.17753511004977754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,768,0.02163377735349867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,1536,0.029134223858515423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,1536,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,4096,0.08812710973951553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,4096,0.27651998731825084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,4096,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,1024,0.018019555343521964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,3584,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,2048,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,51200,1.3757573233710394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,12288,0.026542221506436665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,2048,0.6330675548977321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,5120,0.03687466515435113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,2048,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,12288,0.24509689542982313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,256,0.003644444462325838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,512,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,2048,0.015975111060672335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,1024,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,12288,0.05367111166318258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,1536,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,1536,0.011207111179828644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,512,0.01620444489849938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,2560,0.06841333044899835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,2048,0.016846223009957206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,768,0.015612444943851896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,128,0.0240639994541804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,128,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,128,0.18196800020005968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,2560,0.012835555606418185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,3584,0.2689182228512234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,5120,0.048038221067852445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,256,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,7168,0.045681777927610606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,768,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,12288,0.020264888803164165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,8192,0.018248000078731112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,2560,0.007113777928882175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,6144,0.03207733233769735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,5120,0.12716889381408691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,1024,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,6144,0.13178843922085232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,5120,0.06257955233256023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,64,0.015074667003419666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,5120,0.01922044489118788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,2560,0.11755822764502631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,6144,0.03228088882234361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,10240,0.042281776666641235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,3584,0.028182221783532038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,256,0.05044000016318428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,3072,0.12631200419531927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,6144,0.1669600009918213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,5120,0.03695822093221877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,10240,0.052355554368760854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,2048,0.028819554381900366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,3072,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,256,0.055065777566697865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,512,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,5120,0.036222222778532244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,2048,0.5412915547688802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,7168,0.010687111152542962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,12288,0.02983022067281935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,8192,0.023963555693626404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,7168,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,51200,0.7366479767693414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,128,0.008782222039169734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,2560,0.026212443908055622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,1536,0.03221155537499322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,51200,0.09812621937857734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,4096,0.031039112144046362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,6144,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,6144,0.00981155534585317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,65536,0.06456266509162055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,3584,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,16384,1.498112890455458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,2048,0.036165333456463285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,512,0.003632889025741153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,65536,1.8970027499728734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,512,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,7168,0.17217600345611572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,64,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,256,0.00962222201956643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,5120,0.03746133380466037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,12288,2.1669084760877824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,64,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,4096,0.018038221531444125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,32,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,16384,0.06613066461351183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,12288,0.15818044874403211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,65536,0.06319200330310397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,8192,0.03746577766206529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,2048,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,6144,0.02849422229660882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,7168,0.8386649025811089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,1536,0.005172444300519096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,10240,0.01443288889196184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,12288,0.012260444462299347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,768,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,51200,0.18748444981045195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,8192,0.9147226545545789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,8192,1.2357181972927518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,3584,0.01903288894229465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,5120,0.02025422288311852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,512,0.00351200004418691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,2560,0.05691644218232897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,3584,0.03503555390569899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,51200,0.2842826578352186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,768,0.01961688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,1536,0.032827556133270264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,3072,0.11029599772559272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,4096,0.09563999705844456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,65536,0.1357413397894965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,6144,0.03017066584693061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,10240,0.024693333440356787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,6144,0.019201777047581144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,1024,0.022259554929203455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,128,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,10240,0.09735644525951809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,4096,0.011120888921949597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,4096,0.015066666735543145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,10240,0.10727644628948635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,512,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,2560,0.6693253517150879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,1024,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,256,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,32,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,65536,0.746614244249132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,1536,0.01645955608950721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,128,0.012058667010731168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,10240,0.05102755626042684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,1024,0.012667555775907306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,12288,0.027261333333121404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,1024,0.014175110393100314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,8192,0.8560631010267469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,2048,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,3584,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,128,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,2560,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,1024,0.07041600015428331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,512,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,12288,0.08281155427296956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,7168,0.06491021977530585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,7168,0.18004443910386828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,5120,0.029652443197038438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,4096,0.018203554881943595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,12288,0.01648977730009291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,1536,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,12288,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,65536,0.06281155347824097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,3584,0.039467556609047785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,8192,0.07050310903125338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,16384,0.03386133246951633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,3072,0.12042844295501709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,12288,0.02088177700837453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,768,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,768,0.0031439999325407874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,65536,0.22083644072214761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,512,0.01420088940196567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,16384,2.746424992879232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,1536,0.010282666318946415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,8192,0.08773333496517605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,1024,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,5120,0.04696711235576206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,2560,0.040891554620530866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,32,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,8192,0.048130667871899076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,10240,0.03619733452796936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,7168,0.013560888667901358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,12288,0.018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,16384,0.16968088679843477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,3584,0.01903022163444095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,256,0.00489777781897121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,8192,0.020248888267411124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,1536,0.015903110305468243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,16384,0.23125955793592665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,10240,0.03643111056751675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,8192,0.04092355569203695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,1536,0.03863733344607883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,1024,0.04020177655749851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,65536,0.080530669954088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,1024,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,3072,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,128,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,4096,0.0233297778500451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,3072,0.2862959967719184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,4096,0.02423733307255639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,512,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,8192,0.15578756067487928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,8192,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,65536,0.1428417762120565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,2560,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,32,0.012115555504957834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,512,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,256,0.027505778604083594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,65536,0.028756442997190688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,5120,0.7124364640977648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,7168,0.12690756056043836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,1024,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,65536,0.1276924477683173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,2048,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,5120,0.011312888728247749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,16384,0.019446222318543326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,4096,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,16384,0.02613777750068241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,51200,7.113055759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,7168,0.015410666664441427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,10240,0.14209689034356013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,7168,0.03790488839149475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,5120,0.047192000680499606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,2560,0.6976915465460883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,12288,0.9999057981703016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,16384,0.04177066683769226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,2048,0.10181866751776801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,7168,0.012574222352769641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,64,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,16384,0.07404266463385688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,65536,0.08436089091830784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,4096,0.04055466585689121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,128,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,2048,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,16384,0.005449777675999536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,5120,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,1024,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,2048,0.007091555330488417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,2048,0.01428533262676663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,5120,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,3072,0.00755466686354743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,6144,0.044172445933024086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,16384,0.04485244552294413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,32,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,51200,0.13242222203148737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,2560,0.027267555395762127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,4096,0.020975111259354487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,128,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,4096,0.12104088730282253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,12288,0.030962665875752766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,4096,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,8192,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,16384,5.247234768337674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,1536,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,3584,0.05959288941489326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,4096,0.022544888986481562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,65536,0.09523289071189033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,51200,0.5476737552218968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,3072,0.024384000235133704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,6144,0.012359999948077731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,128,0.009565333525339762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,12288,0.8801120122273763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,6144,0.015026667051845126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,6144,0.09586311048931545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,4096,0.010835555692513784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,3584,0.019152000546455383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,3072,0.3429715633392334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,51200,0.12044533093770345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,3072,0.0737511118253072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,64,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,5120,0.02847111225128174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,1536,0.028377778000301782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,5120,0.011153777440388998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,3584,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,7168,0.052200890249676175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,64,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,768,0.006069333189063602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,51200,0.02444355520937178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,6144,0.1297573381000095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,8192,0.043720000320010714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,12288,0.050350222322675914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,256,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,128,0.16234755516052246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,10240,0.10828889078564113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,64,0.004176889028814104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,2048,3.4608514573838978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,5120,0.29962311850653756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,16384,0.018815999229749043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,2560,0.04740088846948412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,2048,0.040656887822681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,10240,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,7168,0.01567555632856157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,7168,0.01443288889196184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,10240,0.05443111062049866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,4096,0.8077386750115289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,2560,0.02202933364444309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,65536,0.10035910871293809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,768,0.014577777849303352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,512,0.018207111292415194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,32,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,16384,0.052589333719677396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,6144,0.018231999542978074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,51200,0.05379288726382785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,51200,0.054696887731552124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,3584,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,2560,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,256,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,10240,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,65536,0.0710480014483134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,64,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,2048,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,12288,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,768,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,51200,0.22516976462470162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,7168,0.01772444446881612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,2560,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,256,0.03598399957021078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,10240,0.026800000005298193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,51200,0.05062222149637011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,256,0.030573334958818223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,10240,0.016571554872724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,4096,0.10231022040049235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,2560,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,8192,0.022423999177085027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,51200,0.1373902161916097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,64,0.0027546667390399506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,512,0.04695022106170654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,768,0.03245688809288873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,256,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,1536,0.030551112360424463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,768,0.017110221915774874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,2560,0.012958221965365939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,1536,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,512,0.05511022276348538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,16384,0.02240266733699375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,12288,0.053808887799580894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,10240,0.03241511185963949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,16384,0.055445333321889244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,8192,0.018980445133315194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,128,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,4096,0.31159289677937824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,6144,0.06647555695639716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,1024,0.014470222923490735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,2048,0.012984000146389008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,6144,0.07300444444020589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,7168,0.009884444375832876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,2048,0.0053004444473319584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,8192,0.013903111219406128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,10240,0.16174844900767008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,2560,0.01054133309258355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,65536,1.1812221739027235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,7168,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,5120,0.009487999810112847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,2048,0.014311111635631986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,3072,0.0425582230091095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,128,0.006851555572615729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,4096,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,10240,0.09918755292892456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,16384,0.055234667327668935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,4096,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,1536,0.01697422232892778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,768,0.2118071052763197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,5120,0.9743004904852973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,256,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,8192,0.011761777930789523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,51200,0.22889600859747994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,16384,0.026583110292752583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,768,0.019552888141738046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,3072,0.026720000637902155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,3584,0.041420443190468684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,8192,0.09532355599933201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,32,0.00870222184393141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,1536,0.036343110932244196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,5120,0.1808800035052829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,3584,0.03065155612097846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,6144,0.024360888534122046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,5120,0.022281777527597215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,1024,0.02517333295610216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,3072,0.02237155536810557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,51200,0.5935182041592068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,1024,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,10240,0.5748062133789062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,12288,0.06813600328233507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,10240,0.9337031046549479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,16384,0.06292444467544556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,64,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,12288,0.07413778040144178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,16384,0.3878933323754205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,8192,0.02920444475279914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,5120,0.012395555774370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,3584,0.07531733645333184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,3584,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,2560,0.007137777904669444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,10240,0.020928000410397846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,4096,0.6405715412563747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,16384,0.28015467855665416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,32,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,12288,0.04608800013860067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,4096,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,256,0.051870223548677236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,768,0.03537333342764113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,128,0.008781332936551835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,65536,0.0625235570801629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,10240,0.01754666699303521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,12288,0.06020977762010363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,6144,0.2694808906979031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,5120,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,4096,0.011755555868148804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,32,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,6144,0.18263999621073404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,128,0.0517119997077518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,32,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,1024,0.023064000738991633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,12288,0.036666668123669095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,4096,0.02711822258101569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,512,0.006131555471155379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,6144,0.12468532721201579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,768,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,10240,0.06463555494944255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,256,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,8192,1.4480728573269313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,6144,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,51200,0.06514844629499647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,7168,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,768,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,65536,0.11754843923780654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,4096,0.1865075561735365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,1536,0.01425244410832723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,1024,0.0229102224111557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,64,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,2048,0.022653333014912073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,3584,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,64,0.01185155577129788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,51200,0.9086479610866971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,2560,0.0751546687550015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,1536,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,4096,0.011130666567219628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,128,0.025348444779713947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,4096,0.014088888963063559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,768,0.009422221945391761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,5120,0.027452444036801655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,65536,0.0710115565194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,5120,0.2662693394554986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,768,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,3584,0.629627545674642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,3072,0.043990221288469106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,2048,0.013079110946920184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,768,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,2048,0.004544888933499654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,1536,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,3584,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,768,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,2560,0.046932445632086865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,5120,0.12703466415405273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,51200,2.9455786810980906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,1536,0.0548017786608802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,128,0.0732933349079556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,16384,0.10931199789047241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,3584,0.024283554818895128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,128,0.01908622185389201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,512,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,5120,0.015064888530307345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,1024,0.008353778057628209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,4096,0.026191999514897663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,6144,0.10708355241351658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,128,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,2048,0.28965867890252006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,3072,0.045369777414533824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,16384,0.30734043651156956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,8192,0.019048889478047688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,5120,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,1024,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,8192,0.035939554373423256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,2560,0.025514667232831318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,16384,0.01427288850148519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,3072,0.030551112360424463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,12288,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,7168,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,16384,0.10497866736518012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,2560,0.11046489079793294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,512,0.13809511396620008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,256,0.03051999873585171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,4096,0.008099555969238281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,3584,1.3956231011284723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,7168,0.06374666425916883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,51200,0.26319199138217503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,256,0.033568888902664185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,51200,0.04736977815628052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,12288,0.021420444051424663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,512,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,1024,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,3072,0.014542222023010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,2560,0.03360711203681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,128,0.007702222300900354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,16384,0.0239982224173016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,65536,0.05032711227734884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,1024,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,65536,0.025128000312381323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,512,0.055905779202779136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,512,0.01348355578051673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,65536,0.3471208943261041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,512,0.010202666951550378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,512,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,1024,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,2560,0.025096888343493145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,7168,0.020632889535692003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,256,0.003028444531891081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,12288,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,3072,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,3072,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,4096,0.026736888620588515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,6144,0.1290880044301351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,4096,0.27790310647752553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,512,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,512,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,8192,0.030200888713200886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,4096,0.023997333314683702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,1536,0.011904000408119626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,16384,0.19234666559431288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,3072,0.20828444427914092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,51200,0.663973331451416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,1024,0.009913778139485253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,4096,0.008789333204428354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,3584,0.025303110480308533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,3072,0.012706666356987424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,3584,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,6144,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,512,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,10240,0.012850667039553324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,3584,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,8192,0.03863022062513564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,8192,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,256,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,8192,0.033407111962636314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,3584,0.031142221556769475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,8192,0.044509334696663745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,256,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,2560,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,2560,0.13099643919203016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,768,0.0120000003112687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,8192,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,8192,0.09686222341325547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,2560,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,12288,0.03723911113209195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,768,0.01927288870016734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,3072,0.015927111109097798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,5120,0.7998773256937662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,8192,0.010796444283591377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,1536,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,1536,0.007409777906205919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,1536,0.01761244402991401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,10240,0.015916445189052157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,1024,0.017017778423097398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,768,0.0032222221295038858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,256,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,2048,0.03313866588804457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,10240,0.03197066651450263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,768,0.03126488791571723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,16384,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,3584,0.0353724459807078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,1024,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,10240,0.017145777742067974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,12288,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,32,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,768,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,7168,0.03578133384386698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,10240,0.06749511427349515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,2048,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,51200,0.11493066946665446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,1024,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,4096,0.10861244466569688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,2048,0.02531822191344367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,3072,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,3584,0.042860445049073964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,128,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,7168,0.051797333690855235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,10240,0.022713777091768052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,12288,0.013473778135246701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,512,0.009927111367384592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,12288,2.449272791544596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,2560,0.03189955486191644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,6144,0.09409777985678779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,1536,0.36743466059366864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,1024,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,7168,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,65536,0.0195448895295461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,512,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,10240,0.014683556225564746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,3584,0.03340977761480544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,12288,0.008606221940782335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,3584,0.025366221865018208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,512,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,7168,0.3651164372762044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,16384,0.2380666732788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,2048,0.004549333204825719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,4096,0.35825154516432023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,8192,0.011404444773991903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,5120,0.010255110760529837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,2560,0.3465733263227675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,768,0.019292443990707397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,512,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,4096,0.010991999672518836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,768,0.056555555926428906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,512,0.0028942223224374983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,51200,2.5088702307807074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,512,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,7168,0.1899120012919108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,2560,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,12288,0.018426666657129925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,65536,7.607938978407119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,6144,0.022291556000709534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,2048,0.014120888378885059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,2048,0.015746666325463187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,2048,0.009443555441167619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,8192,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,6144,0.8534400198194715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,768,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,6144,0.019744000501102872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,1536,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,64,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,5120,0.014542222023010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,2048,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,512,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,3072,0.023191110955344305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,51200,0.07855111360549927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,768,0.007432888779375289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,12288,1.2274675369262695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,2560,0.04193244377772013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,51200,0.11304444736904568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,51200,0.10817688703536987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,128,0.021282666259341772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,5120,0.016069332758585613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,128,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,2560,0.005713777823580637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,7168,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,10240,0.01758577757411533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,2560,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,7168,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,1024,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,16384,0.13356533315446642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,3584,0.1506684488720364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,10240,0.014509333504570855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,10240,0.020592888196309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,3072,0.07471466726726957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,6144,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,16384,0.11901244852277969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,10240,0.21372800403171113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,128,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,65536,0.639293352762858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,1536,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,65536,0.3308853308359782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,768,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,2560,0.03767911261982388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,3072,0.027801778581407335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,16384,0.08767555819617377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,256,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,2048,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,6144,0.01262399968173769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,16384,0.036826666858461164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,4096,0.01773866679933336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,65536,0.2823546727498372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,8192,0.06762666834725274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,1536,0.009108444054921469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,1024,0.02661955522166358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,512,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,768,0.004559999952713649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,8192,0.1593288845486111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,4096,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,51200,0.10016800297631158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,5120,0.09654844469494289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,2560,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,3072,0.025624889466497634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,3584,0.049207998646630176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,6144,0.01945955554644267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,8192,0.07177155547671847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,4096,0.027653333213594224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,1536,0.01884977850649092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,768,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,51200,0.1705111132727729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,2048,0.004945777770545748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,1024,0.009418666362762451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,5120,0.013136000268989138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,32,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,6144,0.026084444589085046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,2048,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,512,0.01518311103185018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,65536,0.6228444311353896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,2560,0.022667555345429316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,6144,1.3889741897583008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,3072,0.025407999753952026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,2048,0.00684533350997501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,65536,1.125107553270128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,128,0.016968000266287062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,2560,0.028432889117134943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,7168,1.267525355021159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,7168,0.023688889212078516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,3072,0.05669244130452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,256,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,3584,0.49823734495374894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,5120,0.09739910893970066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,3072,0.04428889022933113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,6144,0.050071110328038536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,512,0.024403555525673762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,128,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,8192,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,5120,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,32,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,5120,0.2893795437282986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,3072,0.07051377826266818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,65536,0.06177866458892822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,8192,0.6924666828579373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,65536,0.3436933358510335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,6144,0.10285066896014744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,7168,0.02405422263675266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,65536,13.501688639322916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,51200,0.16755999459160698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,256,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,32,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,2048,0.028113779094484117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,1024,0.0064586665895250105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,3072,0.030600888861550227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,512,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,7168,0.024891555309295654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,10240,0.13655911551581487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,51200,0.12282400661044651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,256,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,12288,0.08505600028567845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,1024,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,65536,1.452262242635091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,3584,0.016024000114864774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,128,0.014889778362380134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,512,0.033850666549470686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,12288,0.023407111565272015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,7168,0.28467199537489146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,51200,0.06273155742221408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,1536,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,65536,0.1457137796613905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,12288,0.5195902188618978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,768,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,128,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,5120,0.13950667116377088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,51200,0.06761688656277128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,8192,0.563194645775689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,1024,0.06697511010699801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,128,0.0022408889813555614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,768,0.028380443652470905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,6144,0.03680355681313409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,128,0.007970666719807519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,512,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,2048,0.016951999730534024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,5120,0.0779706637064616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,128,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,5120,0.015432889262835184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,16384,0.03040177623430888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,7168,0.034196443027920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,6144,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,8192,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,65536,2.255594677395291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,51200,0.06561511092715792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,16384,0.7199315494961209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,5120,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,12288,0.06535022126303779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,51200,0.42201956113179523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,2560,0.00757422215408749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,12288,0.19599466853671602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,64,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,3584,0.014559111661381192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,8192,0.011376000112957425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,65536,0.05040977729691399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,128,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,51200,0.05303111010127597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,8192,0.11012088590198094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,51200,0.7364320225185819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,2048,0.021980444590250652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,7168,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,2560,0.01055111073785358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,5120,0.009783999787436591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,12288,0.08687021997239854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,12288,0.03874489002757602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,65536,0.1289680004119873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,2560,0.029109332296583388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,1024,0.015086222026083203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,128,0.01573422220018175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,768,0.005839111076460944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,12288,1.4237715403238933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,256,0.020571556356218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,51200,0.21399466196695963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,2048,1.0405137803819444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,5120,0.013079110946920184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,32,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,16384,0.2411359945933024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,2048,0.17532177766164145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,128,0.020233778489960563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,3584,0.03272977802488539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,8192,0.07059199942482842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,51200,0.12996355692545572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,3584,0.974289788140191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,65536,0.056906667020585805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,51200,0.24541245566474068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,1536,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,2048,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,65536,0.23400444454616973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,1024,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,1536,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,4096,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,12288,0.04519822200139364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,2048,0.0295688874191708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,6144,0.02141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,256,0.005764444255166584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,512,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,3584,0.05284888876809014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,1024,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,3584,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,2048,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,51200,0.22985956403944227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,768,0.018587556150224473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,768,0.0851866669125027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,3072,0.028472890456517536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,51200,1.0782960255940754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,2048,0.013224889006879596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,4096,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,3584,0.016812443733215332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,512,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,65536,0.04703999890221489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,512,0.0032319999817344877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,12288,0.052947554323408336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,7168,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,2048,0.014456889695591398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,51200,0.11761600441402859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,51200,0.1112568908267551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,512,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,7168,0.052410665485594005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,64,0.004389333228270213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,2560,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,7168,0.042540444268120654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,12288,0.1564693318472968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,8192,0.05016977919472588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,2048,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,32,0.013358222113715278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,128,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,65536,0.18800621562533906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,2560,0.010440000229411654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,7168,0.01425155500570933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,1024,0.020488889680968392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,16384,0.07435555590523614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,2560,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,2560,0.01860000027550591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,5120,0.04298311140802172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,10240,0.033039109574423894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,2560,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,3584,0.024119110571013555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,10240,0.02103022237618764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,2048,0.012848888834317526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,7168,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,64,0.0018622221218215094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,2048,0.004583999928500918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,51200,0.49898846944173175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,6144,0.024903999434577093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,768,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,6144,0.012582222620646158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,7168,0.1567351155810886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,16384,0.021021333005693223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,7168,0.005445333404673471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,1536,0.0057760001056724125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,768,0.024382222029897902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,3584,0.008105778031879002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,7168,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,128,0.051967998345692955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,2048,0.16753333144717744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,10240,0.08468266990449692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,65536,0.1939528915617201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,16384,0.17286400000254312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,12288,0.036329779360029436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,51200,0.4770968755086263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,65536,0.12982577747768825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,4096,0.04009955459170871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,3072,0.0594773358768887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,10240,0.06537422206666735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,10240,0.011115555961926779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,51200,0.7307857937282987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,5120,0.012321777641773224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,3584,0.010086221827401055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,1536,0.03247733248604669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,7168,0.03636711173587375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,512,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,2048,0.025490666429201763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,1024,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,1024,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,3584,0.03451910946104262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,128,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,3072,0.05900888972812229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,5120,0.7748088836669922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,8192,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,8192,0.0210533340771993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,1536,0.02252799934811062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,12288,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,4096,0.016524444023768108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,7168,0.7667368782891167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,5120,0.015786666009161208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,2560,0.010299555957317352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,16384,0.047545777426825635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,32,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,3072,0.01607377827167511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,1536,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,10240,0.06724355618158977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,5120,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,2560,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,6144,0.03053599927160475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,768,0.03012088934580485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,2560,0.011553777588738335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,10240,0.03200177682770623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,65536,0.07230133480495878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,10240,0.045476443237728544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,128,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,3584,0.07560266388787164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,7168,0.018422222799725003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,10240,0.005986666513813867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,1536,0.01239999963177575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,8192,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,768,0.0976782242457072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,7168,0.08275110854042901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,3072,0.035007112556033664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,2048,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,3072,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,1024,0.02212000058756934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,128,0.008595555192894405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,5120,0.018053332964579265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,3584,0.021645334031846788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,51200,0.07982310983869764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,128,0.033391998873816595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,2048,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,65536,0.12563467025756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,64,0.010375111467308467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,512,0.009517333573765224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,2560,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,4096,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,6144,0.02239022155602773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,5120,0.04546844296985202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,2048,0.006554666492674086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,8192,0.05353511042065091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,65536,0.03091111116939121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,5120,0.016333333320087857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,4096,0.011358222199810876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,1024,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,768,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,5120,0.04958577619658577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,10240,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,32,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,1024,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,8192,0.02104088829623328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,10240,0.016555555992656283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,128,0.09252266751395331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,3584,0.017303110824690927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,10240,1.3192604912651908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,1024,0.028414222929212783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,6144,0.014870221416155497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,6144,0.4586284425523546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,10240,0.03724177678426107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,2048,0.020392000675201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,51200,0.14833866225348577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,2048,0.03384000062942505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,8192,0.048150221506754555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,51200,0.06443111101786296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,51200,0.05617955658170912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,2048,0.04929511083496941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,256,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,768,0.37377243571811253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,16384,0.18825688627031115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,12288,0.03569333420859443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,768,0.054692443874147206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,16384,0.020605333977275424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,6144,0.18609066804250082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,4096,0.011858666936556498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,6144,0.00961599995692571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,4096,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,4096,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,768,0.07325066460503472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,8192,0.014984000060293408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,128,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,12288,0.08766400151782566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,12288,0.03193511234389411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,65536,0.06443555487526788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,8192,0.04853244291411506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,768,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,2560,0.08475555313958062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,6144,0.07779644595252143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,1024,0.016341333587964375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,7168,0.030416889323128596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,2048,0.041195554865731135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,12288,0.03736889031198289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,1536,0.24455287721421984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,256,0.012697777814335294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,32,0.0048142220411035745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,3584,0.015935111376974318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,128,0.009631111390060848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,8192,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,256,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,16384,0.19104799959394667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,3584,0.009415110780133141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,10240,0.02884977724817064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,51200,0.14149955908457437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,3584,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,128,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,32,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,5120,0.007112888826264276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,4096,0.39140976799858945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,4096,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,512,0.011336888704035016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,32,0.05087288882997301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,32,0.002013333348764314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,4096,0.018602665927675035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,64,0.306858672036065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,32,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,8192,0.01351999988158544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,2560,0.08811644713083903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,512,0.013809777796268463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,64,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,2560,0.028604444530275133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,16384,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,3584,0.0087013335691558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,7168,0.08156177732679579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,128,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,32,0.01681066718366411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,4096,0.048842665221956044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,7168,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,3584,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,10240,0.02288088864750332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,8192,0.06993777884377374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,128,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,51200,0.047998223039839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,128,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,2048,0.27335911326938206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,256,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,3072,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,12288,0.15307911237080893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,7168,0.2731280061933729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,512,0.013228444589508904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,1024,0.006343110981914733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,1024,0.013032000097963544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,1024,0.017165333032608032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,12288,0.03826933436923557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,2048,0.011111111276679568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,16384,0.05297866794798109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,1536,0.03200533323817783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,6144,0.1077679991722107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,1536,0.020006222857369315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,1536,0.019858666592174105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,64,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,3584,7.050132327609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,2048,0.19020267327626547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,12288,0.053146666950649686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,65536,0.8848053084479438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,10240,0.1066124439239502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,2048,0.0644577807850308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,12288,0.22837511698404947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,4096,0.0079297779334916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,4096,0.03275288807021247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,10240,0.06234044498867459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,6144,0.060271110799577504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,51200,0.08118400308820936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,12288,0.07770844300587972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,3584,0.03987911012437608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,512,0.04251111216015286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,7168,0.022808000445365906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,16384,0.18319555123647055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,12288,0.03392444385422601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,8192,0.1392328871621026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,2048,0.02237244447072347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,16384,0.23106045193142363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,2560,0.006775110960006714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,3072,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,12288,0.013393777940008374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,2048,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,7168,0.03545066714286804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,65536,0.6563430892096626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,16384,0.230292452706231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,768,0.006167999986145232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,64,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,12288,0.18667377365960014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,1536,0.020052444603708055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,5120,0.06232711341645983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,1024,0.04235466652446323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,6144,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,256,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,65536,0.06176355812284681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,10240,0.029866665601730347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,51200,0.1052817768520779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,5120,0.07242488861083984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,2048,0.007481777833567725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,128,0.004513777792453766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,3584,0.09503910938898723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,1024,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,2560,0.036680890454186335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,8192,0.02161955502298143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,2560,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,1024,0.018411555223994784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,2048,0.020423110988405015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,3072,0.02797599964671665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,64,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,512,0.012246222131782107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,64,0.0022328889204396144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,3072,0.026692445079485577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,128,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,1024,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,7168,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,6144,0.05934311283959282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,7168,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,65536,0.08789066473642985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,6144,0.0388426681359609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,5120,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,65536,0.7009404500325521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,16384,0.03812888926929898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,1024,0.033765332566367254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,8192,0.03264266583654616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,7168,0.05375911129845513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,12288,0.27020443810356987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,6144,0.08098666535483466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,3584,0.01424533294306861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,65536,0.27440089649624294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,2048,0.021942221456103857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,7168,0.06932622194290161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,1024,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,3584,0.01184177812602785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,6144,0.010207110808955299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,1024,0.02071555621094174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,32,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,8192,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,3072,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,7168,0.05954044395022922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,512,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,3584,0.03773066732618544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,8192,0.015322667029168872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,6144,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,2048,0.006768888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,1024,0.017129777206314933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,4096,0.04021866785155402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,1024,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,3072,0.019555555449591745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,1536,0.0035706667436493766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,1024,0.019277332557572257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,256,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,8192,0.050423112180497914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,2560,0.04137600130505032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,65536,0.13334222634633383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,4096,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,768,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,12288,0.1062542200088501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,6144,0.012154666913880242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,3072,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,10240,0.012783110969596438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,51200,0.621429337395562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,2560,0.04473511046833462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,7168,1.20614136589898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,1536,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,6144,0.01809155609872606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,10240,0.10412800312042236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,3072,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,1536,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,3584,0.06798488563961454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,4096,0.07599466376834445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,16384,0.01903199983967675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,51200,0.9634417427910699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,5120,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,64,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,6144,0.015121777852376303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,3584,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,65536,0.3095457818773058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,5120,0.2798062165578206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,51200,0.05504088931613498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,512,0.014863999353514777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,1536,0.046796444389555186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,5120,0.019924443629052904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,3072,0.02113955550723606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,768,0.010738666686746808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,8192,0.14896710713704428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,51200,0.06572533316082425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,12288,0.16780711544884575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,512,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,128,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,6144,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,2048,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,10240,0.561832004123264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,7168,0.014564444621404013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,8192,0.03300088975164626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,3584,0.03584533267550998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,256,0.020703110429975722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,10240,0.05094844433996412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,8192,0.07976800203323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,2048,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,16384,0.10237689150704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,6144,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,12288,0.42200620969136554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,512,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,6144,0.04326488905482822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,10240,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,4096,0.09618666436937119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,128,0.014336888988812765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,51200,0.10309422016143799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,128,0.027956446011861164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,16384,0.018072888255119324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,3072,0.07603022125032213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,128,0.031558222240871854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,12288,0.01717866626050737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,4096,0.029930667744742498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,3072,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,4096,0.032207111517588295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,2048,0.01260355528857973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,2048,0.1032808886633979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,8192,0.011353777514563667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,4096,0.12069600158267552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,12288,0.27546845542060006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,2048,0.028918223248587713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,7168,0.02156977852185567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,65536,0.1632817718717787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,3584,0.0625200006696913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,1024,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,2048,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,2048,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,6144,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,3072,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,128,0.015944888194402058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,7168,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,3584,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,12288,0.054232001304626465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,2048,0.03426133261786567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,16384,0.13867644468943277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,3584,0.7490977711147733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,3584,0.03381777803103129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,10240,0.401417785220676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,10240,0.02500177754296197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,3072,0.03315911028120253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,512,0.019148444135983784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,16384,0.030815998713175457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,7168,0.06652888986799452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,5120,0.037610666619406805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,3584,0.014975999792416891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,65536,0.13694489002227783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,2048,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,4096,0.5619466569688585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,65536,0.12398489316304524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,5120,0.039042668210135564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,512,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,768,0.017614222235149808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,2048,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,65536,0.012151111331250934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,2560,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,10240,0.025403555896547105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,512,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,4096,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,7168,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,8192,0.09659110837512547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,1536,0.011004444625642566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,7168,0.1210177739461263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,4096,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,2560,0.014182221558358936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,4096,0.02125688890616099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,32,0.6192062166002061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,256,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,2048,0.06851644648445977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,8192,0.11827199988894993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,128,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,10240,0.30496356222364635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,2560,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,10240,0.045532445112864174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,64,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,768,0.03253511256641812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,2048,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,5120,0.008752000000741746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,65536,0.34371643596225315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,768,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,3072,0.06789511442184448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,4096,0.0351164440313975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,51200,0.06919555531607734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,51200,0.08340089188681708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,3584,0.018425777554512024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,1024,0.024791111548741657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,1024,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,3584,0.09099733167224461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,1024,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,256,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,51200,0.8205626805623373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,16384,3.9443206787109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,5120,0.014017777310477363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,4096,0.026022222306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,128,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,12288,0.5964444478352865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,3584,0.0299217767185635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,1536,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,768,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,128,0.0069884442620807225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,10240,0.07432266738679674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,1536,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,3584,0.010319111247857412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,1536,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,128,0.0790862242380778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,4096,0.014347554908858405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,10240,0.28617599275377065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,2048,0.04011111127005683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,2560,0.02313511073589325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,1024,0.0028702221396896574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,4096,0.05945955382453071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,6144,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,51200,0.1675084431966146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,256,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,1536,0.008413333031866286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,512,0.045203556617101036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,65536,2.273146735297309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,3072,0.007381333245171442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,128,0.07654577493667603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,1024,0.01614755557643043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,3584,0.02683111031850179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,3584,0.014641778336630927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,64,0.00906133320596483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,4096,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,128,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,12288,0.14292177889082167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,3072,0.014183999763594734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,12288,0.054198222027884595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,65536,0.12616799937354192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,10240,0.03402933478355408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,6144,0.1282915539211697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,5120,0.012793777717484368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,1024,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,16384,0.05583910809622871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,1536,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,3584,0.14497866895463732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,2560,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,7168,0.04867466621928745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,4096,0.027705777022573683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,16384,0.03085333440038893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,128,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,1024,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,5120,0.09119555685255264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,3584,0.007531555162535773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,51200,1.3253857294718425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,12288,1.798127916124132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,64,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,8192,0.16906844245062935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,256,0.03303022185961405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,4096,0.02796444296836853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,1024,0.02898755669593811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,2048,0.02501511077086131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,6144,0.03734577695528666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,256,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,768,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,2560,0.01607911123169793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,2048,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,1536,0.01956977778010898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,128,0.02164799968401591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,64,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,64,0.008296888735559251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,8192,0.0417137775156233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,768,0.004535111288229625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,6144,0.05741066402859158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,65536,0.07547555367151897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,12288,0.048057778014077075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,7168,1.3034213383992512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,512,0.04872533347871569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,10240,0.10077510939704047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,512,0.011798222031858234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,1536,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,256,0.002247111043996281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,65536,0.16298311286502415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,6144,0.013933334085676404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,512,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,4096,0.03064089020093282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,10240,0.03442666596836514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,51200,3.8572167290581594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,51200,3.771998087565104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,768,0.026390223039521113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,12288,1.4298729366726344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,768,0.004818666726350784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,3072,0.022612444228596155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,64,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,3584,0.10855288638008966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,65536,0.3756471210055881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,3072,0.03952799903021918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,6144,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,768,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,51200,0.34543466567993164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,128,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,1536,0.006173333360089197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,2048,0.027746667464574177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,51200,0.06970755259195964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,768,0.03234577841228909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,4096,0.08785333236058553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,7168,0.016721778445773654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,1024,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,3584,0.01830755505296919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,4096,0.11304444736904568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,32,0.01406311160988278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,51200,0.04861155483457777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,51200,0.3249315685696072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,7168,0.07760977745056152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,10240,0.03594222333696153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,65536,0.1465244425667657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,16384,0.7179786894056531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,5120,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,512,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,8192,0.03539822167820401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,2048,0.005197333378924264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,512,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,8192,0.04275200102064344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,768,0.12860711415608725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,65536,0.06251733170615302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,256,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,2560,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,3584,0.01756088932355245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,8192,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,7168,0.017480888300471835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,768,0.011039111349317761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,128,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,256,0.024103111690945093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,32,0.0023440000497632557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,256,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,256,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,16384,0.02979555394914415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,51200,0.05603466431299845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,7168,0.34142043855455184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,1024,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,2048,0.013757333159446716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,3584,0.03568266828854879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,2560,0.013728000223636627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,16384,0.09793244467841254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,65536,0.06261066595713298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,2048,0.019932443896929424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,3072,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,6144,0.03968355390760634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,1536,0.027561777167850073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,1024,0.026342221432262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,256,0.011516444385051727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,8192,0.016745777593718637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,65536,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,128,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,512,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,256,0.03894311189651489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,1024,0.012677333421177335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,5120,0.049667557080586754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,32,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,768,0.3108479976654053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,65536,1.444483545091417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,3072,0.021866665946112737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,256,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,51200,0.136153777440389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,5120,0.008900444540712569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,7168,0.03683022326893277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,3584,0.027699554959932964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,7168,0.023081777824295893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,51200,0.07720977730221219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,3072,0.04224444429079691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,128,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,7168,0.02174844510025448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,6144,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,4096,0.029843555556403265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,1536,0.006331555545330048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,2560,0.026664000418451097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,7168,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,10240,0.017197334104114108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,8192,0.17267910639444986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,128,0.014178666803571912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,16384,0.05067200130886502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,6144,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,1024,0.02090577781200409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,16384,0.17055111461215547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,8192,0.025288888149791296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,5120,0.018977777825461496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,1024,0.02279377811484867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,2560,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,768,0.003160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,16384,0.4989031155904134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,2560,0.03319999906751845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,768,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,1024,0.057836446497175426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,512,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,32,0.02370044423474206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,8192,0.030671109755833943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,8192,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,1024,0.006414222220579783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,2048,0.02375288969940609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,10240,0.3702675501505534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,6144,0.01742755538887448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,10240,0.07058666812049018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,8192,0.010948444406191507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,256,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,7168,0.028184887435701158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,51200,0.20536177688174775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,512,0.01499022212293413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,51200,1.5825262069702148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,10240,0.10530489020877415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,3072,0.10493599706225926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,1024,0.012640000217490725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,51200,0.09676355785793728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,5120,0.02565333412753211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,16384,0.05200711223814222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,64,0.00564444437623024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,12288,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,4096,0.07471466726726957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,12288,0.2581004566616482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,1536,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,2560,0.0045422220395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,7168,0.00816533300611708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,6144,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,1536,0.05848711066775852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,1536,0.05871022409862942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,8192,0.29771023326449925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,12288,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,1536,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,10240,0.038153777519861855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,5120,0.030765334765116375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,6144,0.025145777397685583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,1024,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,5120,0.038910223378075495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,65536,0.10266133149464925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,3072,0.011735111474990845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,256,0.013479111095269522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,2560,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,4096,0.049842665592829384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,1024,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,2560,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,7168,0.030118223693635728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,3072,0.0906568898095025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,6144,0.010998222563001843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,1536,0.04411466585265266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,64,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,65536,0.16636978255377874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,6144,0.2302417755126953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,2560,0.9357688691880969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,4096,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,128,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,10240,0.019552888141738046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,64,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,2048,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,5120,0.08771733442942302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,3072,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,2560,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,512,0.005713777823580637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,3072,0.5370035701327854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,128,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,8192,0.025011556016074285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,2048,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,4096,0.01779022150569492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,12288,0.11034578084945679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,2048,0.004936000125275718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,128,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,1536,0.003903999924659729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,10240,0.012737777498033313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,3072,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,1024,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,768,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,16384,0.06457955307430692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,7168,0.0214720004134708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,7168,0.02455111179086897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,2048,0.041794667641321816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,4096,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,16384,0.022882666852739122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,3072,0.2043386697769165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,1536,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,51200,0.06962044371498956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,32,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,7168,0.013184000220563678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,65536,0.37601155704922146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,3584,0.05812888675265842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,7168,0.024380443824662104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,768,0.03162044286727905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,768,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,2048,0.06775289111667208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,32,0.013259555730554791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,256,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,7168,0.044770667950312294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,7168,0.02083911167250739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,256,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,10240,0.04348444607522753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,768,0.020453333854675293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,128,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,8192,0.043600890371534556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,51200,1.7160071267022028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,32,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,8192,1.6255227194892035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,12288,0.021003555920388963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,10240,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,5120,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,3584,0.007698666718271043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,64,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,6144,0.0336079994837443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,512,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,7168,0.00461777796347936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,65536,0.12403910689883763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,7168,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,2048,0.004608889006906086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,16384,0.0251226673523585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,128,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,51200,0.3301066557566325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,2560,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,1024,0.011788444386588203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,64,0.048024889495637685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,6144,0.02772355576356252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,1024,0.015306666493415833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,10240,0.012591111163298288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,2560,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,768,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,8192,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,64,0.018592000007629395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,3072,0.009791110952695211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,2048,0.007048889166778988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,10240,0.05819466378953722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,10240,0.16918578412797716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,5120,0.08850222163730198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,128,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,128,0.01517777807182736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,3584,0.025474665893448725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,256,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,64,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,512,0.056122667259640165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,10240,0.025848888688617285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,768,0.0026586666289303037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,7168,0.024933333198229473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,1024,0.015188443991872998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,32,0.004600000050332812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,2048,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,64,0.015188443991872998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,64,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,2048,0.0070088886552386815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,4096,0.04146044452985128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,3584,0.020579554968410067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,51200,0.0554275545809004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,10240,0.008346666892369589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,16384,0.020557334025700886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,8192,0.2631591161092122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,256,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,768,0.01812088820669386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,2560,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,2048,0.012060444388124677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,64,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,7168,0.0614337788687812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,512,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,4096,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,768,0.017237333787812125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,64,0.007363555332024892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,16384,0.08849955267376369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,2560,0.018872888551818002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,12288,0.03463110989994473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,65536,0.15392266379462347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,6144,0.034416000048319496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,256,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,1024,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,5120,0.02719199988577101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,5120,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,768,0.018888889087571036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,7168,0.6277111371358236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,1536,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,3072,0.005331555588377847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,128,0.016206221448050607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,3072,0.026662222213215295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,65536,0.21678666273752847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,1536,0.08408266968197292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,3584,0.030799110730489094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,768,0.015219555960761176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,2560,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,7168,0.08885244528452556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,51200,0.05135022269354927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,128,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,2048,0.015466666883892484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,256,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,6144,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,4096,0.051056000921461314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,64,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,128,0.00943377779589759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,8192,0.022117333279715642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,4096,0.02147377696302202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,2048,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,3072,0.012347555822796293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,256,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,3584,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,51200,3.415223015679253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,8192,0.09013155433866714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,1536,0.008796444369686974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,10240,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,8192,0.04450044367048475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,8192,0.026183111800087824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,2560,0.016955556141005624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,16384,0.07998400264316134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,65536,0.03703733285268148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,768,0.009173333644866943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,3584,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,51200,0.1079324417644077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,3072,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,8192,0.0410097771220737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,12288,0.1083751122156779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,256,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,3584,0.1519893275366889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,8192,0.023166222704781428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,51200,0.05565155877007378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,5120,0.08275644646750556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,8192,0.05937777625189888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,512,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,65536,0.06256888972388373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,256,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,1024,0.01903377804491255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,128,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,5120,0.27707378069559735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,3584,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,768,0.004281777888536453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,3584,0.040941334433025785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,256,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,32,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,2048,0.029242667886945937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,4096,0.033759110503726535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,51200,0.26885777049594456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,8192,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,51200,1.7440960142347548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,5120,0.15633866522047254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,2048,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,2048,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,5120,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,512,0.020230222079488967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,10240,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,7168,0.12893599934048122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,128,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,2560,0.022061333060264587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,1024,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,2048,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,1536,0.008197333249780867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,2560,0.008573333422342936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,65536,1.0801475312974718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,1536,0.020796444680955675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,7168,0.009737778041097853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,6144,0.2859848870171441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,3072,0.08623111248016357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,6144,0.04640444450908237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,5120,0.027035555905765955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,10240,0.015590222345458137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,16384,0.07548889186647204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,1024,0.019838222199016146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,2560,0.02181422213713328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,32,0.005534222142563925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,8192,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,10240,0.09641866551505195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,2048,0.007702222300900354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,128,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,10240,0.061257779598236084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,768,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,128,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,51200,0.07035377952787611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,128,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,4096,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,6144,0.01888177792231242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,512,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,7168,0.15792622831132677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,10240,0.011855999628702799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,512,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,10240,0.24383733007642958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,3072,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,1536,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,7168,0.09863555431365967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,512,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,2048,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,51200,0.05135022269354927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,2560,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,12288,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,3072,0.5257351133558485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,128,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,64,0.003806222230195999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,1024,0.05258488986227247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,3072,0.16527556048499212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,12288,0.0895964437060886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,10240,0.035944888989130654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,51200,0.10699733098347981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,2560,0.010440000229411654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,5120,0.09017599953545465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,256,0.05000977714856466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,10240,0.032167110178205706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,256,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,2560,0.018810666269726224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,65536,0.18090221616956922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,32,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,4096,0.07293955485026042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,7168,0.032969776127073504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,8192,0.053307556443744235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,10240,0.05200533403290642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,2048,0.00700711127784517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,2048,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,2048,0.02795555525355869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,3072,0.02083288960986667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,3072,0.017304889029926725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,3584,0.008117333054542542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,768,0.011959999799728394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,2560,0.03106044398413764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,6144,0.028852442900339764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,1536,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,3072,0.10354844729105632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,1024,0.011391110718250275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,65536,0.08382311132219102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,256,0.007515555454625024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,5120,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,2560,0.007695111135641734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,32,0.0018906665758954154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,65536,1.0751413769192166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,512,0.06959378057056002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,12288,0.09140711360507542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,3584,0.13192889425489637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,2048,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,65536,0.06170577473110623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,1536,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,3072,0.05559111303753323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,2048,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,12288,0.1320106718275282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,2560,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,51200,0.05085777905252245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,1536,0.03619022170702616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,1024,0.023341332872708637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,7168,0.03460977805985345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,1536,0.017283555534150865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,128,0.013798221945762634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,51200,0.1740871138042874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,1536,0.019242667489581637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,1536,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,512,0.05263288815816244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,8192,0.08475377824571396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,32,0.009875555833180746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,256,0.02083555526203579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,16384,0.13585244284735784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,8192,0.5906720161437988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,51200,0.10487555795245701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,5120,0.012622222304344177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,65536,0.0716995530658298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,256,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,12288,0.013516444298956128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,1024,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,12288,0.1311537822087606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,512,0.0453688899676005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,10240,0.012054222325483957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,10240,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,3072,0.04292800029118856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,1536,0.005773333211739858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,2560,0.009663110805882348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,5120,0.03014933400683933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,5120,0.2513519922892253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,2048,0.020272889071040683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,128,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,6144,0.060277336173587374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,256,0.07428889142142402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,3584,0.09780622190899319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,65536,11.69958750406901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,512,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,768,0.02012800011369917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,128,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,12288,0.04754933383729723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,4096,0.10677778058581883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,16384,0.07805777920616998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,3072,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,10240,0.05713599920272827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,512,0.0026151111556424033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,32,0.0052888890107472735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,128,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,10240,0.11227822303771973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,5120,0.030260443687438965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,1024,0.01109955542617374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,51200,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,7168,0.02869510981771681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,2560,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,7168,0.03492710987726847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,4096,0.016015999846988253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,768,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,51200,0.4786080254448785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,16384,0.0731111102634006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,6144,0.0728728903664483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,5120,8.074176364474827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,3072,0.026338666677474976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,10240,3.363339530097114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,10240,0.33621599939134383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,8192,0.08612000279956394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,12288,0.019924443629052904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,5120,0.013598221871587964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,3584,0.019621334142155118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,2560,0.019914666811625164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,65536,0.12192977799309625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,3072,0.038973334762785174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,4096,0.031587554348839655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,256,0.008022222254011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,65536,0.05711288584603203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,2048,0.020941333638297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,6144,0.49974579281277126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,65536,0.17894577980041504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,8192,0.14395111136966282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,64,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,5120,0.03664977682961358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,7168,0.11917244063483344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,2048,0.027079999446868896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,7168,0.009192888935407003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,4096,0.05709155400594076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,7168,0.12916978200276694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,3584,0.023969777756267126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,5120,0.12802578343285456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,768,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,1024,0.0210515558719635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,4096,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,128,0.014858666393491956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,3072,0.07289777861701117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,2048,0.00792266676823298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,65536,3.2338434855143228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,51200,0.099199997054206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,2560,0.022126222650210064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,2048,0.041871110598246254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,5120,0.01349066694577535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,10240,0.02828977836502923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,8192,0.0266951107316547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,768,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,256,0.01629066632853614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,128,0.027707555227809485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,6144,0.014646222194035849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,6144,0.016156444946924847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,1024,0.06429866949717204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,3584,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,10240,0.0517448882261912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,32,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,51200,0.5745893584357368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,51200,30.857001410590275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,10240,0.07326489024692111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,4096,0.280603567759196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,128,0.022456000248591106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,7168,0.006559999866618051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,6144,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,3584,0.10648177729712592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,64,0.005380444642570283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,16384,0.024158222807778254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,12288,0.04577689038382637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,51200,0.05941244628694323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,6144,0.04894489049911499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,6144,0.11568089326222737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,128,0.01312800000111262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,16384,0.10664266347885132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,51200,0.05008888906902737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,2048,0.20133422480689156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,32,0.00703111125363244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,6144,0.0225564440091451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,768,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,2560,0.14071199629041883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,16384,0.04078133238686456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,2048,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,2560,0.007199111084143321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,12288,0.16525510946909586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,128,0.017451556192504037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,256,0.039334221018685236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,12288,0.0603297750155131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,7168,0.036343110932244196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,8192,0.138410660955641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,1024,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,3072,0.09186933437983195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,6144,0.04720088839530945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,1536,0.019896000623703003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,3072,0.007873777714040544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,5120,0.08015466398662992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,3072,0.07868622408972846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,5120,0.017498667041460674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,768,0.03547911180390252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,1024,0.007387555307812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,5120,0.008945778012275696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,4096,0.10284888744354248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,6144,0.038197331958346896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,3072,0.05966577927271525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,1536,0.02860266632503933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,128,0.0022044444663657085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,10240,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,512,0.005314666777849197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,5120,0.035407112704383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,10240,0.02348533272743225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,51200,0.8242364459567599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,16384,0.054663111766179405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,3584,0.04539022180769178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,4096,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,12288,0.05939200189378527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,16384,0.026514665948020086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,2560,0.06967466407352023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,512,0.0034968890249729156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,5120,0.0565226674079895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,1024,0.014128888646761576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,128,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,1536,0.006340444501903322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,2048,0.3191582096947564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,2048,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,2560,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,12288,0.03679200013478597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,12288,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,1024,0.015612444943851896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,12288,0.026701332794295415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,4096,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,768,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,10240,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,512,0.006189333481921091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,768,0.009870222045315636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,51200,0.24978134367201063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,16384,0.0185217774576611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,2560,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,512,0.140883551703559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,256,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,16384,0.07528799772262573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,5120,0.021065778202480737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,16384,0.1083751122156779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,32,0.004571555389298333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,4096,0.03525777657826742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,3072,0.05468533436457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,10240,0.03273866573969523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,3584,0.46022489335801864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,8192,0.04669510987069872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,2560,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,128,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,3584,0.0081404447555542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,6144,0.0381368895371755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,16384,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,8192,0.012060444388124677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,3584,0.027767111857732136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,5120,0.02018222212791443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,2048,0.06092533138063219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,4096,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,65536,0.08723111285103692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,16384,0.26408089531792533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,7168,0.013933334085676404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,12288,0.035007112556033664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,1024,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,32,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,256,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,3584,0.030672887961069744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,768,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,5120,0.025605332520272996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,2048,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,10240,0.13839111063215467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,8192,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,8192,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,128,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,256,0.03897688786188761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,1024,0.00903555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,5120,0.02903644575013055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,3584,0.02796444296836853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,6144,0.48581245210435653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,512,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,128,0.01447111037042406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,512,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,2560,0.031003554662068684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,768,0.02095111045572493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,1536,0.0041662222809261745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,7168,0.15225777361128065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,4096,0.1003991100523207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,64,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,512,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,12288,0.021008888880411785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,5120,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,6144,0.0743288861380683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,12288,0.020082667469978333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,3072,0.032315555546018816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,65536,0.1524853308995565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,2048,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,1024,0.024683554967244465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,2048,0.019849777221679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,64,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,512,0.027287110686302185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,10240,0.5055386755201552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,1536,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,8192,0.03264177838961283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,10240,0.01605866683853997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,2560,0.03916888766818576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,65536,0.2630639870961507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,1536,0.013314666847387949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,16384,0.04253333475854662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,1536,0.006792000184456508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,10240,0.023386667172114056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,5120,1.054356469048394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,65536,0.06271288792292277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,51200,1.2356523937649198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,512,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,64,0.010507555471526252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,8192,0.05898400147755941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,7168,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,1536,0.0203813330994712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,3584,0.09122933281792535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,5120,0.049323555496003896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,128,0.00812355594502555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,2048,0.4107084539201524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,5120,0.0918346643447876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,12288,0.548967096540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,4096,0.1738293303383721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,7168,0.03194222185346816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,4096,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,2048,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,32,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,128,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,6144,0.024340444140964087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,8192,0.02166755497455597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,12288,0.02719733284579383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,4096,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,4096,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,12288,0.02939644455909729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,6144,0.027105778455734253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,3584,0.04773244592878553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,768,0.017052445146772597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,2560,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,2560,1.3178781933254666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,256,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,512,0.007715555528799693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,1536,0.00849955528974533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,65536,0.31235289573669434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,8192,0.01074044406414032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,256,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,10240,0.05218755536609226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,5120,0.3412764337327745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,51200,0.11626844935946995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,1536,0.02847466535038418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,2048,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,256,0.028505778974956934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,7168,0.014743111199802823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,6144,0.03187555405828688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,7168,0.014088888963063559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,2560,0.011373333632946014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,7168,0.0570711096127828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,128,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,2560,0.029653333955340918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,3584,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,3072,0.04100799891683791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,3584,0.050809777445263334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,12288,0.10684088865915935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,3584,0.023707555400000677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,2048,0.057348443402184375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,51200,1.021176020304362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,3072,0.027560888065232172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,256,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,12288,0.09467377927568223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,5120,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,3584,0.04275111026234097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,4096,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,3584,0.034485333495669894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,16384,0.024043555061022442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,5120,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,6144,0.03494222296608819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,768,0.005720888988839255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,5120,0.00962755580743154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,10240,0.025462221768167283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,51200,0.1464666657977634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,10240,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,2560,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,5120,0.00983911090426975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,65536,0.07475199964311388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,512,0.0038542221817705366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,7168,0.024267555938826665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,512,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,256,0.014169777433077494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,12288,0.042336000336541064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,3072,0.024304888314670984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,3072,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,32,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,256,0.0023004443695147834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,2560,0.01919644408755832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,3072,0.038145777251985334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,128,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,3584,0.01927911076280806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,1024,0.011699555648697747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,64,0.002455111179086897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,512,0.10419644249810113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,3584,0.07126222054163615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,51200,0.049806220663918384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,12288,0.1536533302730984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,768,0.0509786672062344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,1024,0.015423110789722867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,2560,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,1536,0.03286222285694546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,51200,0.18818488385942247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,1536,0.02641155487961239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,8192,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,5120,4.245490603976779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,256,0.0028151110228565005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,2048,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,5120,0.02106488909986284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,512,0.025422222084469263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,65536,0.4843164549933539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,12288,0.05980889002482096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,10240,0.041047109497918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,1536,0.039637333816952176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,10240,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,2560,0.019902222686343722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,1536,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,12288,0.13226311736636692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,256,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,7168,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,16384,0.027178666657871667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,1536,0.029437333345413208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,512,0.011613333390818702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,16384,0.25959022839864093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,10240,0.15165778001149496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,10240,0.30937690205044216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,32,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,12288,0.10976177453994751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,768,0.06833066542943318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,768,0.01627200014061398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,65536,0.06196000178654989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,32,0.01752000053723653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,2048,0.011010666688283285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,768,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,7168,0.017337777548366122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,512,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,3584,0.03871022330390082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,128,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,65536,0.0574586656358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,4096,0.03959200117323134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,8192,0.047832889689339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,8192,0.01349511080318027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,128,0.00563733321097162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,6144,0.17390933301713732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,51200,0.029489778810077246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,6144,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,8192,0.07450933588875665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,3584,0.6691875457763672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,16384,0.058621333705054395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,768,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,4096,0.020632889535692003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,16384,0.054042667150497437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,6144,0.05441866649521721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,10240,0.12988977962070042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,12288,0.1160915560192532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,256,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,32,0.02148088812828064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,4096,0.4044942326015896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,51200,0.09068177806006537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,10240,0.018485332528750103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,2560,0.02195288903183407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,7168,0.027622222900390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,256,0.017235555582576327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,6144,0.11058221922980414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,768,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,64,0.0017911110901170308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,2560,0.04574844572279188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,8192,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,4096,0.14662578370836046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,5120,0.06419377856784396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,512,0.018367111682891846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,51200,0.1133342186609904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,6144,0.019559999306996662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,1024,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,2560,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,10240,0.015528000063366361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,64,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,2560,0.007480888730949826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,12288,0.018331555856598746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,3072,0.8010124630398221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,256,0.0032239999208185407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,65536,0.4479973581102159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,256,0.006024000131421619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,4096,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,1024,0.04141511188613044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,10240,0.018560889694425795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,6144,0.09773955742518108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,512,0.007040888898902469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,2560,0.09490488635169135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,4096,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,51200,0.05996711386574639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,5120,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,2560,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,512,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,5120,0.04786400000254313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,8192,0.4086844391292996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,16384,0.045419553915659584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,8192,0.17007999949985078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,256,0.012470222181744046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,1536,0.010703999963071613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,2048,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,5120,0.009557333257463243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,65536,0.05431288811895582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,128,0.002209777840309673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,32,0.0032133333798911837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,51200,0.8628488646613227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,6144,0.04389511214362251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,51200,0.0655884411599901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,7168,0.1551697783999973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,12288,0.026366222235891554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,768,0.02049688829316033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,3584,0.015967999895413715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,1536,0.13909955819447836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,10240,0.026592888765864905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,5120,0.026962666047943964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,65536,0.029432889488008287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,2560,0.023299554983774822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,10240,0.01844888925552368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,3072,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,5120,0.10186933146582709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,5120,0.014702222413486905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,128,0.015670221712854173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,512,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,3072,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,6144,0.07618844509124756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,7168,0.21868088510301378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,51200,0.047704888714684375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,768,0.01441422270403968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,51200,0.15299199687110052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,2048,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,2048,0.00790666706032223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,65536,3.8121804131401906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,65536,0.06461422310935126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,768,0.05954133139716255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,768,0.019087110956509907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,4096,0.045759111642837524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,7168,0.058984888924492725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,12288,0.05105333195792305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,7168,0.042279998461405434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,3072,0.09234222438600327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,7168,2.4848639170328775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,3584,0.00888177752494812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,3072,0.0646124415927463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,8192,0.038020445240868464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,2560,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,7168,0.016333333320087857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,2048,0.01090222183201048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,256,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,512,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,12288,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,12288,0.1062684456507365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,768,0.028584887584050495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,4096,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,128,0.022965333527988855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,6144,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,3072,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,768,0.020270221763186984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,51200,0.27978489134046763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,16384,0.03816355599297418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,6144,0.046912001238928906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,10240,0.017595556047227647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,10240,0.01737244427204132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,16384,0.0958062211672465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,512,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,256,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,12288,0.23412799835205078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,16384,0.1697324381934272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,1024,0.07475111219618055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,32,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,256,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,65536,0.2686577902899848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,2048,0.008175111479229396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,7168,0.023614222804705303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,65536,0.5924382209777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,3072,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,10240,3.058638254801432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,1536,0.006055111272467508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,10240,0.013810666898886362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,512,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,768,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,128,0.014456889695591398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,10240,0.031926221317715116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,6144,0.011330666641394297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,3584,0.010476444330480365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,16384,0.03276266654332479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,3584,0.03677333394686381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,16384,0.05168088939454821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,2560,0.022220444348123338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,10240,0.043586668041017324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,12288,0.0322106679280599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,5120,0.03375733229849074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,3584,0.04310577776696947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,10240,0.020604444874657523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,3584,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,4096,0.017794667018784415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,5120,0.01605422298113505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,5120,0.020640888147883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,2048,0.016601777738995023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,65536,0.5716284645928277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,51200,0.3597466680738661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,1536,0.08646577596664429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,6144,0.038916445440716214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,8192,0.025343111819691126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,8192,0.055120888683531016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,1024,0.010067555639478896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,6144,0.03552622265285916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,65536,1.0792231029934354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,32,0.027683556079864502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,10240,0.015853333804342482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,3072,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,51200,0.12961333327823216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,51200,0.16011910968356663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,1024,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,768,0.011944000091817645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,4096,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,8192,0.023128888673252527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,3584,0.38474665747748477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,6144,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,1536,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,32,0.02535377773973677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,12288,0.05184622274504768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,512,0.005527110977305307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,1024,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,2560,0.0070426662762959795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,512,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,768,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,1536,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,12288,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,1536,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,1536,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,1024,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,2560,0.08410577641593085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,4096,0.7051119804382324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,65536,0.10349244541592068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,2560,0.1641226609547933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,7168,0.027237332529491846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,8192,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,1024,0.009455111291673448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,2048,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,256,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,16384,3.4767013125949435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,256,0.021987555755509272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,512,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,2560,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,128,0.04634044567743937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,8192,0.451340463426378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,5120,0.008642666869693333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,5120,0.014861333701345654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,5120,0.027246221899986267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,65536,2.096594704522027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,7168,0.015039111177126566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,6144,0.019409777389632333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,7168,0.05774844355053372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,1024,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,3584,0.021964444054497614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,1024,0.022874666584862605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,1024,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,51200,0.8074373139275445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,5120,0.06318666537602742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,4096,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,2560,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,3072,0.023649776975313824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,7168,0.049421333604388766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,16384,0.07993155717849731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,65536,2.460571500990126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,51200,0.19976177480485705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,128,0.0031386667655573953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,65536,0.13524799876742893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,6144,0.03769955701298184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,12288,0.020117332537968952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,512,0.019469334019554984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,10240,0.14641600184970432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,5120,0.032525334093305804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,2560,0.013764444324705334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,6144,0.030031111505296495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,6144,0.03013777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,4096,0.13417422771453857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,3584,0.07515555620193481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,3584,0.07066577672958374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,4096,0.07049777772691515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,8192,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,51200,0.1284995608859592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,3072,0.013829333086808523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,256,0.022171555293930903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,1024,0.020809777908855014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,1024,0.005664000080691443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,10240,0.03547911180390252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,3072,0.012258666257063547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,768,0.152493331167433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,65536,0.06467822525236341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,16384,0.27703822983635795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,32,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,6144,0.06491822004318237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,12288,1.3344720204671223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,65536,0.06372800138261583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,3584,0.06731555859247844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,1024,0.017487999465730455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,768,0.02033066584004296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,8192,0.018162666095627677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,128,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,2048,0.019768888751665752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,3584,0.00964355551534229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,768,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,6144,0.020432000358899433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,5120,0.039656000004874334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,4096,0.00848355558183458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,7168,0.028568890359666612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,4096,0.046148445871141225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,256,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,2048,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,2048,0.019918221566412184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,1536,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,5120,0.012702222499582501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,8192,0.026739555928442214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,256,0.013207111093733044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,7168,0.13225777943929037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,8192,0.2833555539449056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,128,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,51200,0.07092177867889404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,6144,0.04184266593721178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,1536,0.02604177759753333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,64,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,8192,0.011512888802422417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,5120,0.04282933473587036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,512,0.4093937873840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,8192,0.08788622087902492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,2048,0.2808604505327013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,65536,0.29567734400431317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,12288,0.04510044389300876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,65536,0.10031021965874566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,16384,0.02202577723397149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,16384,0.030155556069480047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,768,0.012866666747464074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,64,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,768,0.0063235556913746735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,7168,0.0639973349041409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,3584,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,5120,0.02161866592036353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,65536,0.09522044658660889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,1024,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,16384,0.03573777940538194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,3072,0.01536622146765391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,768,0.03103111187616984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,5120,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,3584,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,65536,0.14671111106872559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,256,0.0034666665726237824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,51200,3.3241074879964194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,2560,0.013288000391589271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,2560,0.12419288688235813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,64,0.01182044463025199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,3584,0.012849777936935425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,8192,0.01088177743885252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,51200,0.29396533966064453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,65536,0.06998755534489949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,2048,0.01182933317290412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,7168,0.013008889224794177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,3072,0.031558222240871854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,65536,0.7415564325120715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,64,0.025478222303920325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,128,0.06829244560665555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,1024,0.015325332681337992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,512,0.005092444519201915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,65536,0.338658677207099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,3584,0.07249422205819024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,32,0.002986666642957263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,8192,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,256,0.03360177742110358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,512,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,2048,0.012802666260136498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,51200,0.11723289224836562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,16384,0.4567386839124892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,10240,0.039846221605936684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,8192,0.08453600274191962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,128,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,65536,0.0810826685693529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,768,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,12288,0.013813333378897773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,7168,0.08828177716996934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,2048,0.02565244502491421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,3072,0.16689956188201904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,51200,0.10427555772993301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,128,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,3072,0.03190222382545471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,768,0.007003555695215861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,4096,0.09731466902626885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,2560,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,8192,0.7846444447835287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,7168,0.03723288906945123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,65536,0.11777599652608235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,10240,0.07961689101325141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,4096,0.008714666797055138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,3072,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,5120,0.026220444175932143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,64,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,3072,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,512,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,65536,1.710784912109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,3072,0.02293955617480808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,32,0.008522666990756989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,768,0.003914666672547658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,6144,0.01566844516330295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,128,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,32,0.08014133241441515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,2560,0.016576889488432143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,256,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,7168,0.07100266880459256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,65536,0.07012355327606201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,65536,0.3597155676947699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,65536,0.9923813078138563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,256,0.016570667425791424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,7168,0.12830044163597956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,768,0.04684799909591675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,16384,0.10973422394858466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,768,0.01387111097574234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,2560,0.04519377814398872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,256,0.16669155491722953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,16384,0.04376711116896736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,256,0.02146400014559428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,4096,0.007469333708286285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,5120,0.01998222205373976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,32,0.012135110795497894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,2560,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,5120,0.03951111104753282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,3072,0.017998221847746108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,768,0.00563733321097162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,10240,0.12373333507113987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,2560,0.033923556407292686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,8192,0.03704266746838888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,5120,0.019588443968031142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,2560,0.07353066735797457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,256,0.005376889059940974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,5120,0.018772444791264005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,10240,0.05134222242567274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,5120,0.019331556227472093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,768,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,256,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,6144,0.03052266769938999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,1536,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,4096,0.015245333313941956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,10240,0.051869332790374756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,7168,1.0144408543904622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,512,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,3072,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,3072,0.009933333430025313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,4096,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,1024,0.002899555489420891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,8192,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,512,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,2048,1.2277439965142143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,6144,0.01016266644001007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,8192,0.026712000370025635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,3072,0.009935999910036722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,768,0.017847999930381775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,65536,1.5820835961235895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,10240,0.13427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,3072,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,8192,0.16657955116695827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,16384,0.018040888839297824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,2560,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,51200,0.05488622188568115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,512,0.16508177916208902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,6144,0.11438843939039443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,1536,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,2560,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,2560,0.019800888167487252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,256,0.03299644589424133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,512,0.012994666894276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,32,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,16384,0.16231732898288304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,512,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,3584,0.030260443687438965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,1536,0.03815911213556925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,3072,0.011694221860832639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,7168,0.08940000004238552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,4096,0.02609777781698439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,768,0.02104444470670488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,12288,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,2048,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,16384,0.21223910649617514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,2560,0.21868977281782362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,7168,0.03436000148455302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,768,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,512,0.003140444349911478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,51200,0.11099466350343491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,6144,0.08298844761318631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,1024,0.04353600078158908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,16384,0.04372355673048231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,10240,0.022284444835450914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,5120,0.01312355531586541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,128,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,51200,0.20050488577948677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,65536,0.6747635735405816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,7168,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,7168,0.13696355289883083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,65536,0.1279253297381931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,51200,0.23144533899095324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,8192,0.0314337776766883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,5120,0.024688889582951862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,2560,0.016530666086408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,1536,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,256,0.013845333622561561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,768,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,256,0.0034906665484110513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,64,0.002209777840309673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,1536,0.0038399998512532976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,8192,0.30376354853312176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,1024,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,256,0.01052266690466139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,51200,0.8172222243414985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,16384,0.20834222104814318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,3072,0.02477866742346022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,256,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,1024,0.005799110978841782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,2560,0.7669102350870768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,3072,0.02123466630776723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,2560,0.023706666297382776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,12288,0.018464888135592144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,5120,0.03413777881198459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,12288,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,512,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,1024,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,32,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,7168,0.0893262227376302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,32,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,8192,0.016650666793187458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,12288,0.01739911072784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,10240,0.013971555564138623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,128,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,1536,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,6144,0.19112888971964517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,2560,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,7168,0.01533333294921451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,7168,0.12803733348846436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,2560,0.007596444752481248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,256,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,512,0.022840888963805303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,4096,0.026339555780092876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,12288,0.033320890532599554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,256,0.010416000253624385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,6144,0.021669333179791767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,8192,0.010797333386209277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,5120,0.023856888214747112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,7168,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,65536,0.13055733839670816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,65536,0.5102204216851128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,16384,0.11205600367652045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,32,0.015061333775520325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,65536,1.3350275887383356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,128,0.007772444850868649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,6144,0.0648791127734714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,65536,0.1525768836339315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,5120,0.14315644900004068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,3584,0.9940773646036783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,6144,0.017315554949972365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,5120,0.06972177823384602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,4096,0.07608977953592937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,4096,0.008774222599135505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,2560,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,1024,0.04727200004789564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,5120,0.08287377489937676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,7168,0.028565333949195013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,10240,0.056440889835357666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,1536,0.24240265952216256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,16384,0.21182400650448271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,1536,0.004313777718279096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,4096,0.038366221719317965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,3584,0.03205688794453939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,6144,0.010142222046852112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,1024,0.020036444067955017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,65536,0.8659528626335992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,16384,4.346756405300565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,6144,0.01222400036123064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,3072,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,10240,0.053017778529061206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,2560,0.004950222041871813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,5120,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,16384,0.3742817772759332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,5120,0.009385777844323052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,32,0.0019075555933846368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,5120,0.5051102108425564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,51200,0.03817333446608649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,2560,0.018555555078718398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,10240,0.013233777549531726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,7168,0.043938666582107544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,7168,0.032221333848105535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,5120,0.017818666166729398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,768,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,1536,0.025263110796610515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,64,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,2048,0.0649413333998786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,8192,0.10601777500576443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,2048,0.028791109720865887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,3072,0.007559110720952352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,32,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,128,0.011901333100265928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,12288,0.016715556383132935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,1024,0.2920088768005371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,2560,0.16428266631232366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,2048,0.025607110725508794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,512,0.004308444344335132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,1024,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,3072,0.026241777671708003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,1536,0.103566222720676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,10240,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,3584,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,65536,0.44177156024509007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,4096,0.058760888046688504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,7168,0.020388444264729817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,32,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,3072,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,8192,0.02199288871553209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,8192,0.05381955703099569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,10240,0.02381955583890279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,3072,0.012311110893885294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,5120,0.019630221856964957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,512,0.018577777677112155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,4096,0.0805600020620558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,4096,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,4096,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,7168,0.010094222095277574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,7168,0.034272889296213783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,7168,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,3584,0.007768889268239339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,5120,0.01644533375898997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,10240,0.05413866705364651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,256,0.029186666011810303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,1024,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,7168,0.05166222320662605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,3584,0.023032888770103455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,65536,0.06060889032151964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,32,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,512,0.01459022197458479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,8192,0.018547554810841877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,51200,0.18386044767167833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,1024,0.020451555649439495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,512,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,3584,0.5500008794996474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,3584,0.15637600421905518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,1536,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,128,0.023785778217845496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,2560,0.33251110712687176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,8192,0.01200177768866221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,4096,0.05574044254091051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,128,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,1536,0.010608000059922537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,8192,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,256,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,10240,0.014379555980364481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,4096,0.033019555939568415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,512,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,1024,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,12288,0.08402933014763726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,51200,0.05823022127151489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,16384,0.1662177774641249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,6144,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,6144,0.22633510165744355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,256,0.02271911170747545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,256,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,12288,0.12762844562530518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,768,0.0038862224254343244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,65536,1.1899742550320096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,256,0.00699733363257514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,51200,0.11215199364556207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,64,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,64,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,6144,0.0092604441775216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,51200,0.02405511173937056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,2048,0.010413332945770687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,16384,0.23200178146362305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,6144,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,6144,0.008368888662921058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,5120,0.017887999614079792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,3584,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,12288,0.014101333088344999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,6144,0.05698133177227444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,65536,0.09514666928185357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,2048,0.010889777706729041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,3584,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,2560,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,3072,0.13937066660987005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,2048,0.016750221451123554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,6144,0.038161777787738375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,65536,1.4779644012451172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,512,0.048057778014077075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,32,0.008648888932334052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,5120,0.02014488809638553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,4096,0.015589333242840238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,10240,0.04099644554985894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,4096,0.02776088813940684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,256,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,32,0.16116266780429417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,6144,0.22401687833997938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,3072,0.02734577821360694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,12288,0.09328977929221259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,2560,0.018936889039145578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,256,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,12288,0.03530577818552653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,7168,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,51200,0.4291173352135553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,1536,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,512,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,10240,0.08156622118420072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,5120,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,2048,0.02773511078622606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,6144,0.03943555553754171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,128,0.012373333175977072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,2560,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,1024,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,12288,0.3188897768656413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,512,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,3072,0.025754666990704004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,51200,1.057077301873101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,65536,0.14860444598727757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,256,0.0053431110249625305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,3584,0.014162666267818876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,51200,0.04063733418782552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,3584,0.015274667077594332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,16384,0.3975004355112712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,6144,0.023759111762046814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,64,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,512,0.01000088867213991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,1536,0.013597333596812354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,4096,0.024400888217820063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,7168,0.011025778121418424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,4096,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,768,0.017265778448846605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,64,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,128,0.013376889129479727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,2560,0.03728355632887946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,65536,0.48611291249593097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,4096,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,2048,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,128,0.013823111024167804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,16384,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,65536,2.341775894165039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,65536,0.07712533076604207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,6144,0.10993866788016425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,65536,0.3489030996958415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,7168,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,768,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,3584,0.007855110698276097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,256,0.017669333351982962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,2048,0.03473866648144192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,51200,0.05168177684148153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,768,0.06497689088185628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,5120,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,3584,0.01315199997689989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,1536,0.00776533368561003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,5120,0.1540773312250773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,51200,0.0694648888376024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,5120,0.009395555489593083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,7168,0.009501333038012186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,8192,0.01651022169325087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,6144,1.576736026340061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,1536,0.005791999813583162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,512,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,7168,0.15498577223883733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,512,0.006835555450783835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,1536,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,128,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,12288,0.013552000125249227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,768,0.01039644413524204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,64,0.015442666080262927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,64,0.05088355475001865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,128,0.016381333271662395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,12288,0.889030244615343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,768,0.020593777298927307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,512,0.13196178277333578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,1536,0.031162665949927434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,2560,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,65536,0.18104178375667998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,2560,0.0350195566813151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,256,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,256,0.03341866532961527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,16384,0.11896177132924397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,8192,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,10240,0.01534577707449595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,4096,0.020184889435768127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,64,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,1024,0.06759199831220838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,3584,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,128,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,5120,0.011971555650234222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,6144,0.026081777281231348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,512,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,2048,0.13324622313181558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,6144,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,10240,0.03720622261365255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,4096,0.01763911048571269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,1536,0.02250755495495266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,16384,0.04400622182422214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,4096,0.04311199982961019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,4096,0.01668355531162686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,3072,0.06458133459091187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,2048,0.0395448903242747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,1024,0.017621333400408428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,5120,0.023994666006830003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,51200,0.9826372994316949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,4096,0.011874666644467248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,512,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,4096,0.008937777744399177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,7168,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,4096,0.0266311119000117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,4096,0.02879555688963996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,2048,0.01459111107720269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,64,0.001785777716173066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,128,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,256,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,2048,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,5120,0.05358666512701246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,12288,0.07029600275887384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,128,0.023976888921525743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,1024,0.006484444356626934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,1536,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,65536,1.2933538224962022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,256,0.011565333439244164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,4096,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,8192,0.014536889062987434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,8192,0.5905475616455078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,32,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,8192,0.24738311767578125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,128,0.00963288876745436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,512,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,3584,0.5283359951443142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,12288,0.018039999736679923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,2048,0.004594666676388847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,3584,0.030392001072565716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,3072,0.031748443841934204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,6144,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,1536,0.004019555532270008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,1024,0.0041911109454101985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,5120,0.05861600240071615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,1536,0.021892444954978094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,10240,0.015517334143320719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,32,0.03918577896224128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,1536,0.08175022072262235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,2048,0.012597333225939008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,12288,0.032933334509531655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,5120,0.03262933426433139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,512,0.02230577833122677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,32,0.021990221407678392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,512,0.011815999945004782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,5120,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,64,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,2560,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,2560,0.020967110991477966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,2048,0.021088000800874498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,16384,0.5077448950873481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,3584,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,3584,0.07230577866236369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,6144,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,1024,0.0034977777136696708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,3072,0.013832888669437833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,768,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,16384,0.25909778806898326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,65536,0.14542222023010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,3072,0.008119111259778341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,3072,0.02086044516828325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,5120,0.04448711209826999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,1536,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,51200,0.07163821988635592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,3584,0.013627555635240344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,8192,0.05927199787563748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,2560,0.012633778154850006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,10240,0.11311199929979111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,2048,0.004530666602982415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,4096,0.040095110734303795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,7168,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,768,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,1024,0.004031110968854693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,10240,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,16384,0.018304889400800068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,6144,0.018391999933454726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,12288,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,768,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,3584,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,1536,0.05482666691144308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,256,0.03887733485963609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,1024,0.005438222239414851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,12288,0.22021778424580893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,2560,0.024999999337726172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,10240,4.132784949408637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,2048,0.030247999562157526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,2048,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,7168,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,5120,0.1764871146943834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,10240,0.05604888995488485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,6144,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,12288,0.028219554159376357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,12288,0.06831022103627522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,2048,0.3471075428856744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,2048,0.004968889057636261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,10240,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,4096,0.12662400139702693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,4096,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,8192,0.14354666074117026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,8192,0.11248977979024251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,768,0.01690933273898231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,768,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,128,0.08097600274615817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,16384,0.3068017694685194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,65536,8.49723646375868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,51200,5.538934071858724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,2560,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,8192,0.03481066558096144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,51200,0.06649866369035509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,12288,0.01501511037349701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,12288,0.013682666752073499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,3072,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,1536,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,3072,0.08808177709579468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,128,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,768,0.032504889700147845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,512,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,1536,0.024412444896168176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,32,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,512,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,65536,0.08670310841666327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,2560,0.018744889232847426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,1536,0.008933333059151968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,3072,0.014141332772043018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,10240,0.018734221657117207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,256,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,51200,0.050415999359554715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,3584,0.018040888839297824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,4096,0.12171378400590684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,128,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,1536,0.04132089018821716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,3072,0.07855733235677083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,6144,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,256,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,7168,0.10280355479982163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,12288,0.030384888251622517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,3584,0.06267199913660686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,128,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,1536,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,1536,0.03860088851716783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,512,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,4096,0.022301332818137273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,10240,0.053027557002173535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,3584,0.04046044415897793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,10240,0.02873777680926853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,7168,0.14002133740319148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,7168,0.02864355511135525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,7168,0.05259733398755392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,10240,0.03561777869860331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,5120,0.031168888012568157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,5120,0.01257333325015174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,512,0.002560888934466574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,512,0.003535110917356279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,3072,0.022015111313925848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,7168,0.9840692943996854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,64,0.1588675578435262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,10240,0.08120266596476237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,5120,0.04061244593726264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,1024,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,128,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,3072,0.01906755566596985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,10240,0.019520888725916546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,512,0.08655289146635269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,12288,0.03439911206563314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,64,0.14467199643452963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,16384,0.02351377738846673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,16384,0.022430222895410325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,512,0.008930666579140557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,16384,0.13994844754536948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,64,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,512,0.05539733171463013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,5120,0.012471111284361945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,64,0.042584889464908175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,3072,0.012682666381200155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,1536,0.011301333705584208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,16384,0.03043466806411743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,2560,0.00700711127784517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,5120,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,51200,1.180580457051595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,12288,0.3393128977881537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,4096,0.02239822182390425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,51200,0.5543022155761719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,768,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,7168,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,1536,0.017210667332013447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,6144,0.17416621579064262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,256,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,6144,0.015538665983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,5120,0.021431111627154883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,2560,0.04509866568777296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,51200,0.050047109524408974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,8192,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,65536,0.17845333947075737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,64,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,3584,0.014955555399258932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,2560,0.03389599919319153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,4096,0.020590222544140287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,256,0.018774222996499803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,6144,0.011771555576059552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,3072,0.013615999784734515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,128,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,256,0.03758755657407973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,4096,0.00868888861603207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,512,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,65536,3.3101147545708547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,12288,0.10203466812769572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,51200,0.15165689256456164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,7168,0.014300444059901767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,65536,0.05267999900711907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,768,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,7168,0.042789333396487765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,16384,0.022125333547592163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,2048,0.0514977773030599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,4096,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,2048,0.027820444769329492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,8192,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,32,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,32,0.0029315555261241067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,3584,0.014576888746685453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,12288,0.18395466274685332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,51200,0.060079998440212674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,16384,0.021702221698231165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,16384,0.22804265552096895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,1536,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,1024,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,51200,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,65536,1.266241815355089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,16384,0.029924445682101782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,8192,0.1290444400575426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,64,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,2560,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,1024,0.021205332544114854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,8192,0.042188445727030434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,768,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,2048,0.02539288832081689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,12288,0.027890665663613215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,256,0.12229066424899632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,768,0.010670222342014313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,51200,0.09193333652284409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,16384,0.030701332622104224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,3584,0.0581608878241645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,512,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,65536,0.06684888733757867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,1024,0.0033573332346147964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,1024,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,768,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,7168,0.02921600143114726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,65536,0.6908702320522733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,5120,0.09147822194629246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,3584,0.017294221454196505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,1024,0.11881511741214329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,2560,0.012276444170210095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,6144,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,2048,0.005128889034191768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,256,0.00776444458299213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,2560,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,7168,0.03457955519358317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,6144,1.7329315609402125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,2560,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,256,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,65536,0.5033929083082411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,10240,0.1892951064639621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,7168,0.027541332774692114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,51200,0.06794222195943196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,3584,0.07088088989257812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,3584,0.017489777670966256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,3584,0.13780889246198866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,8192,1.0904835595024955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,2048,0.19753066698710123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,2048,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,32,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,2560,0.02157955533928341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,2560,0.011376889215575324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,65536,0.14124622609880236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,64,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,3584,0.06313244501749675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,51200,0.40743377473619247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,12288,2.343463897705078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,65536,1.1569929122924805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,256,0.01645155582163069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,768,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,1536,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,128,0.02015822297996945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,128,0.025451555848121643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,256,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,51200,0.035595556100209556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,4096,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,12288,0.04351644383536445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,64,0.08877155515882705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,256,0.07225066423416138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,32,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,128,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,12288,0.2132257752948337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,4096,0.08757599857118394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,65536,0.1271182166205512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,2048,0.021669333179791767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,3072,0.015604444675975375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,32,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,2048,0.01555022266176012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,10240,0.02165688905451033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,5120,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,256,0.005145777844720417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,8192,0.013592888911565145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,3584,0.015754666593339708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,5120,0.05544977717929416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,65536,0.2009751134448581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,128,0.015634665886561077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,2560,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,3584,0.02625688910484314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,10240,0.030472887886895075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,7168,0.014344000154071383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,2048,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,16384,0.025799110531806946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,5120,0.03693066702948676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,4096,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,2048,0.013329777452680798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,4096,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,2560,0.037083556254704796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,3584,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,51200,0.06592177682452731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,3584,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,51200,0.25779822137620717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,7168,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,32,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,768,0.15259377161661783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,128,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,768,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,512,0.0461733341217041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,512,0.014836443795098199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,512,0.005364444520738389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,12288,0.06169066826502482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,12288,0.01997422178586324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,32,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,768,0.01697688963678148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,51200,0.04026666614744399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,65536,0.16129333443111843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,256,0.025360888904995386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,16384,0.010209778116808997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,51200,0.2500115500556098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,65536,0.5826906628078884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,1024,0.30138932334052193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,256,0.011124444504578909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,1536,0.012158222496509552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,1536,0.5304364628261989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,512,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,1024,0.005848000033034219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,65536,0.937863137986925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,12288,0.019653333557976615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,7168,0.44360889328850645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,768,0.020067556036843192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,256,0.13701600498623318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,5120,0.02370400064521366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,128,0.005437333136796951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,512,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,5120,0.019541333119074505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,10240,0.012727999852763282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,3072,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,128,0.05553244551022848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,8192,0.019669334093729656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,7168,3.5519146389431424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,1536,0.028224001328150432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,3072,0.028024001253975764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,16384,2.6554266611735025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,32,0.013044444223244986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,128,0.008940444224410588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,65536,3.899640825059679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,65536,0.2679075664944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,5120,0.034443553951051496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,1536,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,6144,0.02843822373284234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,7168,0.0243848893377516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,768,0.008385777473449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,10240,0.0207360006040997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,2560,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,32,0.005312888986534542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,2560,0.020409777760505676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,256,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,2048,0.034839110241995916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,768,0.1420782274670071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,3584,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,65536,0.166101336479187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,51200,0.5937049123975966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,256,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,8192,0.022287110487620037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,6144,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,7168,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,3072,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,51200,0.05093155635727776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,512,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,8192,0.008672888908121321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,512,0.006519111080302133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,8192,0.026370666093296472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,16384,1.6544710795084636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,6144,0.06611200173695882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,12288,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,4096,0.7196239895290799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,512,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,3584,0.02637333340115017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,2560,0.017823111679818895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,32,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,65536,0.07685866620805529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,10240,0.04938755432764689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,3584,0.022629333866967097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,16384,0.026259554757012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,3072,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,10240,0.026894221703211468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,3072,0.011320888996124268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,3072,0.007844444778230455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,65536,0.410079108344184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,1536,0.018558222386572096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,32,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,1024,0.01629688839117686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,2560,0.46895647048950195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,128,0.01975644462638431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,768,0.006070222291681502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,256,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,65536,0.04893689023123848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,5120,0.011867555479208628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,12288,0.047084444099002414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,256,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,10240,0.0848204427295261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,7168,0.01423111061255137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,5120,0.018613333503405254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,256,0.028527110815048218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,256,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,3072,0.2341902256011963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,1024,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,1536,0.012723555167516073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,512,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,5120,0.06426222456826104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,8192,0.29709066285027397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,32,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,6144,0.0188755558596717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,1536,0.06292355722851224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,7168,0.008580444587601556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,1024,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,1024,0.0195777780479855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,128,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,16384,1.7787049611409504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,1024,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,2560,0.015857777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,128,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,6144,0.33041689130995006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,5120,0.2378044393327501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,128,0.015291555060280694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,51200,0.1323982212278578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,16384,0.020169778002632987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,7168,0.588330692715115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,3072,0.012864888542228274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,1024,0.02327466673321194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,128,0.00872444444232517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,65536,0.42339110374450684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,4096,0.018255111243989732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,16384,0.05759377612007988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,768,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,32,0.002402666749225722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,128,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,256,0.023415111833148535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,1536,0.146814227104187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,64,0.020444444484180875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,3072,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,512,0.0027946666296985415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,12288,0.08304266797171699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,3072,0.017817777064111497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,32,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,256,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,128,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,256,0.005752888818581899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,6144,0.014155555102560254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,256,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,2560,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,16384,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,3584,0.01666933298110962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,51200,13.950062221950956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,2560,0.010660444696744284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,512,0.018594667315483093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,16384,0.34023200141059023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,64,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,2560,0.17063732941945395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,2560,0.02272266646226247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,65536,0.14701422055562338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,1536,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,1536,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,768,0.17032710711161295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,768,0.017992888887723286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,65536,0.08014577627182007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,2560,0.09188977877298991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,256,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,1536,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,7168,0.025389333566029865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,256,0.10486044486363728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,6144,0.041760888364579946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,12288,0.1298560036553277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,65536,0.07711822456783719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,10240,0.30986666679382324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,32,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,128,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,5120,0.12856354978349474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,10240,0.10819022523032294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,3072,0.023307555251651343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,128,0.014513777361975776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,7168,0.01075111081202825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,6144,0.007479999628331926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,1024,0.0092257774538464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,2048,0.03904089000489976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,7168,0.007552000383536021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,1024,0.044347554445266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,512,0.02979733215437995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,512,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,5120,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,65536,0.15115643872155085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,1024,0.05210577779346042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,5120,0.12502488825056288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,64,0.0015786666837003494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,4096,0.010933332973056369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,51200,0.5091822412278917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,2048,0.01959911154376136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,5120,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,7168,0.021747555997636583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,65536,0.5475751029120551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,3072,0.06724444362852308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,3584,0.12751555442810059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,2560,0.01996977792845832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,4096,0.022484444909625586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,512,0.007679111427730984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,51200,0.02385422256257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,8192,0.047841777404149376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,32,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,65536,0.1564355558819241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,5120,0.009720889230569204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,768,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,7168,0.2858960098690457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,1024,0.026081777281231348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,2560,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,51200,1.7502222061157227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,3072,0.01908622185389201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,2560,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,768,0.009762666291660732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,128,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,128,0.021839999490314062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,512,0.006538666784763336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,128,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,12288,0.01721600029203627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,8192,0.011002666420406766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,4096,0.02986933456526862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,6144,0.037520888778898455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,2560,0.015648888217078317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,32,0.001791999985774358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,2560,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,12288,0.06141689088609484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,51200,0.06132444408204821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,768,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,5120,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,12288,0.03804088963402642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,3584,0.01811644434928894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,2560,0.013617777989970313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,10240,0.17728622754414877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,1024,0.0035662220584021676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,3584,0.0910151137246026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,8192,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,1024,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,512,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,12288,0.030815998713175457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,51200,0.06123377879460653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,7168,0.039467556609047785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,512,0.04165244433614943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,256,0.07033333513471815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,8192,0.04191111193762886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,6144,0.01535288823975457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,10240,0.618867556254069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,12288,0.8436977598402234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,128,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,8192,0.01573955516020457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,32,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,51200,0.133206221792433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,10240,0.04623288909594218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,12288,0.17987999651167128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,65536,0.07965777979956733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,2560,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,51200,0.04755022128423055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,65536,0.16643910937839085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,1536,0.03134844369358487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,1536,0.018562666243977018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,7168,0.02088266611099243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,7168,0.20455466376410592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,2560,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,1024,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,7168,0.009597332941161262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,2560,0.02097333305411869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,128,0.014504000544548035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,3584,0.0685173339313931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,64,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,3584,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,1536,0.027162666122118633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,3584,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,12288,0.5325146781073676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,512,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,6144,0.02258133391539256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,128,0.02539288832081689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,64,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,10240,0.05084088775846693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,65536,4.836194780137804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,1024,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,768,0.019559999306996662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,5120,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,7168,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,128,0.010620444185203977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,6144,0.22852712207370332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,1536,0.013175110850069257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,6144,0.2717928886413574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,4096,0.029483556747436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,2048,0.01128799964984258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,1536,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,12288,0.11241155200534397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,7168,0.027803556786643133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,3584,0.02696000039577484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,16384,0.01625777781009674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,768,0.01926577753490872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,512,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,6144,0.012489777472284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,1024,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,8192,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,51200,0.4853573375278049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,2560,0.02531733281082577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,3072,0.042774223619037204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,2048,0.006701333241330252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,65536,0.5530853271484375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,128,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,5120,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,32,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,51200,0.5182275772094727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,2560,0.015471999843915304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,51200,0.05325422353214688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,5120,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,4096,0.008453333543406593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,2048,0.03308355477121141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,6144,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,12288,0.01997599999109904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,3584,0.06140889061821831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,2560,0.3707662158542209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,7168,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,3072,0.06390933195749919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,5120,0.022054221895005967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,768,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,10240,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,12288,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,512,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,512,0.01218577805492613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,4096,0.01347644461525811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,1536,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,3584,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,64,0.0018088887963030075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,512,0.005419555637571547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,5120,0.03789688812361823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,768,0.015808888607554965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,512,0.025940444734361436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,1024,0.021704889006084863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,768,0.05362222260899014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,6144,0.02659911082850562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,1024,0.0229777776532703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,128,0.007559110720952352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,3072,0.030823111534118652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,4096,0.008795555267069075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,7168,0.047079109483295016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,6144,0.0731546680132548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,1536,0.06139466497633192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,16384,0.5213706758287218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,4096,0.02517066730393304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,8192,0.017049777838918898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,51200,0.05269689030117459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,5120,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,768,0.005370666583379109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,128,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,1024,0.02540622154871623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,12288,0.10723288853963216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,6144,0.2932648923661974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,5120,0.014333332578341166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,3072,0.031217777066760596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,5120,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,512,0.020337777005301583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,10240,0.4215119944678412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,16384,0.02604266670015123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,4096,0.011420444481902652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,7168,0.030397332376903955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,8192,0.023338667220539514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,2560,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,1536,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,4096,0.348857773674859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,3072,0.029659556017981634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,512,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,2560,0.022784888744354248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,6144,0.010441777606805166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,768,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,12288,1.077988412645128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,3584,0.04120355513360765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,12288,0.020040889581044514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,1536,0.16598221990797254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,2560,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,65536,0.29791556464301217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,768,0.012342222034931183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,768,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,10240,0.02870133188035753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,2560,0.016170667277442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,65536,0.07552978065278795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,256,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,3584,0.03747999999258254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,5120,1.0294550789727104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,768,0.02609866691960229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,768,0.008360000120268928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,512,0.02156266735659705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,12288,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,12288,0.05924089087380303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,51200,0.33873600429958767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,16384,3.9599261813693576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,256,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,128,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,4096,0.6652720239427354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,5120,0.02564711040920681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,1024,0.008932444784376357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,4096,0.02294755478700002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,5120,0.01182577759027481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,4096,0.01056177748574151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,256,0.01568000018596649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,8192,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,3072,0.04439022143681844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,32,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,256,0.05292355683114794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,2048,0.03493600090344747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,3072,0.011614222493436603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,65536,3.405800077650282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,16384,0.055664002895355225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,16384,0.2233546574910482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,3584,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,2560,0.02342133389578925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,51200,0.055000000529819064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,7168,0.5075955390930176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,4096,0.033225778076383806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,2560,0.19327467017703584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,65536,0.1664497719870673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,51200,0.13357067108154297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,51200,0.5503999922010634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,1536,0.02125599980354309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,12288,0.11657244629330105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,51200,0.11599911583794488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,6144,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,7168,0.42791467242770725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,1536,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,512,0.022328888376553852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,16384,0.0797413322660658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,10240,0.044052445226245456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,10240,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,65536,0.056051552295684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,6144,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,3584,0.007243555453088548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,1536,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,6144,0.21270755926767984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,2560,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,65536,0.0854053364859687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,5120,0.08453155226177639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,3072,0.027256000373098586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,16384,0.05806844764285617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,5120,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,768,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,51200,0.4873261981540256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,6144,0.012876444392734103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,8192,0.37218578656514484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,3584,0.011682666838169098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,51200,0.6737564404805502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,5120,0.028762665059831407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,512,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,6144,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,5120,0.1712391111585829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,3584,0.09039644400278728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,7168,0.6710444556342231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,12288,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,51200,0.0765591131316291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,65536,0.07196178038914998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,16384,0.008212444682916006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,16384,0.016869333055284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,6144,0.012381333443853589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,3072,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,3584,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,12288,0.2844853401184082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,65536,0.38085065947638613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,3072,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,51200,0.05212177832921346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,32,0.007823111282454597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,12288,0.18829245037502715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,8192,0.14914666281806097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,3072,0.018526222970750596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,10240,0.03857422206136916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,16384,0.047394666406843394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,7168,0.03383200036154853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,1024,0.0038373333712418876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,2048,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,4096,0.020587555236286588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,2560,0.11499022112952338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,12288,0.03397333290841844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,5120,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,2560,0.01865688959757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,16384,0.06286488638983832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,32,0.05498577819930183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,256,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,7168,0.03002133303218418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,10240,2.247400919596354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,7168,0.040631999572118126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,128,0.019002666076024372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,4096,0.27231910493638783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,128,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,1024,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,51200,0.014138667119873894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,512,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,65536,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,128,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,10240,0.02498400045765771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,51200,0.2572231027815077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,32,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,3584,0.19639644357893202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,768,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,51200,0.14537421862284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,65536,0.10083733002344768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,3072,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,12288,0.07140444384680854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,8192,0.026406221919589575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,8192,0.3944257895151774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,2560,0.044933332337273486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,16384,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,51200,0.05641688903172811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,2560,0.5382933086819118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,8192,0.0203155560625924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,128,0.015984889533784654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,512,0.0035004446076022256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,512,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,3072,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,3584,0.05102044343948364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,768,0.061287111706203885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,6144,0.14067288239796957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,7168,0.1294728914896647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,10240,0.0997048881318834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,128,0.017509332961506315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,768,0.01763288842307197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,6144,0.036534223291609026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,51200,0.050435556305779346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,768,0.013146667016877068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,1536,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,3072,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,128,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,32,0.003871110992299186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,5120,0.02988977895842658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,128,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,768,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,7168,0.06487288739946154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,5120,0.10860088798734878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,8192,0.044827553961012095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,768,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,65536,0.1253946622212728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,4096,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,16384,0.03122133347723219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,10240,2.4625814225938583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,2048,0.016913778252071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,8192,3.822479248046875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,5120,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,1536,0.028636445601781208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,4096,0.32094399134318036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,5120,0.04683822062280443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,3584,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,128,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,1536,0.008471999731328752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,32,0.14532444212171766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,512,0.0362026658323076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,4096,0.08533244662814671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,12288,0.013738666971524557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,51200,1.1963929070366752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,3072,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,12288,0.19327021969689262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,1024,0.005780444376998478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,3072,0.045921779341167875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,1024,0.01457244488928053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,7168,0.030270222160551283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,2048,0.027310222387313843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,3584,0.10010755724377102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,64,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,3584,0.041377776198916964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,2048,0.03802400165134006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,32,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,2048,0.4162408775753445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,3072,0.02570044497648875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,8192,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,12288,0.054020444552103676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,4096,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,16384,0.04853244291411506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,768,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,16384,0.03307822346687317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,1536,0.027664000789324444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,7168,0.03838488790724013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,512,0.00564444437623024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,768,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,7168,0.06065066655476888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,32,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,4096,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,1536,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,51200,0.05379377802213033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,16384,0.01568000018596649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,2048,0.16519378291236028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,2048,0.010762666662534079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,6144,0.06407022476196289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,128,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,6144,0.0313351121213701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,4096,0.04321777820587158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,65536,0.02221422228548262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,65536,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,2560,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,16384,0.032658666372299194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,128,0.012412444584899478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,10240,0.09228533506393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,256,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,2048,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,2560,0.45856534110175234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,768,0.01685333251953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,51200,0.1943920056025187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,65536,0.1727902226977878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,768,0.029297779003779095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,51200,0.1808640029695299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,16384,0.026383999321195815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,6144,0.010075555907355415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,512,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,768,0.01586844523747762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,512,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,12288,0.10989599757724339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,8192,0.04019111063745286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,12288,0.017204445269372728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,7168,0.04638133446375529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,2560,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,65536,1.4056000179714625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,51200,0.9870337380303277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,768,0.0042160000238153665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,768,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,512,0.008426666259765625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,4096,0.015265777707099915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,16384,0.04029600156678094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,1536,0.01531022290388743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,5120,0.01035199976629681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,6144,0.225600004196167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,16384,0.03125866585307651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,5120,0.03668977816899618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,6144,0.013982221484184265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,10240,0.01607111096382141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,3072,0.007480888730949826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,3584,0.24155378341674805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,51200,0.04615199897024366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,512,0.01756622228357527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,10240,0.02851466668976678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,3584,0.2871626747979058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,7168,0.008647110727098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,1536,0.008935111264387766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,16384,0.17440088589986166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,51200,0.05249599946869744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,2560,0.11699288421207005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,64,0.006089777582221561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,512,0.022614222433831956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,32,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,65536,0.02218399941921234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,6144,0.022088888618681166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,768,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,8192,0.10046488708919948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,5120,0.054970665110482105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,512,0.025612443685531616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,7168,0.1737884415520562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,12288,0.01864177816443973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,4096,0.4581777784559462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,256,0.020599111914634705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,2048,0.01717866626050737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,51200,0.16846133602990043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,4096,0.0140684445699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,512,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,7168,0.047107554144329496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,65536,0.13899821705288357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,4096,0.03152711192766825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,512,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,16384,0.019866666860050626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,512,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,1536,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,2560,0.11449510521358913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,65536,0.9223990970187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,12288,0.3853120009104411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,32,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,65536,2.9662560356987844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,51200,0.34911110666063094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,8192,0.016696888539526198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,1536,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,51200,0.06713955269919501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,65536,0.5198311275906032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,65536,0.12851999865637884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,128,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,256,0.006653333289755716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,51200,0.10310577683978611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,1536,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,768,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,4096,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,128,0.0053075556125905775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,1024,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,16384,0.020693333612547982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,5120,0.008628444539176093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,10240,0.015100444356600443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,6144,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,2048,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,65536,0.10015910863876343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,10240,0.17232267061869302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,6144,0.011377777490350934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,768,0.007354666789372762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,1536,1.2761386235555012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,65536,0.12587111526065403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,51200,0.40055378278096515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,2048,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,2560,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,64,0.0022595555831988654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,6144,0.011045333411958484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,12288,0.041202667686674334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,256,0.021166221963034734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,128,0.01799199978510539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,8192,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,512,0.003633777714437909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,1024,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,3584,0.01819022165404426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,256,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,65536,0.017874666386180453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,128,0.01517955627706316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,7168,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,8192,0.02533866630660163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,51200,0.12489599651760525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,256,0.016199999385409884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,6144,0.06819111108779907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,51200,0.12153600321875678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,1536,0.012795555922720166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,3584,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,10240,0.02415200074513753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,65536,10.267474704318577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,2560,0.01761511133776771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,7168,0.028757333755493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,5120,0.02479822271400028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,6144,0.052904887331856616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,3584,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,1536,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,768,0.016943110360039603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,1024,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,768,0.01662488944000668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,10240,0.014876445134480795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,16384,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,16384,0.07469066646364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,6144,0.3045297728644477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,6144,0.08351022005081177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,32,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,7168,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,768,0.00574311117331187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,2048,0.007223111059930589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,12288,0.07988533046510485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,256,0.0022986665782001284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,65536,0.09208177857928806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,256,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,32,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,64,0.030921777089436848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,4096,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,51200,0.17994222376081678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,64,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,4096,0.4242444568210178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,768,0.003436444534195794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,16384,0.02104355560408698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,6144,0.01145155562294854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,1024,0.017880888448821176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,10240,0.022952889402707417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,51200,0.06191555658976237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,128,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,7168,0.13674667146470812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,65536,0.06502399841944377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,1024,0.0034942221310403612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,5120,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,8192,0.4289697806040446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,1024,0.04622133241759407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,3072,0.09340800179375543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,6144,0.028007109959920246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,8192,0.06973333491219415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,51200,0.14858222007751465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,16384,0.05858400132921007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,3584,0.5313253402709961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,4096,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,7168,0.013869333598348828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,16384,0.016140444411171805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,2560,0.08556266625722249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,6144,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,128,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,2048,0.004592000196377437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,2560,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,4096,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,128,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,512,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,32,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,512,0.014641778336630927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,51200,0.04502933224042257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,8192,0.011517333487669626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,768,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,1536,0.019235556324323017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,3072,0.048397332429885864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,8192,0.023010666171709698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,128,0.0631564458211263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,8192,0.08591555224524604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,256,0.0282008879714542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,1024,0.011384888655609555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,32,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,16384,0.08479555447896321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,1024,0.05287999908129374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,65536,2.8365484873453775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,5120,0.0163093325164583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,8192,0.00942844483587477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,6144,0.026391999589072332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,8192,0.03940977652867635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,32,0.029790222644805908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,51200,0.32530755466885036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,768,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,3584,0.02680622206793891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,1024,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,1536,0.0288080010149214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,12288,1.7736115985446508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,6144,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,768,0.006163555714819167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,8192,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,5120,0.027152000202072993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,16384,0.02660266723897722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,3584,0.06744177473915948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,7168,0.020101333657900494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,7168,0.12491644753350152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,512,0.007330666813585494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,4096,0.0255031105544832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,1024,0.017543110582563613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,7168,0.010668444136778513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,1024,0.028188443846172754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,2048,0.0345164438088735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,3072,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,8192,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,768,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,3072,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,512,0.01405066748460134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,65536,0.13885866271124944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,65536,25.099965413411457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,2560,0.03185866607560052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,1024,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,6144,0.019931554794311523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,10240,0.09395111269421047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,16384,0.0579342246055603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,5120,0.03876711262596978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,2048,0.01442311041884952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,768,0.017553778158293832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,51200,0.04652000135845608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,768,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,1024,0.014379555980364481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,1536,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,6144,0.11724177996317546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,2560,0.02011644509103563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,2560,0.017839110559887357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,768,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,3584,0.03732000125779046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,10240,0.8232328626844617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,512,0.15903022554185656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,6144,0.06728266345130073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,10240,0.030403554439544678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,10240,0.01385244478782018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,2048,0.026157332791222468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,2560,0.035740445057551064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,1024,0.016894222961531747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,7168,0.13097155094146729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,128,0.014745778507656522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,2560,0.053416000472174756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,2560,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,3584,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,5120,0.28854044278462726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,7168,0.018567111757066514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,10240,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,2048,0.01074044406414032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,51200,0.22465867466396758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,6144,2.272452460394965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,6144,0.049599998527103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,12288,0.027552889453040227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,10240,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,8192,0.03460977805985345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,16384,0.2329031096564399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,12288,0.05990488661660088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,10240,0.8255884382459852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,4096,0.048488888475630015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,256,0.018609777092933655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,12288,0.013410666750537025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,6144,0.052658667167027794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,3072,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,51200,0.23303466373019746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,512,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,4096,0.021939555803934734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,16384,0.0337066650390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,1024,0.010583110981517367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,2560,0.0254711111386617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,10240,3.926964653862847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,12288,0.043507556120554604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,1536,0.038047999143600464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,5120,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,5120,0.01391999920209249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,3584,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,3072,0.031812445984946355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,6144,0.008448888858159384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,128,0.005761777775155173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,65536,0.17527377605438232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,3072,0.029383109675513372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,768,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,768,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,7168,0.8416079945034451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,3072,0.03541244400872125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,3072,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,7168,0.053230222728517324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,10240,0.0620595547888014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,768,0.006170666466156642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,256,0.05511377917395698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,256,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,8192,0.023382221659024555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,1536,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,256,0.0069884442620807225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,10240,0.9644382264879016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,1024,0.0211191111140781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,51200,0.8803546693589953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,12288,0.06086755461162991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,5120,0.01181688904762268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,2048,0.02088977727625105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,128,0.014500444134076437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,512,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,5120,0.06673244635264079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,32,0.001998222122589747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,32,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,10240,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,2048,0.011795555551846823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,12288,0.29308001200358075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,2560,0.014913777510325113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,768,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,3584,0.01185155577129788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,8192,0.13626489374372694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,6144,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,12288,0.04725689027044508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,65536,12.989165412055122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,7168,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,5120,0.1420231130388048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,2560,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,2048,0.025988444685935974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,6144,0.10773066679636638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,65536,0.07401955789989896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,2560,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,5120,0.030403554439544678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,12288,0.3064640098147922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,3584,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,6144,0.029952890343136255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,2560,0.04558933443493313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,16384,0.05204000075658163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,64,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,4096,0.018039110634062026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,768,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,2048,0.004354666504595014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,5120,0.05676266882154676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,12288,0.10280622376335992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,3072,0.04394311043951246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,2560,0.01884355478816562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,1536,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,3072,0.03009066647953457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,768,0.01791822248035007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,512,0.004567111117972268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,8192,0.19188443819681802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,1536,0.016840888394249808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,1024,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,10240,0.0122079998254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,2560,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,65536,0.12372355990939671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,64,0.07388889127307467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,128,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,768,0.18676088915930855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,3072,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,65536,0.05332533187336392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,12288,0.20026222864786783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,2560,0.07618933253818087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,128,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,256,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,12288,0.00850755555762185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,8192,0.06276000208324857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,2560,0.6420399877760146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,10240,0.029682666063308716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,1536,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,128,0.0195777780479855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,10240,0.019066666563351948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,3584,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,2560,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,5120,0.1172906690173679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,8192,0.018415111634466384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,512,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,2560,0.020766221814685397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,64,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,16384,0.10099822282791138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,4096,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,128,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,65536,0.050592889388402305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,16384,3.93088870578342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,7168,0.027782221635182697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,12288,0.015076445208655464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,1024,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,2048,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,2048,0.019558222757445443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,65536,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,256,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,6144,0.011785777906576792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,51200,0.4079306655459934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,6144,0.02667644454373254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,8192,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,1024,0.017571555243598092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,65536,0.058423108524746366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,128,0.009381333159075843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,128,0.005643555687533484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,10240,0.15502044889662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,2560,0.007134222322040134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,16384,0.02604266670015123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,768,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,5120,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,6144,1.132231076558431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,6144,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,768,0.02989333205752903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,256,0.04553688897026909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,5120,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,4096,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,7168,0.0645573337872823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,3584,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,3072,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,1024,0.03636888994110955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,32,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,2560,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,16384,0.020921778347757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,10240,0.008642666869693333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,6144,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,2048,0.019388443893856473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,10240,0.08845333258310954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,128,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,16384,0.34552354282803005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,51200,0.0414373344845242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,1024,0.014746665954589844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,1536,0.04823911190032959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,10240,0.030569778548346624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,65536,2.9426604376898875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,2048,0.27684177292717826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,2560,0.007535999847782983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,6144,0.3380684322781033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,8192,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,1024,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,5120,0.08579288588629828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,10240,0.4103662172953288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,64,0.0019431112127171622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,512,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,3072,0.015903110305468243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,4096,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,128,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,768,0.008240000241332585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,16384,0.3672782315148248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,768,0.03269333309597439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,512,0.1543004380332099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,2560,0.05294933252864414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,12288,0.047297779056761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,32,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,512,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,65536,2.3813902537027993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,16384,0.06864977545208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,128,0.00198755558166239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,512,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,16384,0.08637422323226929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,256,0.005385777602593104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,2560,0.654345777299669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,16384,0.018961777289708454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,5120,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,4096,0.106678220960829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,2048,0.029578665892283123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,51200,0.19559021790822348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,256,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,2560,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,32,0.003911111089918348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,128,0.020033778415785897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,3584,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,1536,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,65536,0.41255643632676864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,5120,0.24744621912638345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,64,0.09593155648973252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,7168,0.016357334123717416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,1536,0.016524444023768108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,65536,0.06468711296717326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,2560,0.017141333884663053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,1024,0.00654755574133661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,32,0.005671111245950063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,1536,0.019573332534896005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,6144,0.0669084456231859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,2048,0.01868355605337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,2048,0.02827199962404039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,8192,0.01348000019788742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,65536,0.05269244313240051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,512,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,5120,0.01923466722170512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,3072,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,3584,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,256,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,768,0.019860444797409903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,256,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,1536,0.037716444995668195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,512,0.015078221758206686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,512,0.01900622248649597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,6144,0.28662310706244576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,2560,1.278347545199924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,64,0.005623111294375525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,128,0.020901333954599168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,512,0.0074826669361856245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,51200,0.020780444145202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,65536,0.46205780241224503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,32,0.006529777828190062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,64,0.009461333354314169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,1536,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,768,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,2048,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,12288,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,1024,0.016025778320100572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,512,0.04720088839530945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,5120,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,3584,0.004855999930037392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,1536,0.4175866709815131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,16384,0.02597155504756504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,3072,0.0210515558719635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,51200,1.0517653359307182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,10240,0.03420889046457078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,7168,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,10240,0.08517333534028794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,128,0.017109332813156974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,32,0.014703111516104804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,16384,0.032207111517588295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,512,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,768,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,3072,0.058592001597086586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,2560,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,10240,0.11910311381022136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,32,0.10017510917451646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,768,0.025039111574490864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,65536,0.1444337765375773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,512,0.03085244364208645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,65536,0.9327981736924914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,256,0.0032826666202810076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,51200,0.09540710846583049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,4096,0.1574524508582221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,7168,0.03337600164943271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,1536,0.009888889061080085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,51200,0.08893955416149563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,2560,0.04565599891874525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,5120,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,51200,0.050203555160098605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,256,0.026625777284304302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,2560,0.02195288903183407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,2560,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,10240,0.0240000006225374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,3072,0.01366311146153344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,10240,0.26668087641398114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,256,0.025191111697090998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,2560,0.027201778358883325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,5120,0.03203644355138143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,512,0.08615377876493667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,12288,0.02071288890308804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,2048,0.10127288765377468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,512,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,10240,0.04283644424544441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,6144,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,16384,0.020407110452651978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,3072,0.009908444351620143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,1536,0.009115555220180089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,4096,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,12288,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,768,0.008711111214425828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,10240,0.0468773345152537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,64,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,2560,0.04683466752370199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,10240,0.02648177742958069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,2048,0.05189333359400431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,1536,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,2560,0.010327999790509542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,4096,0.01668444441424476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,1536,0.09927466842863296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,8192,0.024842666255103216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,7168,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,16384,0.054495112763510815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,256,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,2560,0.029490666257010564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,6144,0.25176800621880424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,6144,0.01919911139541202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,768,0.08934311072031657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,5120,0.031269331773122154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,10240,0.02753688891728719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,51200,0.05490488807360331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,5120,0.022086222966512043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,1024,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,256,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,5120,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,2560,0.06574400266011556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,12288,0.03677155574162801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,4096,0.011005333728260465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,51200,0.4456471337212457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,128,0.028939555088678997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,768,0.01614755557643043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,8192,0.030511998467975195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,8192,0.04040711124738058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,2560,0.004208000169859992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,2048,0.06832177771462335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,65536,0.0891653365559048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,256,0.11217333210839166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,5120,0.02184088859293196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,1536,0.043235556946860425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,5120,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,768,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,64,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,12288,0.1165884468290541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,512,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,7168,0.08211911386913724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,4096,0.011145778000354767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,65536,0.06327111191219754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,10240,0.02358755634890662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,1536,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,16384,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,5120,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,32,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,128,0.04217955470085144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,3072,0.02581688927279578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,51200,0.16191289159986708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,10240,0.012095111111799875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,65536,0.3311413394080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,128,0.02421066661675771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,65536,0.22093955675760904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,5120,1.373555607265896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,1536,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,4096,0.012330667012267642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,768,0.027747554911507502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,3072,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,2048,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,10240,0.03578133384386698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,1024,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,32,0.09333688682980007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,12288,29.56241183810764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,3584,0.01495911180973053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,6144,0.030105776256985132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,7168,0.014717333846622042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,128,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,3584,0.02086488902568817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,512,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,10240,0.26744090186225045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,12288,0.08598577976226807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,51200,0.2374995549519857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,51200,0.05981510877609253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,12288,0.01368444412946701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,16384,0.020964443683624268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,256,0.03653511073854234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,64,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,16384,0.09626844194200303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,3072,0.03126400046878391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,512,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,2048,0.004221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,8192,0.1443315611945258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,4096,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,512,0.16717510753207734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,2560,0.03846755623817444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,32,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,10240,0.02609688871436649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,6144,0.013214222258991666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,5120,0.010000000397364298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,512,0.028119110398822363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,64,0.025182222326596577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,32,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,64,0.03443911009364658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,64,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,2048,0.007484444313579136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,6144,0.1778577831056383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,8192,0.17537777953677705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,768,0.7668462329440646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,768,0.009033777647548253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,512,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,8192,0.028985778490702312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,2560,0.021653332644038733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,3072,0.03037066592110528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,3584,0.198052446047465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,1536,0.021946666969193354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,8192,0.006719999843173557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,64,0.06823911269505818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,7168,0.09155111180411445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,256,0.011786667009194693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,2560,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,4096,0.026365333133273657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,16384,0.017257778180970084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,12288,1.695560031467014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,256,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,512,0.09510311153199937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,1024,0.2869395679897732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,65536,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,512,0.02777333392037286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,10240,0.15189422501458064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,6144,0.026476444469557867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,3584,0.02898755669593811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,1536,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,2048,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,32,0.01592266725169288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,2560,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,2048,0.11716711521148682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,32,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,32,0.0016942221878303422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,1536,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,4096,0.008216888540320927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,4096,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,1536,0.019641778535313077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,16384,0.027199111051029626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,51200,1.0138764911227756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,7168,0.6595786412556967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,2560,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,16384,0.05058044526312086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,128,0.01408266690042284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,128,0.023307555251651343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,12288,0.016167999969588388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,1024,0.018790221876568265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,1536,0.026924444569481745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,51200,0.2345893383026123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,16384,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,2048,0.020819556381967332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,12288,0.11032533645629883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,5120,0.01830044388771057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,3584,0.03158400124973721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,3584,0.03496088915401035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,51200,0.20501332812839082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,3072,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,16384,0.0255031105544832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,3584,0.1956319941414727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,10240,0.01202400028705597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,32,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,3584,0.023137778043746948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,1536,0.018426666657129925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,2048,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,32,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,512,0.015395555231306287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,128,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,12288,0.11354311307271321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,512,0.08570844597286648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,512,0.00244711111817095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,51200,1.0636986626519096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,5120,0.07789333661397298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,4096,0.794872866736518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,3584,0.03797511259714762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,5120,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,65536,0.27852622667948407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,51200,0.06633689006169637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,51200,0.2165128919813368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,65536,0.32511289914449054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,6144,0.11822044849395752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,10240,0.04230133361286587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,6144,0.031446221801969744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,51200,0.03191466795073615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,512,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,3584,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,12288,0.017859554953045316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,51200,6.528714497884114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,2560,0.007346666521496243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,3584,0.015899555550681222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,32,0.07433688640594482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,10240,1.7117724948459203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,16384,0.026770666241645813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,2048,0.022983999715911016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,2048,0.042358222934934825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,32,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,768,0.06555200285381742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,32,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,16384,0.19910400443606904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,4096,0.12240266799926758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,65536,0.884768909878201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,6144,0.03608533408906724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,1024,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,768,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,5120,0.03052177694108751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,2048,0.029143999020258587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,256,0.033923556407292686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,256,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,1024,0.03462222218513489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,10240,0.031207111146714952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,4096,0.03735111157099406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,768,0.003485333174467087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,2048,0.011491555306646558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,2560,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,16384,0.04722399844063652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,12288,0.030809776650534734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,10240,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,10240,0.38760799831814235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,3584,0.03800977932082282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,10240,0.04027555717362298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,65536,0.6423519982231988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,32,0.024313777685165405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,64,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,3072,0.04373777906099955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,12288,0.09077955616845025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,3584,0.029007110330793593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,6144,0.03680177860789829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,51200,0.06751377715004815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,12288,0.018887110882335238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,768,0.014663111832406787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,1536,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,256,0.002226666650838322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,128,0.023010666171709698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,1536,0.018384888768196106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,512,0.005655999812814925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,7168,0.0418177776866489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,768,0.016029333074887592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,768,0.0373564428753323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,7168,0.030400001340442236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,16384,0.2087848848766751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,1536,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,51200,0.12337156136830647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,65536,0.3395635551876492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,8192,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,3584,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,2048,0.01902577777703603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,5120,0.05367822117275662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,10240,0.03458666801452637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,1536,0.011903111305501727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,16384,0.384185791015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,2048,0.023707555400000677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,12288,0.033801777495278254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,3584,0.007320000065697565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,10240,0.25054311752319336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,6144,0.06670044528113471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,256,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,5120,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,12288,0.014259555273585849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,512,0.033166223102145724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,7168,0.02055644492308299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,128,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,16384,0.09808978107240464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,6144,0.03736799955368042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,3072,0.0332622230052948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,16384,7.108744727240668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,3072,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,64,0.004239999999602635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,12288,0.015493333339691162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,1536,0.011948444777064852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,7168,0.03800977932082282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,7168,0.010060444474220276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,128,0.0022053333620230355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,256,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,3584,0.006887110984987683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,256,0.002542222125662698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,10240,0.06577689117855497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,65536,0.5462657610575358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,256,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,51200,1.086806191338433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,768,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,4096,0.016351110405392118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,768,0.021706667211320665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,3584,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,3584,0.024156444602542456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,6144,0.3507546583811442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,256,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,7168,0.026927111877335444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,3584,0.03007022208637661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,1024,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,32,0.007440889047251807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,256,0.024567999773555334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,16384,0.15779733657836914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,16384,0.020211555891566806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,6144,0.06005599763658312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,2048,0.02404533326625824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,7168,0.07880889044867621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,768,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,1024,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,6144,0.024569777978791132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,32,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,10240,0.08159822225570679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,7168,0.020438222421540152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,1536,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,8192,0.011198221809334226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,1024,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,2560,0.08878666824764675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,3072,0.013389333254761167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,65536,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,512,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,512,0.006816888848940532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,5120,0.008270222279760573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,16384,0.07710578044255574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,65536,0.11577155854966904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,6144,0.04589510957400004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,1024,0.034507556094063654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,1536,0.003842666745185852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,5120,0.017391110459963482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,4096,0.14531732930077448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,5120,0.011984888878133563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,3072,0.08928000264697605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,1536,0.004401777767472797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,128,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,10240,0.016026667422718473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,768,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,3072,0.08248977528678046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,128,0.014885332849290637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,4096,0.030870222383075293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,2048,0.007162666983074612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,2560,0.020121778051058452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,6144,0.018194667167133756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,256,0.1038844452963935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,51200,0.1224479940202501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,32,0.0033315554675128725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,51200,0.2373208999633789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,128,0.015065777632925244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,1536,0.010279999838935005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,1536,0.06737422280841403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,2560,0.02271733350223965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,1024,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,64,0.0019431112127171622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,4096,0.35270577006869847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,1536,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,2048,0.014104000396198697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,16384,1.6023324330647786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,3584,0.015660444895426433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,2048,0.006124444305896759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,32,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,6144,0.20640177196926543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,768,0.022587555978033278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,51200,0.04895022180345324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,5120,0.017161778277821012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,65536,2.6843155754937063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,64,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,12288,0.6715857717725965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,512,0.019077334139082167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,64,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,2560,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,51200,0.6600391070048014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,5120,0.016229333149062265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,65536,0.04202933443917168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,512,0.012420444852775998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,10240,0.07409599754545423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,1536,0.018872888551818002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,12288,0.018548443913459778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,2560,0.03295466634962294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,10240,0.020907556017239887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,3584,0.1444222264819675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,4096,0.11501955986022949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,1024,0.02528000043498145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,12288,1.1484595404730904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,64,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,8192,0.019496000475353666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,51200,0.10115466515223186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,768,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,6144,0.017864889568752713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,768,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,65536,0.06448533137639363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,2048,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,2048,0.0041751112374994485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,128,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,512,0.005714666512277391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,3584,0.02509244448608822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,3584,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,16384,0.041950222518708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,16384,0.028247998820410833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,3584,0.012663111090660095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,128,0.013209777573744455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,1024,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,64,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,128,0.03255555695957608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,128,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,65536,1.8042034573025172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,3584,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,65536,4.9463929070366754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,768,0.06194666359159681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,2560,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,4096,0.029783109823862713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,8192,0.02169599963559045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,7168,0.01200977795653873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,3072,0.08472177717420791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,2560,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,512,0.005670222143332164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,51200,0.2470560073852539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,16384,0.023852444357342188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,1024,0.03966666592491998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,12288,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,256,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,7168,0.020612445142534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,5120,0.017222222354676988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,128,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,1024,0.039510223600599505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,3072,0.0818471113840739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,128,0.06427022483613756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,16384,0.13718310991923013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,3584,0.4902026918199327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,128,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,32,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,3584,0.007896000312434303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,8192,0.0495377779006958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,7168,0.05710311068428887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,8192,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,10240,0.03360177742110358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,51200,0.027263111538357202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,7168,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,16384,0.06378577815161811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,256,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,16384,0.19267733891805014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,4096,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,512,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,3072,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,7168,0.12369333373175727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,6144,0.06222489145067003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,64,0.006853333363930385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,10240,0.38523822360568577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,8192,0.019259555472267997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,7168,0.10087111261155869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,51200,0.6025013393825954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,512,0.016541333662139047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,512,0.011570666399266986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,512,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,1536,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,51200,3.9870834350585938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,1024,0.012520000338554382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,12288,1.7835138108995225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,5120,0.03254666593339708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,12288,0.017306667235162523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,12288,0.01977155605951945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,16384,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,8192,0.15039200252956814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,2560,0.32951733801099986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,6144,0.01554666625128852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,768,0.03298399845759074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,1024,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,3584,0.025801777839660645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,16384,0.20111378033955893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,256,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,768,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,1536,0.04991199904017978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,5120,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,5120,0.02948444419436985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,512,0.025659556190172832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,8192,0.016473778420024447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,10240,0.012785777449607849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,12288,0.03386666708522373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,3072,0.009401777552233802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,128,0.012630222572220696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,12288,0.3413120110829671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,16384,0.04987466666433546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,8192,0.07388622230953641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,3584,0.02701866626739502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,768,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,12288,0.018577777677112155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,51200,0.049385776122411094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,4096,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,7168,0.16343911488850912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,4096,0.04165777895185683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,10240,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,512,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,12288,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,3072,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,128,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,4096,0.056700441572401256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,32,0.005852444304360284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,1024,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,1024,0.022783999641736347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,768,0.023914666639433965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,7168,0.0709857808219062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,65536,1.397481812371148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,128,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,768,0.0036968886852264404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,12288,0.6488897535536025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,7168,0.018796443939208984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,256,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,256,0.020269332660569083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,5120,0.05591289202372233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,8192,0.013712888790501488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,2560,0.04875822199715508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,10240,0.1554444498485989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,5120,0.01795288920402527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,1024,0.01812622282240126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,3584,0.02930755416552226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,2560,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,12288,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,3072,0.013707555830478668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,65536,0.06632355848948161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,1536,0.07022666931152344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,64,0.013906665974193148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,6144,0.016419554750124615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,6144,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,3584,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,1536,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,4096,0.10566311412387425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,12288,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,1024,0.006671999891599019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,512,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,256,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,65536,0.06428622537189059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,2048,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,3584,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,65536,0.22884088092380098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,2560,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,32,0.0027555556346972785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,1024,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,4096,0.019638222124841478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,51200,0.021993777818149988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,1024,0.017454221844673157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,4096,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,7168,0.012837332983811697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,6144,0.009375111096435124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,4096,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,512,0.0026071110947264563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,51200,0.12784266471862793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,3584,0.015063110325071545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,16384,0.05318133367432488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,6144,0.0706968903541565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,65536,0.07305777735180326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,1536,0.05400977863205803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,128,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,16384,0.04641511042912801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,10240,0.08808710840013291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,16384,0.04102755586306254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,128,0.011724444727102915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,7168,0.024680000212457445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,1536,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,2560,0.006792888873153263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,10240,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,65536,1.4680017895168727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,51200,0.35272977087232804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,12288,0.8779662450154623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,16384,0.03786133395300971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,51200,0.2067751089731852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,1024,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,3072,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,2560,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,128,0.010552888943089379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,2560,0.031609776947233416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,512,0.030613332986831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,8192,0.030151999659008447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,768,0.03071555495262146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,128,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,5120,0.009116444322797988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,3072,0.022690667046440974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,7168,0.029806223180558946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,768,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,2048,0.034267554680506386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,256,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,10240,0.04542577928966946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,1536,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,10240,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,51200,0.13160178396436903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,6144,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,1024,0.005967999911970562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,2048,0.028421332438786823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,2560,0.018882667024930317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,51200,1.8202595180935328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,51200,2.0525040096706815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,3584,0.051029334465662636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,5120,0.06757244136598375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,2048,0.02840177880393134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,768,0.059655997488233775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,8192,0.020595555504163105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,51200,0.2711128923628065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,16384,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,51200,0.04041599896219041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,8192,0.03036266565322876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,12288,0.014644443988800049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,6144,0.06607288784450956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,256,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,2560,0.008887999587588841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,3072,0.02569777766863505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,51200,0.03827910953097873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,32,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,2560,0.007247111035717859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,3072,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,2048,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,1024,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,4096,0.02944444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,6144,0.026968000663651362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,3584,0.01862044466866387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,5120,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,1024,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,16384,0.06464178032345243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,3072,0.033505777517954506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,3584,0.022639999787012737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,2560,0.007679999702506595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,16384,0.04311111238267687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,1536,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,51200,0.11508888668484157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,8192,0.01406044430202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,32,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,3584,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,64,0.035387555758158364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,51200,0.07466310924953885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,6144,0.011620444556077322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,5120,0.31456976466708714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,8192,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,65536,0.4439111020829942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,64,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,3072,0.0690222250090705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,2560,0.022295110755496558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,1536,0.00923822240697013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,768,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,128,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,512,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,8192,0.011423110961914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,256,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,7168,0.14067644543117946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,32,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,4096,0.02390755547417535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,5120,0.10184977451960246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,128,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,1536,0.0032826666202810076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,5120,0.018981332580248516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,768,0.005445333404673471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,32,0.2820168866051568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,3072,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,2048,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,51200,10.309437221950954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,1536,0.012114666402339935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,1024,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,512,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,51200,0.10136533445782131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,1024,0.014269333746698169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,12288,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,16384,0.06857688559426202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,16384,0.04117600123087565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,12288,0.04881866772969564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,3072,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,5120,1.0637324651082356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,16384,0.018047110901938546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,12288,0.09886666801240708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,1536,0.13229778077867296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,512,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,7168,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,16384,0.06623822450637817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,16384,0.03986755675739712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,4096,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,7168,0.025107555919223364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,65536,0.5855395529005263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,768,0.041477332512537636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,64,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,1024,0.022574222750133936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,2048,0.015936000479592215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,2560,0.023665777511066858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,65536,0.0767315559917026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,6144,0.05502577622731527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,256,0.01463733282354143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,2048,0.014113777213626437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,6144,1.8007653554280598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,65536,0.6709395514594184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,65536,2.4552480909559464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,2560,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,12288,3.9209378560384116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,64,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,6144,0.1828835540347629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,51200,0.2112720012664795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,16384,0.30494933658176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,51200,0.0689893364906311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,16384,0.01772711177666982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,256,0.01891644464598762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,768,0.007368889119890001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,51200,0.06044533517625597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,12288,0.09131644169489543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,4096,0.05775644381841024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,10240,0.06984621948666044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,2048,0.032058666149775185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,51200,0.34270310401916504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,4096,0.20386221673753527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,10240,0.03308799862861633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,512,0.010626667075686984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,6144,0.01146133326821857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,51200,1.1012052959865994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,768,1.2208933300442166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,12288,0.17972444163428414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,5120,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,3584,0.04391111267937554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,12288,0.18001777595943877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,4096,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,5120,0.06128799915313721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,3072,0.11319822735256618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,16384,0.02323466704951392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,10240,0.05252444412973192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,4096,0.028746667835447524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,2560,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,512,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,8192,0.19562666945987275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,2048,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,3584,0.2056959999932183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,5120,0.010605333579911126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,1024,0.004176889028814104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,4096,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,6144,0.009720000127951304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,3072,0.03160088923242357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,3072,0.018064889642927382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,256,0.0028702221396896574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,256,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,65536,0.10692622264226277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,2048,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,6144,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,10240,0.02882666720284356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,4096,0.012839111189047495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,1024,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,3584,0.00434755575325754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,32,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,4096,0.020000888241661918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,1536,0.05890044238832262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,12288,0.23828000492519805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,1536,0.03397333290841844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,12288,0.5562933286031088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,16384,0.031070222457249958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,768,0.0042133331298828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,768,0.0181688881582684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,16384,0.1806631088256836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,3584,0.04469244347678291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,3072,0.06156177653206719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,6144,0.04110755523045858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,10240,0.10108266936408149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,16384,0.1759271091885037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,16384,0.30797600746154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,512,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,1536,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,768,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,1024,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,12288,0.03426577647527059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,51200,0.053363555007510714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,128,0.00699733363257514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,1536,0.028120001157124836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,3072,0.02176888949341244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,2048,0.04917066627078586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,1024,0.021864889396561518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,2560,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,6144,0.06308088699976604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,1024,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,3584,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,64,0.005033777819739448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,7168,0.041482667128245033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,1024,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,1536,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,4096,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,2048,0.01348533315791024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,12288,0.14405600229899088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,512,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,16384,0.018002667360835604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,768,0.01754399968518151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,64,0.002789333255754577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,3584,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,256,0.018831110662884183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,512,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,1024,0.031756444109810725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,3072,0.26033422682020396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,128,0.024576000041431848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,12288,0.10148533185323079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,8192,0.09727111127641465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,8192,0.018580444984965853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,65536,0.06576355298360188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,6144,0.019482667247454327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,5120,0.06622044245402019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,512,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,8192,0.03243822190496657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,3072,0.017667555146747164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,2560,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,16384,0.4062799877590603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,768,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,10240,0.5302177535163032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,3072,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,1536,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,6144,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,6144,0.026664889521068994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,16384,0.05511733227305942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,512,0.005741333381997214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,3584,0.05258488986227247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,5120,0.021887999441888597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,768,0.017163554827372234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,2048,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,16384,0.11820799774593777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,128,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,768,0.04222666554980808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,1024,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,65536,1.8527083926730687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,2048,0.15914311673906115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,2560,0.017872888180944655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,12288,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,6144,0.03838933176464505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,256,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,6144,0.013304000099500021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,128,0.015171556009186639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,32,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,768,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,768,0.08689866463343303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,16384,0.33458489841885036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,8192,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,7168,0.06212177541520861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,12288,0.037992000579833984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,256,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,12288,0.966865751478407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,256,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,6144,0.03333866596221924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,128,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,10240,0.10778933101230198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,65536,0.14731733004252115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,8192,20.19500562879774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,128,0.13255111376444498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,65536,3.575445387098524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,5120,0.23239021831088594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,3072,0.01498133275243971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,2048,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,512,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,256,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,128,0.012178666889667511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,3072,0.02988622254795498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,4096,0.00869599978129069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,5120,0.016795555750528973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,256,0.012368888490729861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,2048,0.05084266596370273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,6144,0.4468266699049208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,128,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,8192,0.018601778480741713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,65536,0.16707021660274932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,6144,0.029567999972237483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,3584,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,12288,0.03403733174006144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,2048,0.03513333201408386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,65536,0.017507554756270517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,51200,0.8259315490722656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,5120,0.03918044434653388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,32,0.0063324446479479475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,128,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,8192,0.031324442889955305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,6144,0.016558221644825406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,128,0.037428445286220975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,3584,0.018939556346999276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,768,0.0028008888992998335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,64,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,512,0.11638755268520778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,1536,0.02941511074701945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,4096,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,7168,0.12207378281487359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,3072,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,256,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,5120,0.01850133306450314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,32,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,4096,0.02968533171547784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,1536,0.013479111095269522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,2560,0.023623110519515142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,768,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,2560,0.027238221632109746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,16384,0.01075644459989336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,32,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,3584,0.04477333360248142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,6144,0.033497777250077986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,3072,0.07654222514894274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,6144,0.32778043217129177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,3072,0.016010666886965435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,1536,0.07651022407743666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,51200,2.5940924750434027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,65536,13.191144307454428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,2560,0.16697955131530762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,51200,0.06821599933836195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,7168,0.12775554921891955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,3072,0.03643466697798835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,32,0.006855999843941794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,51200,14.914087931315104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,7168,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,3072,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,16384,0.018560889694425795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,16384,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,3072,0.03490044342146979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,2560,0.0075804442167282104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,4096,0.09735911422305638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,7168,0.02152266601721446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,1536,0.028304000695546467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,8192,0.1213795608944363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,4096,0.06905333201090495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,51200,0.5427733527289497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,1024,0.021613332960340712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,5120,0.11115022500356038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,1024,0.02713866697417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,7168,0.03917510973082648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,51200,0.7137022018432617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,51200,0.17876088619232178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,12288,0.061257779598236084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,3072,0.01925511161486308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,64,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,51200,0.07626489135954115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,1024,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,6144,0.024728889266649883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,128,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,32,0.007872888611422645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,1536,0.003519999898142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,12288,0.112500442398919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,3072,0.05785955323113335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,6144,0.16610222392612034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,65536,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,512,0.09126844671037461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,12288,0.020855110552575853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,256,0.03215466605292426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,128,0.0021235555824306277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,768,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,64,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,4096,0.1769191159142388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,8192,0.016581333345837064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,5120,0.040229333771599665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,256,0.005296000176005893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,12288,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,4096,0.035840001371171736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,6144,0.03836799992455377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,2560,0.017549332645204332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,3584,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,3584,0.007711999946170383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,8192,0.04080977704789903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,12288,0.21669688489702013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,12288,0.051107555627822876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,3072,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,7168,0.09704889191521539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,7168,0.019928000039524503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,12288,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,64,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,512,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,2048,0.05144177873929342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,2048,0.004358222087224324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,5120,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,3584,0.005970666805903117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,3072,0.008112000094519721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,8192,0.011170667078759937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,128,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,16384,0.014897776974572076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,7168,0.08882400062349106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,768,0.02442399991883172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,1024,0.0037342223028341928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,12288,0.013398222625255585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,10240,0.05195288856824239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,8192,0.02535999980237749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,1536,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,7168,0.1254835526148478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,1536,0.03816355599297418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,2560,0.10624800125757854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,5120,0.07766399780909221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,8192,1.2347449196709526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,2560,0.09764089186986287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,768,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,65536,0.15953866640726724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,8192,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,1024,0.07197599940829806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,3072,0.05594755543602837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,12288,0.021728888154029846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,768,0.006396444307433234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,7168,0.018946665856573317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,256,0.02862311071819729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,512,0.0042204442951414315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,8192,0.028705779049131606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,16384,0.03811111052831014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,256,0.011678222152921887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,12288,0.030926220946841772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,1024,0.015611555841233997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,10240,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,1536,0.004671999977694618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,8192,0.07091022200054593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,10240,0.06506844361623128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,2048,0.01277599980433782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,6144,0.01552977826860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,1024,0.011490667031870948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,4096,0.03870222303602431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,32,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,16384,0.03526044554180569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,512,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,2560,0.016591111818949383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,1024,0.08141866657469007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,128,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,2560,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,51200,0.056127998563978404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,10240,0.016688000824716356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,8192,0.027854220734702215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,256,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,2560,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,65536,0.15572444597880045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,1536,0.02776088813940684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,4096,0.031075557072957356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,3584,0.09161866373485988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,3584,0.06786577569113837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,12288,0.01899377836121453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,1536,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,5120,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,16384,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,1024,0.01148888882663515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,2560,0.024889777104059856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,2560,0.034689777427249484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,3584,0.013312888642152151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,3584,0.0105262224872907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,128,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,1536,0.06781244277954102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,5120,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,256,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,512,0.01761155492729611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,1024,0.055247111452950366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,10240,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,64,0.0018835556175973679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,16384,0.018203554881943595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,3072,0.03638222151332431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,7168,0.010866666833559671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,512,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,2048,0.011208000282446543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,6144,0.05888355440563626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,2048,0.24190044403076172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,1536,0.00462755560874939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,4096,0.021147555775112573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,10240,0.22624354892306855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,3584,0.034119112624062434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,3584,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,51200,0.17174222734239367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,3072,0.013050666285885705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,7168,0.06613866488138835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,2048,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,128,0.005494222044944763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,7168,0.07761866516537137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,3072,0.0075457774930530125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,3072,0.030936890178256567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,4096,0.026776888304286536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,2048,0.01754044493039449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,8192,0.03456266721089681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,3072,0.022987556126382615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,512,0.014845333165592618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,3072,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,16384,0.0457839998934004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,1024,0.01760444376203749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,4096,0.06876266664928861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,51200,0.4257226785024007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,12288,0.03601600064171685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,1024,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,51200,0.13466310501098633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,3072,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,3584,0.035678221119774714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,3584,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,10240,0.22177333301968047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,3072,0.017234666479958426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,10240,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,16384,0.03271555569436815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,4096,0.027638221780459087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,16384,0.47979646258884007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,128,0.019545776976479422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,12288,1.4538240432739258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,1536,0.2777022255791558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,3072,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,8192,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,10240,0.04267200165324741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,4096,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,256,0.019541333119074505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,16384,0.3355084525214301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,7168,0.051813334226608276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,12288,0.040046221680111356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,7168,0.014867555763986377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,10240,0.1324453353881836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,4096,0.007117333511511485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,64,0.0034764446318149567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,256,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,256,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,3584,0.14585067166222468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,64,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,2048,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,1536,0.012845333251688214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,3072,0.10087200005849202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,1024,0.016009777784347534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,256,0.009376000199053023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,3584,0.012792000340090858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,128,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,3072,0.33360976643032497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,3072,0.008608889248636033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,65536,0.06964977582295735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,12288,0.0606844425201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,65536,0.06577599710888334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,512,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,3584,0.01995733380317688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,768,0.03052977720896403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,6144,0.0458586679564582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,128,0.013068444199032254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,5120,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,4096,0.026370666093296472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,4096,0.023188443647490606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,4096,0.012286221815480126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,8192,0.029787556992636785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,7168,0.4930933316548665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,16384,0.11081600189208984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,5120,0.027072888281610277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,8192,0.6348400115966797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,16384,0.13313599427541098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,2048,0.03808355662557814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,4096,0.12221866183810765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,16384,0.024373332659403484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,2048,0.04495199852519565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,3072,0.9268888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,3584,0.011025778121418424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,512,0.08224533663855658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,51200,10.872250027126737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,10240,0.2618853251139323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,7168,0.031153778235117596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,4096,0.01054222219520145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,128,0.02459200057718489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,6144,0.015847999188635085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,128,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,512,0.024062222904629175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,768,0.028799109988742407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,512,0.029870222012201946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,64,0.00816888858874639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,8192,0.07407555315229628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,256,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,4096,0.07334133651521471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,16384,0.03756533397568597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,10240,0.4666302469041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,3584,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,6144,0.03288622366057502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,1536,0.019514666663275827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,128,0.015293333265516492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,65536,0.05855911307864719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,4096,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,3072,0.07505333423614502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,1536,0.004559999952713649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,2560,0.058224889967176646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,1024,0.012773333324326409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,512,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,51200,0.4134719901614719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,12288,0.004953777624501122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,2048,0.04000799854596456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,3072,0.25226132074991864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,5120,0.0188737776544359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,512,0.003212444484233856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,128,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,5120,0.0184871107339859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,3584,0.03346844514211019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,64,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,256,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,256,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,3584,0.12844178411695692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,3072,0.42009690072801376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,2048,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,8192,0.02542488939232296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,64,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,64,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,51200,0.054195556375715465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,8192,0.015089778436554803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,8192,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,128,0.49563201268513996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,1024,0.011707555916574268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,6144,0.013750221994188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,2560,0.012461333639091916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,8192,0.014560000763999091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,128,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,1536,0.012057777908113269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,2048,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,3584,0.007228444847795699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,2048,0.010286221901575724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,10240,0.016188444362746347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,10240,0.012146666646003723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,256,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,7168,0.021368889345063105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,6144,0.05553866757286919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,3072,0.2985404332478841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,65536,0.28260623084174263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,10240,0.025747555825445387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,16384,0.2901617685953776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,128,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,16384,0.17677066061231825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,64,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,128,0.030079109801186457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,768,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,4096,0.01902844508488973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,4096,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,16384,0.04311644368701511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,256,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,512,0.013742222554153867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,10240,0.03832888934347365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,2560,0.019197333190176222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,1024,0.0034862222770849862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,128,0.011704888608720569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,3072,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,4096,0.9964532852172852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,8192,0.2286897765265571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,2048,0.01054933336046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,51200,0.0803440014521281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,3584,0.30731201171875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,65536,0.06237510840098063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,2560,0.06347378094991048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,3072,0.00739911115831799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,8192,0.057532442940606006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,10240,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,256,0.013847110999955071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,512,0.042021334171295166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,1024,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,51200,0.052935110198126904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,1024,0.1050106684366862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,1536,0.02309866746266683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,256,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,32,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,3072,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,10240,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,51200,0.9301172892252604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,128,0.01201244443655014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,256,0.0064399999876817065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,2048,0.018949333164427016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,16384,0.23678665690951875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,6144,0.10845510827170478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,8192,0.012058667010731168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,2048,0.02826933397187127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,512,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,1536,0.031806220610936485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,1024,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,512,0.005423999908897612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,768,0.0074684446056683855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,512,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,4096,0.008265777594513362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,10240,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,6144,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,8192,0.10956444342931111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,3584,0.02362222141689724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,12288,9.847533331976996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,16384,0.024321777953041926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,5120,0.0566213329633077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,10240,0.04059200154410468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,512,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,65536,0.8193973435295953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,256,0.019175110591782462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,7168,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,64,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,6144,0.03228977653715346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,3584,0.014682667122946845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,2048,0.013722666435771518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,51200,0.05860977702670627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,12288,0.011689778003427716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,65536,0.12746932771470812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,7168,0.018005333013004728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,5120,0.010066666536860997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,256,0.010403555300500657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,8192,0.039270222187042236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,6144,0.024319110645188227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,256,0.3162142170800103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,7168,0.13876267274220785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,6144,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,32,0.006440889090299606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,51200,0.063009778658549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,5120,0.06422666708628337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,1024,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,2560,0.014070222775141398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,6144,0.08838755554623073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,1536,0.18056889375050864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,1024,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,128,0.006207110981146495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,2048,0.03353510962592231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,5120,0.11674843894110785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,1024,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,256,0.00535377777285046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,10240,0.7246941990322537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,5120,0.4215964476267497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,4096,0.024732444021436904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,7168,0.16062045097351074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,256,0.006490666833188798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,65536,0.05730311075846354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,256,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,5120,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,6144,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,64,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,2048,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,128,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,16384,0.0210524449745814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,2048,0.026752889156341553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,8192,0.025207999679777358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,2048,0.06179910898208618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,7168,0.01351199961370892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,512,0.01776266594727834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,65536,0.06878133614857991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,1024,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,51200,0.08120444748136732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,7168,0.013013333082199097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,512,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,3584,0.028427554501427546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,3072,0.025946666797002155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,2560,0.03458311160405477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,16384,0.12206844488779704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,1024,0.035644445154401995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,10240,0.14795377519395617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,7168,0.010533333652549319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,8192,0.08906222052044338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,2560,0.012471111284361945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,4096,0.026862222287389968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,8192,0.03630400035116408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,512,0.06476977798673841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,65536,0.9171733856201172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,3072,0.04307199849022759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,12288,0.01682400041156345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,128,0.055903110239240855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,51200,0.04717155628734165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,10240,0.023035556077957153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,8192,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,65536,0.06272266970740424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,12288,0.04299911194377475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,2560,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,4096,0.20304532845815024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,5120,0.021631111701329548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,1536,0.00870577742656072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,5120,0.22806223233540854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,51200,0.6268311076694065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,3072,0.05506666501363119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,5120,0.04914044340451559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,1536,0.015513777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,5120,0.011483555866612328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,512,0.020169778002632987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,256,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,2560,0.051932444175084434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,2048,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,512,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,3584,0.026053332620196875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,128,0.014893333117167154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,256,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,3584,0.007793777518802219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,7168,0.022071111533376906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,1024,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,256,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,128,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,8192,0.024991111622916326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,2048,0.012432888978057437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,2560,0.00489155575633049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,2560,0.020950223008791607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,6144,0.02426311042573717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,51200,0.05616622500949436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,5120,0.15517332818773058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,3584,0.004950222041871813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,16384,0.03288711110750834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,7168,0.06521955463621351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,512,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,51200,0.06743911239835951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,7168,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,7168,0.02010933392577701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,32,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,10240,0.009729777773221334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,2048,0.010524444282054901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,768,0.021623111433453027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,10240,0.026686223016844854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,10240,0.011170667078759937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,2048,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,10240,0.030566222137875024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,7168,0.022711111439598933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,3584,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,256,0.005462222215202119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,1024,0.01681777834892273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,512,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,768,0.005635555419656966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,1536,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,4096,0.24250488811069063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,3072,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,65536,0.06298844681845771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,16384,0.0225600004196167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,16384,0.3704977830251058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,6144,0.058745774957868785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,3584,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,6144,0.06129688686794705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,10240,0.023432888918452795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,8192,0.015678221980730694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,512,0.018431999617152743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,65536,0.19807910919189453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,12288,0.03643111056751675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,2560,0.036637332704332136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,128,0.006517333288987477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,256,0.03011200163099501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,65536,0.024414221445719402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,4096,0.035727110173967146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,8192,0.08301955461502075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,2560,0.027653333213594224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,10240,0.33895111083984375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,32,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,12288,0.37928088506062824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,10240,0.013904000322024027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,65536,0.19399555524190268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,1536,0.006514666808976068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,256,0.015916445189052157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,256,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,512,0.017027555240525138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,16384,1.0966053009033203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,10240,0.03077600068516201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,7168,0.020351111888885498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,8192,0.1493697828716702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,3072,0.05366311139530606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,3072,0.10726933346854316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,512,0.002332444406217999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,768,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,51200,14.257037692599825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,3072,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,51200,0.4078017870585124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,16384,0.48285600874159074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,12288,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,512,0.014231999715169271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,768,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,768,0.028232001596026953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,16384,0.29286932945251465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,8192,0.036333332459131874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,5120,0.18917777803209093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,7168,0.21274399757385254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,8192,0.03216800093650818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,2048,0.012376888758606382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,5120,0.01759466694460975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,3584,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,7168,0.013227555486891003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,256,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,768,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,512,0.023634667197863262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,16384,0.05495377712779575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,3072,0.01607022186120351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,4096,0.014877332581414117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,32,0.0018097776919603348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,128,0.0025555555605226094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,2560,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,8192,0.018106665876176622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,4096,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,64,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,7168,0.06176533301671346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,8192,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,256,0.0032151111712058387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,256,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,10240,0.08351022005081177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,10240,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,1024,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,1024,0.023153778579499986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,4096,0.0403164459599389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,3072,0.012822222378518848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,128,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,6144,0.04099199838108487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,6144,0.036972443262736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,10240,0.026761778526835974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,12288,0.030823998981051978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,3072,0.037754668129815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,512,0.03477955526775784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,16384,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,1024,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,2560,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,6144,0.02058577703105079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,12288,0.026376888155937195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,7168,0.025395555628670588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,12288,0.018826666805479262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,2560,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,8192,0.023636443747414485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,512,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,2048,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,1024,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,16384,0.05817066960864597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,1024,0.017080888152122498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,3584,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,3584,0.04597599969969856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,4096,0.019409777389632333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,1536,0.008430221842394935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,16384,0.030076444149017334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,2560,0.010327999790509542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,8192,0.07804799742168851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,6144,0.015459555718633862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,1024,0.24975821706983778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,32,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,10240,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,256,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,4096,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,2560,0.05043288734224108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,32,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,2048,0.026259554757012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,10240,0.027626666757795546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,256,0.012474666866991254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,3584,0.0105608892109659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,1536,0.007732444339328342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,65536,0.22425956196255156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,1536,0.024523556232452393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,7168,0.011938666303952536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,51200,0.0982088910208808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,12288,0.05996888875961304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,512,0.01893422173129188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,128,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,8192,0.03049244483311971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,51200,0.04718577861785889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,128,0.01612622208065457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,2048,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,16384,0.07508533530765109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,4096,0.09523289071189033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,3584,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,2560,0.04789688852098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,2048,0.051197333468331226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,8192,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,6144,0.03724177678426107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,2048,0.07641066445244683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,65536,0.06385066774156359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,10240,0.11746488677130805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,8192,0.9096435970730252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,7168,0.05008000135421753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,8192,0.02796622117360433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,3584,0.03938488827811347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,16384,2.3695074717203775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,3072,0.06400799751281738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,256,0.00998933364947637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,2560,0.06888977686564128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,128,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,10240,0.30314577950371635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,3072,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,51200,0.020971554848882887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,65536,0.32919377750820583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,8192,0.03836799992455377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,51200,0.26380978690253365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,2048,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,128,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,768,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,64,0.005863111052248213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,64,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,512,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,64,0.005535111245181825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,2048,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,51200,0.11523289150661892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,7168,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,7168,0.08595289124382867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,51200,0.05387466814782885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,65536,0.3254622353447808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,3584,0.008943110704421997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,32,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,51200,0.5467333263821071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,128,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,3584,0.00489777781897121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,128,0.006157333238257303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,6144,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,256,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,51200,0.07355111175113253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,7168,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,4096,0.041783110962973706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,256,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,2048,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,65536,0.0666159987449646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,2560,0.1317760017183092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,3584,0.016623111234770883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,8192,0.07974400122960408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,5120,0.08933955430984497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,5120,0.04634222057130602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,8192,0.026135111848513286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,10240,0.09362133344014485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,1536,0.14202400048573813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,4096,0.010207999911573198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,4096,0.052579555246565074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,256,0.02162933349609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,16384,0.03228177626927694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,65536,0.06293600135379367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,3072,0.007800000409285228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,128,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,7168,0.014225777652528552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,7168,0.012851555314328937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,2048,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,2048,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,7168,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,10240,0.021947556071811255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,2048,0.012344000240166983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,51200,0.5883982446458604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,6144,0.012233778006500669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,256,0.0025475554996066624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,768,0.016532444291644625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,3072,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,16384,0.057016889254252114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,4096,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,3072,0.005530666559934616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,5120,0.0543484456009335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,2048,0.006559999866618051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,3072,0.02196888956758711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,3072,0.01867644488811493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,768,0.02106399999724494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,512,0.05715377463234795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,2048,0.028008000718222723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,1536,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,12288,0.029743999242782593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,3072,0.03325600094265408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,5120,0.007864889171388414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,128,0.010256888965765635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,3072,0.011094222466150919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,6144,0.009896000226338705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,7168,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,1536,0.02103022237618764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,10240,0.12259911166297065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,768,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,1536,0.011446221835083432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,256,0.08640088637669881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,12288,0.03591022226545546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,2048,0.0053582220441765254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,3584,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,5120,0.027990221977233887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,1024,0.019638222124841478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,16384,0.06674666537178887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,3584,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,512,0.013381332986884646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,5120,0.10634755425983006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,10240,0.0468773345152537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,4096,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,3072,0.023183999790085688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,5120,0.024889777104059856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,1536,0.04429777794414096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,1024,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,3072,0.004961777892377642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,128,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,3072,0.00508266646001074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,2048,0.006414222220579783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,10240,0.012239110966523489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,5120,0.03757777810096741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,4096,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,128,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,1024,0.006182222316662471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,3072,0.0358168880144755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,128,0.046684443950653076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,768,0.02472177810139126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,65536,0.045716444651285805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,64,0.0018542222678661346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,10240,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,2560,0.021336888273557026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,768,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,51200,0.604240894317627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,128,0.005738666488064661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,2560,0.04994933472739326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,12288,0.023669333921538457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,10240,0.041278223196665444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,51200,0.2685884369744195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,2048,0.04554666744338142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,8192,0.0444088876247406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,3072,0.052730666266547314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,3072,0.02808088726467556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,10240,0.011923555698659686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,6144,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,32,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,6144,0.11393688784705268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,7168,0.12161333031124538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,2048,0.018345778187115986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,2560,0.016940444707870483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,2048,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,4096,0.007873777714040544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,1536,0.010994666980372535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,12288,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,1024,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,4096,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,4096,0.08859822485182021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,10240,0.0376800000667572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,64,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,512,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,16384,0.13078578313191733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,4096,0.07395022445254855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,1536,0.008877333667543199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,64,0.013087111214796701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,7168,0.12213777171240912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,1536,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,2560,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,3072,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,3072,0.021822222405009802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,51200,0.05006044440799289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,256,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,8192,0.15390400091807047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,1536,0.008757332960764566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,256,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,1024,0.019673777951134574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,12288,0.2772168848249647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,512,0.08906666437784831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,2048,0.013183111117945777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,6144,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,5120,0.029095109966066148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,6144,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,768,0.002663111107216941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,3072,0.012243555651770698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,12288,0.029903110530641343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,768,0.00960000024901496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,512,0.007705777883529663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,2560,0.045844445625940956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,1536,0.012132444315486483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,16384,0.052278220653533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,5120,0.05937066343095568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,3584,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,128,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,65536,0.19737777445051405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,2560,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,65536,0.06451911396450467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,1024,0.03197866678237915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,6144,0.02251022226280636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,6144,0.023016888234350417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,3072,0.03600000010596381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,2560,0.01993688941001892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,10240,0.022892443670166865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,65536,0.6355315314398872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,3072,0.017557332913080852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,3072,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,6144,0.2570719983842638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,256,0.019896888070636325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,4096,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,1536,0.08570310804578994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,7168,0.028140445550282795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,1024,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,51200,0.19308800167507598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,512,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,8192,0.03110755483309428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,64,0.0022542222092549005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,16384,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,2560,0.02029600077205234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,3072,0.025516443782382544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,2560,0.030807998445298936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,16384,1.7131715350680883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,7168,0.15886310736338297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,4096,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,10240,0.04076177875200907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,5120,0.05982933441797892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,2048,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,128,0.011147555377748279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,1024,0.015857777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,768,0.07279022534688313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,3072,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,16384,0.030805332793129817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,51200,0.10382755597432454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,3584,0.07641510830985175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,768,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,256,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,1536,0.0069013333155049225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,2048,0.019906666543748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,3072,0.03425688876046075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,1536,0.013061333033773633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,6144,0.011792889071835412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,128,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,6144,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,5120,0.03648266527387831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,5120,0.017269333203633625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,3072,0.0203804439968533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,4096,0.023737778266270954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,6144,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,64,0.34935821427239316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,1024,0.020842666427294414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,10240,0.012741333080662621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,4096,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,128,0.014914666612943014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,1024,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,2048,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,6144,0.09950577550464207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,128,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,51200,0.043066667185889355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,5120,0.00814577771557702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,65536,0.06547644403245714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,1024,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,8192,0.14315644900004068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,4096,0.3503893216451009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,6144,0.018935999936527677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,12288,0.0418888893392351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,8192,0.05557688739564684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,32,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,16384,0.057392888598971896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,4096,0.008055999875068665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,32,0.02383733292420705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,4096,0.12828089131249323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,3584,0.05453244513935513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,7168,0.027074666486846075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,128,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,5120,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,51200,0.706025759379069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,64,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,768,0.021662222014533147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,768,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,4096,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,768,0.01825777855184343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,2560,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,2560,0.057914667659335665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,6144,0.05572977993223402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,12288,0.021126222279336717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,64,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,5120,0.030172444052166406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,51200,0.11506577332814534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,3584,0.02053777707947625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,32,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,128,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,3584,0.049426668220096164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,6144,0.038374221987194486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,512,0.010838222172525195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,128,0.0022888889329300984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,64,0.005708444449636672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,6144,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,6144,0.10637688636779785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,8192,0.4264657762315538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,65536,2.8062720828586154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,3072,0.031888888941870794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,65536,0.09380977683597141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,256,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,4096,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,512,0.006077333456940121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,7168,0.036454220612843834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,3072,0.02199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,3584,0.017165333032608032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,16384,0.06568177541097005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,768,0.051847110191980995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,32,0.0018951110541820526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,8192,0.06979288657506307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,2560,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,65536,0.14879288938310412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,768,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,1536,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,5120,0.10054044591055976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,16384,0.13521867328219944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,8192,0.01898755629857381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,32,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,12288,0.5508862071567112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,65536,0.5613164371914333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,1024,0.018591110905011494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,16384,0.33765334553188747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,768,0.05307022068235609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,2560,0.07852266894446479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,3072,0.013575110998418597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,1024,0.042287998729281955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,32,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,1024,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,2048,0.006871999965773688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,1024,0.018516444497638278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,256,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,1536,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,5120,0.07494222455554538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,4096,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,7168,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,16384,0.09953066375520493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,12288,0.4001439942253961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,3072,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,10240,0.01257955531279246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,768,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,1536,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,4096,0.918476422627767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,51200,0.19471555285983613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,51200,0.14064621925354004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,768,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,5120,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,2048,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,8192,1.459232012430827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,128,0.13247466087341309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,7168,0.015467555986510383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,4096,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,512,0.056177775065104164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,12288,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,8192,0.03701511025428772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,51200,0.09505777888827854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,3072,0.253224876191881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,8192,0.2303395536210802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,65536,0.12167555756039089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,3072,0.17140177885691324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,2048,0.023387556274731953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,3072,0.06049955553478665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,3072,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,12288,0.021936888496081035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,65536,0.07486133442984687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,5120,0.012840888566441007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,10240,0.2034693294101291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,3072,0.08164088593588935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,12288,0.0628568861219618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,10240,0.015637333194414776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,2048,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,7168,0.03151377704408433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,6144,0.02639555599954393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,12288,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,5120,0.280786673227946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,16384,0.01831644442346361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,12288,0.00886044485701455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,5120,0.019707555572191875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,64,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,256,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,12288,0.1051822238498264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,1024,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,6144,0.3075982199774848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,3072,0.013384888569513956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,1024,0.07051022185219659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,8192,0.03428355521625943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,7168,0.01941333380010393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,1536,0.011647111011876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,51200,0.023177777727444965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,2560,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,2048,0.020847110284699336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,256,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,51200,2.393506579928928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,256,0.02640177806218465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,5120,0.004555555681387584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,12288,0.10086311234368218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,512,0.05112266540527344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,16384,0.04955910974078708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,256,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,7168,0.03383644421895345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,7168,0.01663466625743442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,4096,0.03974311219321357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,64,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,256,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,4096,0.03674488928582933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,6144,0.01739911072784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,16384,0.06428088744481404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,3584,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,3072,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,1536,0.024922667278183833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,8192,0.04301777813169691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,51200,0.3192124366760254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,7168,0.2453662289513482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,6144,0.010456889039940305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,6144,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,10240,0.012373333175977072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,16384,0.042657779322730176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,2560,0.022663111488024395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,32,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,512,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,256,0.02880355715751648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,3072,0.018996444013383653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,65536,2.33931032816569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,256,0.025607999828126695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,3584,0.28672801123725045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,7168,0.013256889250543384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,5120,0.15080000294579401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,512,0.01776444415251414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,10240,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,2560,0.028752889898088243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,128,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,16384,0.025366221865018208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,2560,0.023481777972645227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,16384,0.13349066840277776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,8192,0.05968799855973986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,256,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,1536,0.029993779129452173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,16384,0.2770995563930935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,3072,0.014859555496109856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,4096,0.10344355636172825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,16384,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,2560,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,256,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,768,0.06716533501942952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,7168,0.011182222101423474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,7168,0.045765333705478244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,512,0.012834666503800286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,6144,0.026280888252788123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,51200,0.5481831232706705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,6144,0.06423911121156481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,2048,0.025753777888086107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,65536,0.057761775122748486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,128,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,512,0.0031946665710873077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,512,0.009917333722114563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,2560,0.1262257761425442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,5120,0.07422755824195014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,7168,0.06331466966205172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,64,0.0018782222436534034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,4096,0.011167999770906238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,768,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,1024,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,1536,0.0047146665553251905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,5120,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,3584,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,6144,0.019116444720162284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,2048,0.03306044472588433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,1024,0.02255288925435808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,256,0.02257866660753886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,65536,0.01688888834582435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,3584,0.026777777406904433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,8192,0.10956799983978271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,2560,0.024193776978386775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,2048,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,6144,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,16384,0.02011999984582265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,512,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,65536,0.7673048973083496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,768,0.008485332959228093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,10240,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,1024,0.00720355576939053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,16384,0.03187288840611776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,7168,0.011063999599880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,16384,0.0229768885506524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,8192,0.08074755801094903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,8192,0.01591377788119846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,64,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,65536,2.3348070780436196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,8192,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,10240,0.012478222449620565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,6144,0.6475306616889106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,5120,0.01711466742886437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,3584,0.04653866754637825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,2048,0.034055110481050276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,128,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,6144,0.01884888940387302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,1536,0.025065778030289546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,16384,0.1665653387705485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,512,0.01870844430393643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,3584,0.055514666769239635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,3584,0.06565866867701213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,8192,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,128,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,3584,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,16384,0.2159271107779609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,512,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,2048,0.014464888307783337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,256,0.006370666540331311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,12288,0.3050684399074978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,4096,0.21338489320543078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,3072,0.076928887102339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,2560,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,512,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,3584,0.02658044464058346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,128,0.0020373333245515823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,128,0.08869510889053345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,51200,0.04683644572893778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,64,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,12288,0.017761778500345018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,7168,0.027913779020309448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,51200,0.10271733336978489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,4096,0.011200000014570026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,3072,0.00888177752494812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,51200,0.11253066857655843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,12288,0.07835377587212457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,4096,0.010442666709423065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,16384,0.041859553919898145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,7168,0.043262223402659096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,10240,0.015188443991872998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,7168,0.029119110769695703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,3584,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,51200,1.2952488793267143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,6144,0.01695999999841054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,8192,0.049955556790033974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,7168,0.07480266359117296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,3072,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,5120,0.05746399694018894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,7168,0.04162755608558655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,6144,0.17436533504062227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,8192,0.020995555652512446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,1024,0.02435644467671712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,8192,0.015384000208642749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,7168,0.03392533461252848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,768,0.014747555057207743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,64,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,16384,0.0630177789264255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,7168,0.06473422050476074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,4096,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,1536,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,512,0.003936888857020272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,7168,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,5120,0.016210666961140104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,4096,0.023079110516442194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,5120,0.022566222482257422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,32,0.002238222294383579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,16384,0.020908445119857788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,1536,0.07377333111233182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,128,0.011411555939250521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,65536,0.03927733169661628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,12288,0.005336888962321811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,128,0.005180444568395615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,5120,0.016813332835833233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,3584,0.013384888569513956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,10240,0.01520088811715444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,7168,0.13037689526875815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,10240,0.0422435568438636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,64,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,128,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,7168,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,7168,1.790427525838216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,1536,0.008747555315494537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,1536,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,256,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,128,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,1536,0.019377777973810833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,512,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,5120,0.1640133327907986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,512,1.8294444613986547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,512,0.019983111156357657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,51200,0.5225111113654243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,128,0.015014222926563688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,2048,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,12288,0.02831377916865879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,3072,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,12288,0.013264888690577613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,16384,0.022113778524928625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,51200,0.07929511202706231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,3584,0.0732311142815484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,6144,0.031496001614464655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,64,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,16384,0.9618061913384331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,16384,0.03805244300100539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,51200,0.33006400532192653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,6144,0.04109155469470554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,2560,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,6144,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,5120,0.021846221552954778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,768,0.037321776151657104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,3072,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,16384,0.03470755616823832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,256,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,768,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,1536,0.0353724459807078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,7168,0.04158488909403483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,4096,0.026565333207448322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,4096,0.012355555262830524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,1024,0.02682311170630985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,1536,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,16384,0.10387821992238362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,32,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,51200,0.11600177817874485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,2560,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,8192,0.03052355514632331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,2048,0.07524977789984809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,4096,0.020258666740523446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,16384,0.08050844404432508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,1024,0.01372977760103014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,512,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,5120,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,3072,0.012716444830099741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,512,1.3899102740817602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,7168,0.011367999845080905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,768,0.016688000824716356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,3584,0.013776000175211163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,3584,0.009280000295903947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,2048,0.12936622566647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,8192,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,4096,0.03810933232307434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,1024,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,128,0.02324177821477254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,512,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,8192,1.4030115339491103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,128,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,12288,1.2729288736979167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,256,0.02219911085234748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,51200,0.2667795552147759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,51200,0.10482577482859294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,8192,0.16695733865102133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,3072,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,65536,0.4313608805338542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,64,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,16384,0.04422933194372389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,51200,0.0579048858748542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,12288,0.023359111613697473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,128,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,768,0.006337777607970768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,6144,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,6144,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,51200,0.18589866161346436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,10240,0.05193155672815111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,8192,0.13966133859422472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,2560,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,128,0.010408889088365767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,256,0.015588444140222339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,10240,0.030480888154771592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,10240,0.03674577673276266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,16384,0.016345777445369296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,256,0.09845244222217137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,6144,0.04397777716318766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,128,0.021545777718226116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,256,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,16384,0.015823110938072205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,768,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,3584,0.10641066895590888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,51200,0.12514755460951063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,65536,0.06426399946212769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,7168,0.10608800252278645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,2560,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,512,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,1536,0.029461334149042766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,3584,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,4096,0.02277066641383701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,64,0.005432888865470886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,3072,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,32,0.021204445097181533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,32,0.0020026666008763844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,3072,0.1874693367216322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,128,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,8192,0.19920978281233045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,32,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,1536,0.3155084451039632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,16384,0.018553778529167175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,6144,0.12457156181335449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,1536,0.024079110887315538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,12288,0.2311671045091417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,51200,42.618384467230904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,6144,0.015576889117558798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,3584,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,65536,0.5632328987121582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,1536,0.03532622257868449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,8192,1.4877440134684246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,10240,0.2262506749894884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,65536,0.6309048864576552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,16384,0.0222053329149882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,16384,0.023623999622133043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,512,0.008951110972298516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,6144,0.010070222119490305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,3072,0.008270222279760573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,512,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,3584,0.038648890124426946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,5120,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,1024,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,8192,0.02202933364444309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,16384,0.07197688685523139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,16384,0.021307556165589228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,4096,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,4096,0.009529777699046666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,3072,0.2869902186923557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,16384,0.05123466584417555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,7168,0.024995555480321247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,3072,0.023752000596788194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,8192,0.13958933618333605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,768,0.03864266806178623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,3072,0.027512888113657635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,256,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,6144,0.03534311056137085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,1024,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,5120,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,65536,0.1694577799903022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,64,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,2560,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,12288,0.020611556039916146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,1536,0.0346995559003618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,768,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,768,0.02539199921819899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,10240,0.0950124462445577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,1536,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,5120,0.4423271020253499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,8192,0.021528000632921856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,1536,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,3584,0.03860622313287523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,7168,0.02160888910293579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,12288,0.040056887600157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,16384,0.015784889459609985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,512,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,768,0.019340443942281935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,3072,0.07193777958552043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,128,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,3072,0.007015110717879401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,16384,0.13720266024271646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,768,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,3584,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,16384,4.3265503777398004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,64,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,7168,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,65536,0.02068622244728936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,5120,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,768,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,5120,0.4196586608886719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,768,0.0034835553831524322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,3072,0.21468888388739693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,512,0.013603555659453074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,32,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,2560,0.009493333597977957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,10240,0.004601777841647466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,7168,0.06682222419314914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,51200,0.2988453441196018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,51200,7.939679463704427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,2048,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,8192,0.3638604482014974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,10240,0.025047111842367385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,51200,0.5638471179538304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,768,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,10240,0.02141511109140184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,3072,0.46852535671657985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,128,0.02264533274703556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,2560,0.01834755473666721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,10240,0.03955111238691542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,12288,0.0599493318133884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,65536,0.030673778719372217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,10240,0.049555556641684644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,2048,0.006456888798210356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,12288,0.07599555783801608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,12288,0.16370044814215765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,12288,0.04054044352637397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,3584,0.020928000410397846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,128,0.0972248911857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,768,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,2048,0.005335999859703912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,51200,2.1718578338623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,3072,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,3584,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,128,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,1536,0.006414222220579783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,256,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,4096,0.0322053333123525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,2560,0.01055111073785358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,2560,0.01331822243001726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,64,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,10240,0.2554115454355876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,768,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,1536,0.027495111028353374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,3584,0.01888088881969452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,1536,0.0543502238061693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,32,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,3072,0.0661804477373759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,7168,1.7155200110541449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,256,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,768,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,4096,0.018076444665590923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,10240,0.0749066670735677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,64,0.0018817778262827133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,16384,0.17985778384738496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,16384,0.03018666638268365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,768,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,256,0.03540888759824965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,5120,0.021027555068333943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,2560,0.02234488891230689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,4096,0.02546399997340308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,768,0.016795555750528973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,32,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,1024,0.016162667009565566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,65536,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,65536,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,1536,0.28149154451158315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,32,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,6144,0.040207998620139226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,51200,2.4053031073676214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,65536,0.14733866850535074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,12288,0.14713422457377115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,6144,0.14471466011471218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,3584,0.08819822470347087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,16384,0.009913778139485253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,128,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,10240,0.04378044605255127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,65536,5.047852410210504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,10240,1.788321812947591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,256,0.021143999364640977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,64,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,7168,0.08556889163123237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,10240,0.025384000606007043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,128,0.01074577785200543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,16384,0.03326311045222812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,10240,0.11272444989946152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,5120,0.03849333193567064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,2048,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,12288,0.014367999302016364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,5120,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,512,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,12288,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,3072,0.028466665082507666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,768,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,10240,0.02215911116864946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,768,0.017889777819315594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,4096,0.006992888947327931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,4096,0.016650666793187458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,51200,0.026574222577942744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,3072,0.03089511063363817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,6144,0.029896888468000624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,51200,0.04952711198065016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,51200,0.0630577802658081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,128,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,1536,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,1024,0.02293955617480808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,3072,0.012155555188655853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,12288,0.023669333921538457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,128,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,7168,0.10539911190668742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,16384,0.019270223047998216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,512,0.004029333177540037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,7168,0.057799114121331104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,128,0.07315911187065972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,8192,0.007836444510353936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,128,0.012456888953844706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,12288,0.06517422199249268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,128,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,768,0.22456267144944933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,12288,0.05327466792530484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,1536,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,3072,0.03143289022975498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,7168,0.011379555695586734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,7168,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,10240,0.024384000235133704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,3584,0.19769867261250815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,7168,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,5120,0.014198222094111972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,2560,0.2920142279730903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,768,0.018246221873495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,3584,0.08921155664655898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,10240,0.05439111259248522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,512,0.0738568902015686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,51200,0.10530933406617905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,2048,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,3584,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,7168,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,16384,0.051593777206208974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,128,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,16384,0.016428444120619033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,12288,11.935189988878038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,4096,0.039797332551744245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,65536,0.06001600291993883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,6144,0.02149066660139296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,512,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,2048,0.06552710798051622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,3584,0.02422933280467987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,7168,0.39059999254014754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,64,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,128,0.020431111256281536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,4096,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,768,0.007912000020345053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,51200,0.49948713514539933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,1536,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,5120,0.0913155542479621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,2560,0.060284442371792264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,128,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,5120,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,5120,0.01958044370015462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,768,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,2048,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,1024,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,3584,0.011202666494581433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,10240,0.023045332895384893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,12288,0.024148444334665935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,4096,0.09216177463531494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,6144,0.1397262281841702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,1536,0.03158222304450141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,10240,0.015491555134455362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,3584,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,2048,0.018194667167133756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,4096,0.02064622276359134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,51200,0.564284430609809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,2560,0.07404444615046184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,2048,0.04640266630384657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,4096,0.057216889328426786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,8192,0.03261244297027588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,32,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,2560,0.004943111290534337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,7168,0.03684444559945001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,128,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,256,0.006081777728266186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,768,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,51200,0.04920533299446106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,7168,0.03207111027505662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,2048,0.08126133018069796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,4096,0.062014222145080566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,3072,0.014718222949239941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,128,0.03273511264059279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,8192,0.2001724508073595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,6144,0.01197866681549284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,1024,0.016545777519543965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,3072,0.022780444886949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,768,0.0026693333768182327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,768,0.0027840000887711844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,7168,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,10240,0.06051377455393473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,10240,1.9444107479519313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,6144,0.1577546728981866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,4096,0.009468444519572789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,51200,4.622698465983073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,5120,0.5313439899020725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,2560,0.01202044470442666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,8192,0.5333306524488661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,5120,0.14289600319332546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,3072,0.03541333476702372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,16384,0.04684888985421923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,4096,0.5929617881774902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,5120,0.04741333259476555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,1536,0.004504888835880491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,8192,1.0399502648247612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,6144,0.9179137547810873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,3072,0.010423999693658618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,51200,0.1382719940609402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,16384,0.03427733315361871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,2560,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,51200,0.6274497773912218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,3584,0.007438222567240397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,768,0.0034097778714365433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,512,0.056369774871402316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,12288,0.028840889533360798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,16384,0.03049244483311971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,8192,0.004548444516128964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,512,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,65536,0.3542524443732367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,12288,0.02978133492999607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,512,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,768,0.005737777799367905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,64,0.0760177771250407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,128,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,2560,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,3072,0.014628445108731588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,2560,0.5495688650343153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,65536,0.0773271123568217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,128,0.013804444836245643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,12288,0.03401333424780104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,512,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,7168,0.13983378145429823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,4096,0.03738488753636678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,4096,0.01147377739350001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,8192,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,51200,0.04890933301713732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,512,0.003063999944263034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,16384,2.9477386474609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,3584,0.07258844375610352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,6144,0.3931199974483914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,1024,0.024851555625597637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,7168,0.02068533334467146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,6144,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,3072,0.10928710963990952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,3072,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,1024,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,3072,0.6065262158711752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,256,0.02258666687541538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,128,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,768,0.01940444442960951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,6144,0.054063111543655396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,51200,0.04594044552909004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,1024,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,2560,0.004970666848950916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,12288,0.037610666619406805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,32,0.022356443934970435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,12288,0.01733688844574822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,7168,0.08415289057625665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,64,0.011129777464601727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,4096,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,16384,0.05843289030922783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,10240,0.03940088881386651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,8192,0.05798044469621447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,4096,0.019501333435376484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,3072,0.010890666809346942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,512,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,6144,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,2048,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,3584,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,512,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,1536,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,7168,0.027430221438407898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,512,0.007157333195209503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,256,0.020059555768966675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,4096,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,10240,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,16384,0.7530915472242566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,256,0.007384888827800751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,12288,2.0401492648654513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,32,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,7168,0.15402311748928493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,16384,0.038394666380352445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,128,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,7168,0.017253332667880587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,1536,0.008729777402347988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,65536,0.07240088780721028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,2560,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,256,0.009751111268997192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,10240,0.11939467324150933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,128,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,16384,0.04739200075467428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,8192,0.02363377809524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,65536,0.05632177988688151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,4096,0.026798221800062392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,6144,0.08491377698050605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,3072,0.015228443675571017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,51200,0.07701866494284736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,64,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,64,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,10240,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,2048,0.03287911083963182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,3584,0.022720888257026672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,7168,0.22690489557054308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,3072,0.008574222524960836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,7168,0.010430222584141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,51200,0.1409040027194553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,3584,0.013835555149449242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,8192,0.12142044968075222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,10240,0.020255110330051847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,12288,0.03901777664820353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,1024,0.046560890144772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,8192,0.015778667396969266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,512,0.014615999327765571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,8192,0.20428266790178087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,256,0.015435554915004306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,1536,0.008168000313970778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,65536,0.08738133642408584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,3072,0.09302844603856404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,128,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,6144,0.005668444352017508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,12288,0.035826666487587824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,51200,0.6842471228705512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,7168,0.12295288509792751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,768,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,768,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,7168,0.010057777994208867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,1536,0.022838221655951604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,2560,0.011805333197116852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,65536,2.0837156507703996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,2048,0.10804533296161228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,8192,0.07076622380150689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,2048,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,51200,0.28632622294955784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,51200,1.2581715053982205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,3072,0.06490400102403429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,65536,0.23119023111131456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,2560,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,1536,0.010943999720944298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,1536,0.00848355558183458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,65536,0.3869955539703369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,4096,0.038113776180479265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,768,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,32,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,3584,0.021710221966107685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,2560,0.005148444324731827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,2560,0.025803556044896443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,4096,0.028453333510292902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,1536,0.02623822291692098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,12288,0.018840889135996502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,65536,0.15872000323401556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,1024,0.015508444772826301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,16384,0.3264995680914985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,5120,0.014522666732470194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,3584,0.02843022346496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,2048,0.06290666924582587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,1024,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,6144,2.3983332316080728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,2560,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,768,0.03096888793839349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,3584,0.028365333875020344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,2048,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,1024,0.1354924440383911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,6144,0.03268266717592875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,256,0.03229422370592753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,5120,0.08854222297668457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,5120,0.01332266628742218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,2560,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,256,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,6144,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,8192,0.012968888713253869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,4096,0.12960710790422228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,512,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,65536,1.4863004684448242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,65536,0.1306613286336263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,7168,0.012852444416946836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,128,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,10240,0.030105776256985132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,1536,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,1024,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,10240,0.05616177452935112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,12288,0.050795555114746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,7168,0.026153778036435444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,3072,0.1290080017513699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,128,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,7168,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,51200,0.20110044214460585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,4096,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,12288,0.028869334194395278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,512,0.0026391111314296722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,256,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,5120,0.7103573481241862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,2048,0.01923644377125634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,3072,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,12288,0.031776888502968684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,6144,0.031154665682050917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,65536,3.08294317457411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,7168,0.034520887666278414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,7168,0.009524444739023844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,128,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,10240,0.15966667069329157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,128,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,65536,1.1892560323079426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,4096,0.023268444670571223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,1024,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,5120,0.03200088938077291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,768,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,3584,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,8192,0.03767733441458808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,5120,0.035064889325035944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,768,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,7168,0.010932444698280759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,6144,0.09127733442518447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,128,0.02276800076166789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,2048,0.020580444071027968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,6144,0.03643466697798835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,128,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,12288,0.03345600101682875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,256,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,51200,0.048960887723498873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,128,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,16384,0.01368711143732071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,16384,0.024967110819286768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,16384,0.2684595584869385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,64,0.011352000137170156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,256,0.012274666792816587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,10240,0.020785777105225455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,51200,0.6158071094089085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,16384,0.2135395606358846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,2560,0.03276977936426798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,2560,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,4096,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,6144,0.1275040043724908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,128,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,512,0.028591109646691218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,768,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,7168,0.04651733239491781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,64,0.02187555531660716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,3072,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,51200,0.21646666526794434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,6144,0.009915555516878763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,512,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,5120,0.018575110369258456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,32,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,128,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,51200,0.10728800296783447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,12288,0.04980355501174927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,512,0.002565333412753211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,64,0.0019146667586432563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,65536,0.3944186634487576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,1536,0.014865777558750577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,10240,0.01535999940501319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,3072,0.09360888931486343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,51200,0.4479546546936035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,32,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,5120,0.020768889122539096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,5120,0.033775111039479576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,128,0.16708089245690239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,51200,0.060290667745802135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,1024,0.012462221913867526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,5120,0.18246400356292725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,256,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,256,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,5120,0.027830223242441814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,7168,0.10303022464116414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,3072,0.09094755517111884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,10240,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,1024,0.012115555504957834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,3584,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,512,0.024067555864651997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,512,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,2048,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,51200,10.26816643608941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,5120,0.09347733524110581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,51200,0.49216265148586696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,2048,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,8192,0.045801778634389244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,256,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,12288,0.08321955468919542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,512,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,16384,0.055574225054846875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,5120,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,6144,0.015184000134468079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,65536,0.07413778040144178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,256,0.022825777530670166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,1024,0.007442666424645319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,2048,0.004604444321658877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,4096,0.024869332710901897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,1024,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,16384,0.019460444649060566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,51200,0.5005360179477268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,3072,0.18001333872477213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,1024,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,1024,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,512,0.017123555143674214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,1024,0.016030222177505493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,768,0.01777066621515486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,2560,0.012831110921170978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,12288,0.01646311084429423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,2560,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,768,0.013225778109497495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,12288,0.8974533081054688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,65536,0.22755911615159777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,10240,0.019131556153297424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,7168,0.14016000429789224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,10240,0.27725421057807076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,2048,0.06930666499667697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,3584,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,4096,0.057257778114742704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,256,0.0034373334298531213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,3072,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,7168,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,256,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,16384,0.04336533281538221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,12288,0.1555831167432997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,2048,0.007689778175618913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,65536,0.0743866695298089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,4096,0.023015111684799194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,3584,0.0292595558696323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,128,0.056311998102400035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,8192,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,65536,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,6144,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,1024,0.00886044485701455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,768,0.022916444473796423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,768,0.011953777737087674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,65536,0.06638311015235053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,768,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,3584,0.008583111067612966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,3584,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,6144,0.014009777042600842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,768,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,128,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,256,0.010710222025712332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,5120,1.2227520412868924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,2560,0.029145777225494385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,5120,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,64,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,7168,0.09061333205964829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,16384,0.19531377156575522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,512,0.039143999417622886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,256,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,3584,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,5120,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,2560,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,4096,0.019152889649073284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,6144,0.019546666079097323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,2560,0.08649066421720718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,12288,0.10844533310996161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,1536,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,7168,0.0314355558819241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,6144,0.12457422415415446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,7168,0.024417777856190998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,6144,0.2601199944814046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,1536,0.006350222147173352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,2560,0.006748444504208035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,51200,0.1970640023549398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,2048,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,128,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,512,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,65536,0.09654400083753799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,1024,0.06252800093756782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,16384,0.36318667729695636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,51200,2.684385723537869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,1024,0.04615733358595106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,3072,0.27617867787679035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,12288,0.02310488952530755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,128,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,7168,0.07679822047551473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,8192,0.028712888558705647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,1024,0.032991111278533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,2048,0.01717511150572035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,2560,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,6144,0.07409155368804932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,3072,0.012372444073359171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,256,0.007711110843552484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,51200,7.501426696777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,16384,0.1887742280960083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,4096,0.07151910993787977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,1024,0.005702222386995952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,2048,0.1432195504506429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,2048,0.012696888711717395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,768,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,2048,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,7168,0.011061333119869232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,3072,0.03187822302182516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,51200,0.061894224749671094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,12288,0.02162666618824005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,128,0.006533333410819371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,16384,0.46692175335354275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,10240,0.11485955450269912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,12288,0.0994488861825731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,2560,0.28071822060479057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,65536,0.6720328860812717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,256,0.010949333508809408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,51200,0.1759208838144938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,1536,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,128,0.021617778473430212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,16384,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,512,0.01637066735161675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,6144,0.01036444471942054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,65536,0.12361066871219212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,2560,0.0503422220547994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,4096,0.09832177559534709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,1536,0.06741244263119169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,51200,0.06637066602706909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,64,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,2560,0.033202668031056724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,65536,0.45729155010647243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,768,0.017723555366198223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,32,0.012098666694429187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,10240,0.05307288964589437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,6144,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,1024,0.047203554047478564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,12288,0.017052445146772597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,7168,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,128,0.014164444473054675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,8192,0.030638221237394545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,16384,15.667546590169271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,12288,0.10331200228797065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,16384,0.020607110526826646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,3584,0.04254577888382805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,32,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,10240,0.02253688871860504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,256,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,10240,0.11188799805111355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,2048,0.008296888735559251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,6144,0.13107200463612875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,2048,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,512,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,51200,0.3096240096622043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,1536,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,512,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,16384,0.05279111199908786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,3584,0.03200799889034695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,768,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,5120,0.027808000644048054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,5120,0.022640888889630634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,512,0.056200000974867076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,256,0.013721778160995908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,1536,0.018616888258192275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,256,0.008378666308191087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,32,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,65536,0.13135822614034018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,256,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,16384,0.06298400296105279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,12288,3.4961876339382596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,5120,0.03489244315359328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,128,0.0129457778400845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,2048,0.02472177810139126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,12288,0.44387290212843156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,12288,0.21519466241200766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,3584,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,4096,0.15460177262624106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,768,0.031108445591396753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,5120,0.027662222584088642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,256,0.012892444100644855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,10240,0.01685422162214915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,512,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,2048,0.023741333021057978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,1024,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,4096,0.036185777849621244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,2560,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,512,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,32,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,1024,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,256,0.03953333364592658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,12288,0.05080711179309421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,12288,0.038057777616712786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,3072,0.0492142207092709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,4096,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,256,0.005993777679072485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,12288,0.10739377472135757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,1536,3.798437330457899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,2048,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,256,0.046896000703175865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,3072,0.4539955457051595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,3072,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,1536,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,1024,0.02401955591307746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,128,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,1024,0.013812444276279874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,8192,0.0365866654449039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,1536,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,51200,0.04833955566088358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,1024,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,10240,0.6703528828091092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,10240,0.04407199886110094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,128,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,8192,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,2560,0.010292444792058732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,7168,0.027217777238951787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,768,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,6144,0.05310755636956957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,4096,0.14632444911532932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,7168,0.024714666936132643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,2048,0.2202648851606581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,3584,0.0848826633559333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,10240,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,256,0.07122311327192518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,3072,0.012669333153300814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,32,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,12288,0.03734222385618422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,7168,0.03840533230039809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,256,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,768,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,128,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,65536,0.019927110936906602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,32,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,512,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,128,0.04462933209207323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,65536,0.18349244859483507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,7168,0.07475199964311388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,512,0.0029804443733559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,10240,0.0567493306265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,7168,0.012147555748621622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,256,0.012239110966523489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,51200,0.06603378057479858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,1536,0.009409777820110321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,65536,0.06873777839872572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,16384,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,3072,0.02160533269246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,16384,0.0417848891682095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,1024,0.008482666479216682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,65536,0.3291520012749566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,512,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,7168,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,16384,0.06514310836791992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,1024,0.005707555760939916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,256,0.015279110934999255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,512,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,4096,0.046375112401114575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,5120,0.018290667070282828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,7168,0.04595111144913567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,256,0.03259199857711792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,3584,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,512,0.03504622313711379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,1536,0.029586666160159644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,3584,0.09805422359042698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,256,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,128,0.018751111295488145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,32,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,3584,0.022983110613293115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,1024,0.006405333264006509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,65536,2.0636418660481772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,128,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,5120,0.06945777601665921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,64,0.0694426695505778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,10240,0.020276443825827706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,8192,0.043812443812688194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,8192,0.024473778075642053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,16384,0.15559466679890951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,16384,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,6144,0.09415288766225179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,3072,0.014228444960382251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,7168,0.020279111133681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,64,0.04490400022930569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,16384,0.01906044450071123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,7168,0.06313066350089179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,1024,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,12288,0.019886222150590684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,6144,0.21757600042555067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,256,0.09175111187828912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,768,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,12288,0.02606222199069129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,7168,0.12655733691321477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,1024,0.023885332875781592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,512,0.010114666488435533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,7168,0.029166221618652344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,2048,0.014878221684032016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,768,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,65536,0.17661421828799775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,1536,0.030230220821168687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,2048,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,16384,0.04032266802257962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,768,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,7168,0.014728888869285583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,128,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,768,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,512,0.0028373334142896864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,3072,0.3983448876274957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,51200,0.10011111365424262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,2560,0.007501333124107785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,2560,0.024680889315075342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,3584,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,5120,0.015184889237085978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,16384,0.34639377064175075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,128,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,6144,0.012963555753231049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,16384,0.26655199792650014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,51200,0.07729511128531562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,64,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,512,0.00776622196038564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,4096,0.011045333411958484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,7168,0.013904889424641928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,5120,0.029517332712809246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,128,0.008195555872387355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,128,0.00647288892004225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,512,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,1536,0.01995733380317688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,32,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,6144,0.66357421875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,128,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,512,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,12288,0.15313243865966797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,256,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,2048,0.6417733298407661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,3072,0.03251466817326016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,6144,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,2048,0.0964319970872667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,2048,0.005535111245181825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,10240,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,10240,0.028487109475665625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,5120,0.15859199894799128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,128,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,4096,0.010448888772063784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,256,0.0053013331360287136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,1024,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,12288,0.05779822005165947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,2560,0.06124711036682129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,16384,0.026714666022194758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,1536,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,2048,0.06432800160513984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,6144,0.1960657835006714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,7168,0.06653066476186116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,64,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,1024,0.040447112586763166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,512,0.03178400132391188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,12288,0.02035466664367252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,1024,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,128,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,16384,0.019766221443812054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,128,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,16384,0.022320000661744013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,3584,0.045203556617101036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,3072,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,5120,0.047220445341534086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,3072,0.010423999693658618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,4096,0.03865244322352939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,5120,0.028130667077170476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,64,0.051262223058276706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,12288,0.6531270874871148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,3584,0.03949422306484646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,5120,0.06890044609705608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,8192,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,256,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,16384,0.6969528728061252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,16384,0.2914053334130181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,256,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,7168,0.005017777697907554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,768,0.009029332962301042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,12288,0.01314933349688848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,2048,0.012133333418104382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,256,0.012014221813943652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,64,0.25708267423841685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,4096,0.0318195554945204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,4096,0.024674667252434626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,3584,0.03839911023775736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,2048,0.027296889159414504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,8192,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,2560,0.05687822235955132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,51200,0.8157217767503527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,6144,0.6904097663031684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,12288,0.02985066506597731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,12288,0.027793778313530818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,4096,0.12789866659376356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,2048,0.04080977704789903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,65536,0.28754933675130206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,1536,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,2048,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,3072,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,3072,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,4096,0.00907555553648207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,768,0.054895109600490995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,32,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,512,0.011854222251309289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,5120,0.011445333560307821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,32,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,7168,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,16384,0.03844711184501648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,768,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,5120,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,51200,0.10199555423524644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,5120,0.18039111296335855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,12288,0.5772577921549479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,3584,0.011261333194043903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,7168,0.14883200327555338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,5120,0.02442933287885454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,32,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,64,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,512,0.0214035560687383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,4096,0.022662222385406494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,768,0.005405333307054307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,32,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,7168,0.0688577757941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,3072,0.025591999292373657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,128,0.009490666290124258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,5120,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,7168,0.037760890192455716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,16384,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,6144,0.11862488587697347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,512,0.9604987038506402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,32,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,32,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,64,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,2048,0.00666933341158761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,16384,0.06309777498245239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,6144,0.02479377720091078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,32,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,5120,0.009099555512269339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,12288,0.04986844460169474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,7168,0.01240711079703437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,16384,0.16938932736714682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,6144,0.3010355631510417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,2560,0.04227733280923632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,4096,0.03688622183269925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,3072,0.016010666886965435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,6144,0.012353777885437012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,1024,0.0800115532345242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,10240,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,1536,0.04043911231888665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,3584,0.023289778166347083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,3584,0.024456888437271118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,10240,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,2560,0.031896889209747314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,128,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,32,0.005687110953860813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,768,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,2560,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,512,0.005379555539952383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,8192,0.17499021689097086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,5120,0.09362844626108806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,6144,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,1024,0.02941511074701945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,5120,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,512,0.02032088902261522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,10240,0.03245155678855048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,16384,0.023822221491071913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,2560,0.0460382236374749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,16384,0.026712000370025635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,5120,1.0478542115953233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,2048,0.024937777055634394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,768,0.2531920009189182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,3584,0.05471466647254097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,4096,0.11368266741434734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,1024,0.0028817777832349143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,6144,0.2946053345998128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,64,0.006653333289755716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,3584,0.007730666961934831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,3584,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,2048,0.0618835555182563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,32,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,16384,0.17139200369517008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,2048,0.012376888758606382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,65536,0.41655556360880536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,1024,0.026891556051042344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,65536,1.0249030854966905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,768,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,4096,0.02720622221628825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,65536,0.20687288708157012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,512,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,8192,0.10353600316577488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,1024,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,7168,0.019839111301634047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,3072,0.006204444501135085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,1024,0.012148444851239523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,3072,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,12288,0.03435555431577895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,128,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,64,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,3072,0.005416000054942236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,3584,0.008135110967689091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,10240,0.017521777086787753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,5120,0.02436977790461646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,12288,0.03082044588194953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,2048,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,16384,0.11904533704121907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,768,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,256,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,64,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,51200,0.14172266589270696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,4096,0.27170933617485893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,64,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,768,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,768,0.009736888938479954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,7168,0.034076445632510714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,2048,0.009174221919642555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,12288,0.272237327363756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,256,0.1642400026321411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,128,0.041328890456093684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,768,0.017066667477289837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,16384,0.09487199783325195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,3584,0.020256888535287645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,2048,0.031223111682467993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,7168,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,16384,0.198834670914544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,512,0.007243555453088548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,5120,0.17426844437917074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,5120,0.046246220668156944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,5120,0.1877866718504164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,51200,0.38834312227037215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,4096,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,7168,0.10966489050123428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,7168,0.04418666826354133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,768,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,128,0.0035217776894569397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,128,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,128,0.00923999978436364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,32,0.0022373333987262514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,32,0.002114666625857353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,8192,0.17827644613054064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,16384,0.09118222528033787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,5120,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,10240,0.027666666441493567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,8192,0.03179644544919332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,51200,0.10886310868793064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,7168,9.01914554172092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,1536,0.022954665952258643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,16384,0.12710132863786486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,128,0.0707004467646281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,51200,0.25916798909505206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,32,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,8192,0.026759111218982275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,1024,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,16384,0.0888613329993354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,1536,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,12288,0.1653368870417277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,512,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,2560,0.028901331954532202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,128,0.03238222334120009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,4096,0.33289066950480145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,5120,0.03719022207789951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,4096,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,2560,0.0075502221783002215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,65536,0.08266044325298734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,10240,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,16384,0.36963465478685165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,2048,0.006715555571847492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,2560,0.027031110392676458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,2560,0.022272000710169475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,16384,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,128,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,65536,1.163608021206326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,65536,36.11005995008681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,4096,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,128,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,65536,4.891601986355252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,65536,0.2851395606994629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,5120,0.017332444588343304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,4096,0.06321333514319526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,32,0.0032675556010670135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,2048,0.04343822267320421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,256,0.02437955637772878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,4096,0.02236266599761115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,3584,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,16384,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,5120,0.09580533372031318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,65536,0.13586755593617758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,3584,0.027093332674768236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,1024,0.03941866755485535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,16384,0.5512293179829916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,128,0.017941332525677152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,32,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,51200,0.3107191191779243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,65536,1.5139457914564345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,128,0.03731377919514974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,4096,0.05821600225236681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,3584,0.04065422217051188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,32,0.023943111300468445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,64,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,10240,0.5000569025675455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,4096,0.018631999691327412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,768,0.01657777859105004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,2048,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,8192,0.36762134234110516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,2560,0.1050880021519131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,3072,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,2560,0.022463111413849726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,16384,0.01588088936275906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,1024,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,1536,0.0088755554623074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,2048,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,1536,0.014148443937301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,6144,0.030164443784289893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,12288,0.06918933656480578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,3072,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,64,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,1024,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,7168,0.7372693485683866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,2560,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,1024,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,16384,0.12441421879662408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,1024,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,6144,0.034049779176712036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,8192,0.07876888910929362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,2560,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,8192,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,768,0.04272888766394722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,128,0.005413333161009683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,64,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,12288,0.42452889018588597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,768,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,51200,1.9265528784857855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,128,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,51200,0.06481155421998766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,256,0.007190222541491191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,768,0.010784000158309937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,16384,0.07479733228683472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,512,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,12288,0.04527644316355387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,256,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,1024,0.0184542222155465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,128,0.001968888979819086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,65536,0.09988977511723836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,1536,0.030288888348473444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,3072,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,7168,0.02832533253563775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,51200,0.049088888698154025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,512,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,6144,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,3584,0.00942133367061615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,2560,0.027821332216262817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,16384,0.04429955614937676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,65536,0.08857955535252889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,7168,0.11696622106764053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,128,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,768,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,4096,0.02055999967787001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,64,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,8192,0.033640000555250384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,1536,0.022759111391173467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,51200,0.1564071045981513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,2048,0.018594667315483093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,1536,0.008460444708665213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,32,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,2560,0.08706577618916829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,2560,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,1536,0.00741155528359943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,12288,0.03795199924045139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,16384,0.12386755148569743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,16384,0.09187733464770848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,16384,0.020978666014141504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,8192,0.2807706726921929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,6144,0.012880889077981314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,2048,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,2560,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,5120,0.04748533169428507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,32,0.04131111171510484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,512,0.03068800104988946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,6144,0.51511197619968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,65536,0.06326844294865926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,64,0.016591111818949383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,128,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,3584,0.010064889159467485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,1024,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,512,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,2048,0.0101724440852801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,64,0.004201777693298128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,65536,0.30758221944173175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,7168,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,51200,0.7588648796081543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,1536,0.047503110435273915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,6144,0.01089333328935835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,51200,1.9658436245388453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,3584,0.06529333194096883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,128,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,512,0.04642311069700453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,2560,0.00871733327706655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,51200,0.105003555615743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,512,0.01626044511795044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,1024,0.014163555370436775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,128,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,8192,0.05906666649712456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,3072,0.038768000072903104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,10240,0.021306667062971327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,256,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,768,0.003220444545149803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,3584,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,3584,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,768,0.02834133307139079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,768,0.011963555382357704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,1536,0.004249777644872665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,10240,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,256,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,10240,0.02052622205681271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,4096,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,8192,0.048287110196219556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,8192,0.024880889389250014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,8192,0.026543999711672466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,512,0.11006310913297866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,256,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,256,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,1024,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,65536,0.2275102138519287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,256,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,3584,0.01804266704453362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,3072,0.03866577810711331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,51200,4.8623402913411455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,2048,0.03750400079621209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,7168,0.010189333723651038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,12288,0.093322667810652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,51200,0.13505155510372585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,6144,0.014175999495718213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,3072,0.05041955577002632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,512,0.0041662222809261745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,10240,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,512,0.011164444188276926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,16384,0.08581066793865627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,32,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,3584,0.025174222058720056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,4096,0.01830400029818217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,12288,0.29996710353427464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,65536,61.797329372829864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,2048,0.01053600013256073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,3584,0.026575111680560645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,6144,0.04219377703136868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,512,0.0026364444444576898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,4096,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,51200,0.060136000315348305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,2048,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,768,0.022294221652878657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,128,0.09035733011033799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,4096,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,6144,0.021242666575643752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,768,0.03594755464129978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,8192,0.1610906653934055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,16384,0.009593778186374241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,10240,0.012643555800120035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,10240,0.08927910857730442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,512,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,10240,0.09997333420647515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,7168,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,8192,0.014479110638300577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,5120,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,5120,0.19539999961853027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,768,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,512,0.04746577805942959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,768,0.023562666442659166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,2048,0.014059555199411182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,10240,0.0712968905766805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,5120,0.008903111020723978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,2560,0.08502044280370076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,1536,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,3584,0.010497777826256223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,65536,0.8592098024156359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,16384,0.029948443174362183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,3584,0.08372000190946792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,65536,0.4574871063232422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,256,0.01199288914601008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,512,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,256,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,7168,0.0587937765651279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,8192,0.03212177753448486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,2048,0.02957688768704732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,128,0.0058551111982928375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,65536,0.1990577777226766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,256,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,12288,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,1024,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,12288,0.05564888980653551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,16384,0.02088000045882331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,12288,0.34031200408935547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,512,0.027621333797772724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,5120,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,2560,0.06675289074579875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,4096,0.023248889380031165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,7168,0.009763555394278632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,4096,0.046888887882232666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,10240,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,65536,0.5201386875576443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,4096,0.02700800034734938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,2560,0.08723466926150852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,128,0.009503111243247986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,4096,0.024127999941507976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,2048,0.017370666066805523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,2048,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,256,0.009391110804345872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,5120,0.0224897778696484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,128,0.007694222033023834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,32,0.001822222231162919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,64,0.0018808889306253856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,7168,0.03457155492570665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,512,0.02152266601721446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,16384,0.033952001068327166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,3584,0.013816888961527081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,768,0.01927911076280806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,51200,0.06845777564578585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,512,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,64,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,51200,0.40979644987318253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,4096,0.017584888471497428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,4096,0.039684444665908813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,768,0.004334222111437055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,256,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,3072,0.012881778180599213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,65536,1.0488630930582683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,2560,0.009747555686367882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,1024,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,256,0.03216177887386746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,768,0.014246222045686511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,1536,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,8192,0.24153510729471842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,4096,2.012362586127387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,8192,0.022848889231681824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,1024,0.007054222126801808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,2048,0.06243555413352119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,4096,0.37914310561286074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,6144,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,16384,0.04316799839337667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,2560,0.010098666780524785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,4096,0.008441777692900764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,12288,0.045751998821894325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,2560,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,51200,0.08023111025492351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,3072,0.0109120003051228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,3072,0.2230008973015679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,7168,0.054210666153166026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,7168,0.13372445106506348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,3584,0.01590044465329912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,8192,0.020590222544140287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,5120,0.013726222018400827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,12288,0.042040000359217324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,3072,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,6144,0.0373342235883077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,1024,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,3072,0.03873422079616123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,768,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,12288,0.04629777868588766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,768,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,128,0.02327288852797614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,3584,0.1742853323618571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,5120,0.016935111747847665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,10240,0.016214221715927124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,4096,0.01643111142847273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,16384,0.05063822203212314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,2560,0.019661333825853135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,2560,0.09034844239552815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,10240,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,3072,0.1977555619345771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,12288,0.7589208814832898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,64,0.005490666462315454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,5120,0.022175111704402502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,3584,0.05458933446142408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,2560,0.19314222865634492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,16384,0.021320000290870667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,8192,0.4487279786003961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,512,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,64,0.02834933333926731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,7168,0.1371119949552748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,12288,0.059789333078596324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,8192,0.39092355304294163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,3072,0.022907555103302002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,768,0.02148177723089854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,3584,0.03444088829888238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,10240,0.016568000117937725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,8192,0.012915555801656512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,1536,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,32,0.024448888169394598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,65536,0.04908355408244663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,768,0.014501333236694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,128,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,512,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,5120,0.09740266535017227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,5120,0.4593386650085449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,32,0.023408888114823237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,7168,0.02273599969016181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,1536,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,768,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,65536,0.2821253405676948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,3072,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,64,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,5120,0.4262231190999349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,65536,0.0770791106753879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,768,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,5120,0.006527110934257507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,8192,0.11660799715254043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,5120,0.013840888937314352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,1536,0.3987662262386746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,768,0.03643911083539327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,3072,0.04007200068897671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,16384,0.0654639999071757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,1536,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,5120,0.02184355590078566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,2048,0.015749333633316886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,512,0.010811555716726514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,10240,0.041307555304633245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,51200,0.5158444510565864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,256,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,3072,0.027393778165181477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,32,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,256,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,512,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,16384,0.021278222401936848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,3584,0.07514578104019165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,128,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,12288,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,1024,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,51200,0.05310666561126709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,512,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,64,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,1536,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,3584,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,5120,0.0622311102019416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,65536,0.21720266342163086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,51200,0.4977848794725206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,4096,0.016063110695944894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,7168,0.041109333435694374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,7168,0.17324977450900605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,16384,0.06528000036875407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,51200,0.1236471070183648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,10240,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,3584,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,12288,0.03882133298450046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,2560,0.032187554571363665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,64,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,256,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,64,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,65536,0.060918225182427295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,8192,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,1536,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,768,0.20913333363003203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,65536,0.07865599791208903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,8192,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,8192,0.13896444108751085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,1536,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,12288,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,8192,0.6387404335869683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,4096,0.02660711109638214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,128,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,2048,0.009357333183288574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,128,0.004960000101062986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,4096,0.0389395554860433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,128,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,4096,0.0147724449634552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,1536,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,6144,0.42918756273057723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,2048,0.011612444288200803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,65536,0.025456888808144465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,4096,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,3072,0.016516443755891588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,4096,0.015430221954981485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,3072,0.04538133409288195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,2560,0.01143022212717268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,8192,0.06606844398710462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,3584,0.04383466641108195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,8192,0.01994044416480594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,1024,0.025860443711280823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,1024,0.009292444421185387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,16384,0.03351199958059523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,65536,0.01593955523437924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,2560,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,65536,0.0966000027126736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,2560,0.3108631239997016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,5120,0.11132266786363389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,2048,0.020249777370029025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,64,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,6144,0.02896266513400608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,12288,0.5121359825134277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,12288,0.02499733368555705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,128,0.010442666709423065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,256,0.014797333214018079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,51200,0.24702933099534777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,16384,0.10841777589586045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,10240,0.011557333171367645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,4096,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,128,0.0026142222599850762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,3072,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,8192,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,4096,0.007178666690985362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,1024,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,12288,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,12288,0.10477955473793878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,3072,0.01993600030740102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,3072,0.006196444647179709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,4096,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,4096,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,1536,0.009872000250551436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,768,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,128,0.01019200020366245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,5120,0.07174755467308892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,3584,0.01886577738655938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,3584,0.06857599814732869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,1536,0.05814844369888306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,3584,0.047876444127824574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,12288,0.05637866920895047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,4096,0.00813066628244188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,1024,0.017450667089886136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,3072,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,12288,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,2048,0.04099466734462314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,512,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,256,0.011689778003427716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,10240,0.028399109840393066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,768,0.005151111218664381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,1024,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,2560,0.006824888702895906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,1536,0.012688000169065265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,3072,0.013349333571063148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,10240,0.05931911203596327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,2048,0.05870488617155287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,10240,0.012940444052219391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,1536,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,12288,0.020607999629444547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,65536,0.5404844284057617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,3584,0.06270933151245117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,3072,0.019323555959595572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,64,0.008812444077597724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,64,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,3072,0.020241777102152508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,12288,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,2560,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,4096,0.08191911379496257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,6144,0.009637333452701569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,2560,0.08589600192175971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,51200,0.10109421941969131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,4096,0.01314933349688848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,1536,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,64,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,1536,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,128,0.01015644437736935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,2560,0.022658665974934895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,64,0.003474666840500302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,128,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,8192,0.08063466681374444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,2048,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,5120,0.0344142218430837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,10240,0.02215733296341366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,64,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,512,0.005491555564933353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,1024,0.3041200107998318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,512,0.007329777710967594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,512,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,12288,0.2230559984842936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,6144,0.03492444422509935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,512,0.17820710606045195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,64,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,7168,0.013914666242069669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,10240,0.03422844409942627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,256,0.009527111219035255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,5120,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,16384,0.3466595543755426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,3072,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,32,0.0020879999630981022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,768,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,51200,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,16384,0.035544001393847995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,1536,0.013386666774749756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,5120,0.03772977656788296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,256,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,4096,0.04320533408059014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,1024,0.015172445111804538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,768,0.028833776712417603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,512,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,65536,0.21890578005048963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,3584,0.04314133193757799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,5120,0.02212000058756934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,768,0.03568444318241543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,16384,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,32,0.007585778004593319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,6144,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,5120,0.0210533340771993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,10240,0.06739377975463867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,7168,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,3072,0.027081777652104694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,6144,0.16177777449289957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,12288,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,6144,0.03306666678852505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,768,0.025722665919197932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,7168,0.3439653449588352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,3584,0.02154133386082119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,64,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,256,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,512,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,8192,0.013008000122176277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,12288,0.018260444204012554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,3584,0.07783289088143243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,2048,0.03686133358213637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,256,0.02163999941613939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,12288,0.03237066666285197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,1024,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,51200,1.5055288738674586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,16384,0.10989777909384833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,64,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,3072,0.015394666128688388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,3072,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,65536,0.3184346622890896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,1536,0.018761777215533786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,768,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,10240,0.01236088905069563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,32,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,5120,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,2048,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,512,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,256,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,3584,0.254493342505561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,12288,0.22907643847995332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,1024,0.06552533308664958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,2048,0.025427555044492085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,3072,0.16455554962158203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,12288,0.07206933365927802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,65536,1.5336799621582031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,12288,0.01679377754529317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,64,0.0017608888447284698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,2560,0.015712888704405892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,12288,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,2560,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,3584,0.015772443678643968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,10240,0.019169777631759644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,1024,0.006432000133726332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,3584,0.012111999922328524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,1536,0.006207110981146495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,6144,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,65536,0.06742488675647311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,3584,0.07831644349628024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,768,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,2048,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,32,0.06789688931571113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,1024,0.03132088979085287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,16384,0.32194222344292533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,3584,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,2048,0.07047644588682386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,128,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,512,0.03394844465785556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,512,0.020277332928445604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,6144,0.20669866932762992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,1536,0.0038311113086011675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,65536,0.09028889073265924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,768,0.017503110898865592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,51200,0.4918950928582086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,512,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,5120,0.007889777421951294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,3072,1.8856258392333984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,1024,0.007506666911972894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,64,0.011764444410800934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,5120,0.016055110428068373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,768,0.015903110305468243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,512,0.013364444176355997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,4096,0.03827644387880961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,65536,0.21819289525349936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,12288,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,51200,0.02073955535888672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,8192,0.14168443944719103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,7168,0.22353688875834146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,65536,0.542285336388482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,10240,0.04837422238455879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,7168,0.014167111780908374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,4096,0.09467289182874893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,6144,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,1536,0.012019555601808759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,7168,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,7168,0.05463022324774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,16384,0.03922933340072632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,32,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,512,0.0481608874268002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,512,0.04656444324387444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,8192,0.023026666707462732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,5120,0.05482044484880236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,64,0.10826488998201157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,768,0.007124444676770105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,5120,0.8777999877929688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,65536,0.057796445157792836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,512,0.0196142229768965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,10240,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,512,0.02387111054526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,7168,0.019221333993805777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,6144,0.025191111697090998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,64,0.03439022103945414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,8192,0.010497777826256223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,4096,0.01479111115137736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,128,0.0025742221623659134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,64,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,256,0.025316443708207872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,7168,0.03928088810708787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,64,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,2560,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,32,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,1536,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,3072,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,6144,0.008019555773999956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,128,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,65536,0.6441111034817165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,12288,0.058063109715779625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,4096,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,1024,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,7168,0.02349777850839827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,256,0.0038817777401871155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,6144,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,3584,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,12288,0.2584799925486247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,2048,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,10240,0.016833777228991192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,6144,0.040419555372662015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,1536,0.012599999705950419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,64,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,6144,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,1536,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,16384,0.029623111089070637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,65536,0.03480444351832072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,16384,15.157480875651041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,5120,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,1536,0.017265778448846605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,7168,0.011398221883508893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,3584,0.5165395736694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,64,0.00240622212489446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,51200,1.1598196029663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,8192,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,16384,0.25004445181952584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,4096,0.011392888923486074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,51200,0.19254488415188262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,51200,0.717374218834771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,512,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,16384,0.29599377844068736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,8192,0.11931733290354411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,8192,0.011315555208259158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,10240,0.059841778543260365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,512,0.028802666399214003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,10240,0.07010310888290405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,3584,0.0440942214594947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,51200,0.24448710017734102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,12288,0.36192888683742946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,4096,0.06735555330912273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,512,0.011439999772442712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,5120,0.01718755563100179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,128,0.012283555335468717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,7168,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,256,0.17626578278011748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,65536,1.694210688273112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,2560,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,3072,0.3140710989634196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,51200,0.7158782217237684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,7168,0.010608889162540436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,512,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,16384,3.303475485907661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,1536,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,6144,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,128,0.01704266667366028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,16384,0.029504001140594482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,1536,0.006745777610275481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,1536,0.003527111063400904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,2560,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,256,0.017234666479958426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,65536,2.381284501817491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,2048,0.3892177740732829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,1024,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,32,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,65536,0.060510224766201444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,2560,0.32332711749606663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,16384,0.07611199882295397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,16384,0.0636319981680976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,2560,0.10759022500779893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,1536,0.0038159998754660287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,5120,0.03922577699025472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,10240,0.03577422102292379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,51200,0.6449208789401585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,3584,0.03239999877081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,2560,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,1536,0.009924444059530893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,768,0.007424888511498769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,51200,0.8569466802808973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,3072,0.052520000272327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,256,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,6144,0.02162488963868883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,12288,0.23589865366617838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,32,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,16384,2.2816694047715926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,2048,0.031501332918802895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,8192,0.0679253339767456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,10240,0.033137778441111244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,512,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,512,0.10014222065607707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,2560,0.021634666456116572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,3072,0.028753777345021565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,51200,4.510133531358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,2560,0.011224889092975192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,6144,0.04480888777308994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,51200,6.160804324679905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,8192,0.13720622327592638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,128,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,4096,0.021186666356192693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,256,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,512,0.018851555056042142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,5120,0.07675733168919881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,5120,0.017543110582563613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,5120,0.9173297882080078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,32,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,256,0.023364444573720295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,256,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,768,0.02794044547610813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,768,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,12288,0.004957333207130432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,3584,0.016021332807011075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,2560,0.19400444295671251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,128,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,6144,0.02016177773475647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,768,0.018629334039158292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,10240,0.03942577706442939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,10240,0.3968782160017226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,10240,0.01953155630164676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,16384,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,2048,0.01351466692156262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,6144,0.10180799828635322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,4096,0.006666666517655055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,10240,0.09843555423948501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,1024,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,128,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,12288,0.031103110975689355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,7168,0.029486222399605647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,7168,0.16759555869632295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,512,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,51200,0.677696016099718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,512,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,1536,0.04215111003981697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,1024,0.017494221528371174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,256,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,65536,0.2771911091274685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,10240,0.11028177870644464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,7168,0.021006221572558086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,10240,0.10840177536010742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,768,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,6144,0.3311084376441108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,512,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,2048,0.004293333325121137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,16384,0.5158329010009766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,3072,0.04141689009136624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,7168,0.13282488452063665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,7168,0.04895022180345324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,6144,0.03144977821244134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,65536,0.0568444463941786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,12288,0.0147724449634552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,128,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,512,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,3584,0.027866668171352808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,4096,0.026586666703224182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,7168,0.08043911059697469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,4096,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,8192,0.06429955694410536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,5120,0.28014132711622447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,10240,0.011187555889288584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,512,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,512,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,5120,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,1024,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,512,0.013869333598348828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,1024,0.0046453335218959385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,5120,0.046550220913357206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,1024,2.557552973429362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,10240,0.11532089445326064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,64,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,51200,0.12807377179463705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,2560,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,3584,0.024632889363500807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,256,0.0348364445898268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,1536,0.009884444375832876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,65536,0.12944977813296848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,12288,0.049569778972201876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,65536,0.1926595502429538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,2560,0.056367112530602344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,128,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,5120,0.060284442371792264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,3584,0.3170479933420817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,65536,0.14703555901845297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,7168,0.649504025777181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,768,0.013221333424250284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,12288,0.04305244485537211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,128,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,65536,0.07700977722803752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,16384,0.016009777784347534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,2560,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,128,1.0789066950480144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,3072,0.0090844440791342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,32,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,5120,0.05172444383303324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,3584,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,2048,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,5120,0.015556444724400839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,1536,0.005315555466545953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,3584,0.014018666413095264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,51200,0.3643066618177626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,3072,0.03977333174811469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,12288,0.054309331708484225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,16384,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,3072,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,64,0.016295111841625638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,4096,0.025683555338117812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,16384,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,768,0.0026657777941889236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,7168,0.018927999668651156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,1024,0.019492444064882066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,128,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,2560,0.2396248976389567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,3584,0.013942221800486246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,1536,0.019108444452285767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,512,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,16384,0.1320693360434638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,4096,0.027886221806208294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,2048,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,7168,0.030047999487982854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,1024,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,512,0.0074133334888352295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,16384,0.4973137643602159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,128,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,16384,0.06710666418075562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,256,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,128,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,3584,0.017567111386193168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,64,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,256,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,1024,0.01683200067943997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,7168,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,12288,0.026339555780092876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,2560,0.06528888808356391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,16384,0.023578666978412207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,1536,0.012732444538010491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,3584,0.007888889147175683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,1024,0.22628090116712782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,3584,0.08364799949857925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,10240,0.07333955499860975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,64,0.005991999887757831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,64,0.004456888884305954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,1024,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,2048,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,768,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,512,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,12288,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,3584,0.03878133495648702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,51200,0.06014044417275322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,512,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,2560,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,2560,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,1536,0.020234665936893888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,8192,0.03197244471973843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,128,0.011863110793961419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,5120,0.0633831090397305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,8192,0.08560888634787665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,512,0.023983110984166462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,2560,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,768,0.021031111478805542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,6144,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,6144,0.2545822196536594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,5120,0.011341333389282227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,12288,0.20851466390821669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,768,0.0038240001433425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,3584,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,1024,0.018047110901938546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,512,0.09724533557891846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,7168,0.06759644216961332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,3072,0.03747466537687514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,3584,0.10338577959272598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,512,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,3584,0.004921777794758479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,64,0.0053280000057485365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,3072,0.14883822864956325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,128,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,4096,0.019352888067563374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,2048,0.014928888943460254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,2048,0.05203910999827915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,256,0.012160888976520963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,65536,0.3282079961564806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,6144,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,256,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,1536,0.013969777358902825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,7168,0.017581333716710407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,4096,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,65536,0.127329773373074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,2048,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,1024,0.02737066646416982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,16384,0.42968000306023496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,1536,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,3072,0.05734399954477946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,8192,0.01148088855875863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,256,0.025732444392310247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,12288,0.018444443742434185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,32,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,10240,0.01793511046303643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,2048,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,128,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,12288,0.02416177756256527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,6144,0.13263111644321018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,8192,0.02300888962215847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,65536,0.0655955539809333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,3072,0.022152889106008742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,4096,0.007719111111429002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,256,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,768,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,4096,0.05884710947672526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,2560,0.03921688927544488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,32,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,6144,0.13854311572180852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,1536,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,4096,0.03631644447644552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,5120,0.08597422308391994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,2048,0.018535999788178336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,4096,0.3682071102990045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,2560,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,16384,0.04074044360054864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,128,0.010083555347389644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,512,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,12288,0.018383999665578205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,51200,0.05096266667048136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,12288,0.018746667438083224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,2048,0.007532444265153673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,16384,1.3302115334404838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,7168,0.05109599894947476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,12288,0.02741866641574436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,768,0.01891644464598762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,8192,0.2740746604071723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,2048,0.006515555497672822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,51200,0.6183910899692112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,4096,0.27373867564731175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,10240,0.017907554904619854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,7168,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,51200,0.25337423218621147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,7168,0.019923556182119582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,1024,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,2560,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,8192,0.2183155616124471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,4096,0.04424533247947693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,32,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,2048,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,6144,0.03600355651643541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,1536,0.0035786665976047516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,6144,0.30742311477661133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,16384,0.03258133265707228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,16384,0.05000266763899061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,51200,0.039557334449556135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,512,0.01773599949147966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,51200,0.08315733406278822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,1024,0.016398222910033334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,51200,0.09760710928175186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,7168,0.08923555745018853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,32,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,10240,0.03534311056137085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,4096,0.01924088928434584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,256,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,128,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,1024,0.008182221816645728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,4096,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,3072,0.07169510920842488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,10240,0.0173653331067827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,51200,0.16651999950408936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,2560,0.07750311162736681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,8192,0.09866844283209907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,3072,0.009194666312800514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,2048,0.017608889275126986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,128,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,5120,0.01018044435315662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,16384,0.19406044483184814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,1536,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,768,0.011400889191362592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,16384,0.02401155564520094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,2048,0.04389422138532003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,65536,0.15151111284891763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,2560,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,3072,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,5120,1.295540491739909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,7168,0.022221333450741235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,32,0.005724444571468566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,6144,0.01573422220018175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,7168,0.2545493443806966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,512,0.005421333428886201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,1024,0.0030008889734745026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,10240,0.36050844192504883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,512,0.05720444520314535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,256,0.022034666604465906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,65536,0.08783644437789917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,128,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,2560,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,4096,0.23407377137078178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,1024,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,2048,0.03565422362751431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,5120,0.01614755557643043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,16384,0.02996444371011522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,5120,0.6711929109361437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,2048,0.04337244563632541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,3072,0.009147555463843876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,2048,0.031160000297758315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,512,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,1536,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,1024,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,3584,0.06057422028647529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,3072,0.06346578068203397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,2560,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,51200,0.332623110877143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,1536,0.07092710998323229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,256,0.014230221509933472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,65536,1.2984497282240126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,3072,0.06109866831037733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,512,0.021493333909246657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,6144,0.16865511735280356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,16384,0.023040889037979975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,2048,0.03541422221395705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,65536,0.04722133278846741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,16384,0.0552151103814443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,16384,0.13431199391682944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,1536,0.012789333032237159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,256,0.018485332528750103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,256,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,256,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,16384,0.03158844510714213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,2560,0.15552622742123073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,4096,0.02366399930583106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,4096,0.06270933151245117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,32,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,2560,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,1536,0.08224355512195163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,7168,0.010786666638321347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,51200,0.11455821990966797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,6144,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,4096,0.008026666939258575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,6144,0.014352889524565803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,12288,0.08238933483759563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,256,0.0035075553589397003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,8192,0.046562665038638644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,256,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,65536,0.07185244560241699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,10240,0.027719999353090923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,6144,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,16384,0.015777778294351365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,2560,0.011168888873524137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,1024,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,65536,19.50216335720486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,65536,0.12730311022864446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,16384,0.07274311118655734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,8192,0.033928000264697604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,768,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,5120,0.016068443655967712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,5120,0.011639110743999481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,32,0.08488000101513332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,10240,0.010026666853162978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,12288,0.11379822095235188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,16384,0.023215111758973863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,16384,0.028601778878106013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,6144,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,7168,0.01427288850148519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,256,0.01585155559910668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,128,0.01145155562294854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,1024,0.013812444276279874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,51200,0.017528888252046373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,2560,0.005152889009979037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,256,0.005311111195219888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,4096,0.02144355575243632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,2048,0.022115555074479844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,65536,0.05728977918624878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,5120,0.029862221744325426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,12288,0.42171822653876406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,32,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,128,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,10240,0.5673537784152561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,2560,0.09357688824335735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,16384,0.009257777697510189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,65536,0.33230310016208225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,8192,0.1452488899230957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,1536,0.017138666576809354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,12288,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,4096,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,12288,0.2717813385857476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,128,0.01715555621518029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,512,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,5120,0.08861244387096828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,1536,0.007706666986147563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,2048,0.007416889071464539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,2048,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,2560,0.030615998639000788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,4096,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,7168,0.010239111052619087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,1024,0.03749688797526889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,128,0.0022462221483389535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,256,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,512,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,2048,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,10240,0.017337777548366122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,51200,0.012659555508030785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,5120,0.01607822212908003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,1536,0.05062311225467258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,2560,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,768,0.006339555399285422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,256,0.005696889013051987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,2048,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,768,0.011572444604502784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,2560,0.02094399929046631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,32,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,12288,0.006770666688680649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,512,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,4096,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,4096,0.034767998589409724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,2560,0.010056888891590966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,5120,0.011914666328165265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,12288,0.37145421240064835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,5120,0.08571733368767633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,2048,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,32,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,3584,0.4491386413574219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,51200,4.528304206000434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,12288,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,16384,0.41417778862847227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,65536,0.147871110174391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,5120,0.04500888784726461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,12288,0.11653333240085179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,1536,0.003593777616818746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,5120,0.045478221442964345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,6144,0.12054667207929824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,3584,0.027608889672491286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,5120,0.054642667373021446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,8192,0.01959911154376136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,16384,0.12514044178856745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,2560,0.014366222752465142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,4096,0.02775466607676612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,3584,0.032481776343451604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,65536,0.07918310827679105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,4096,0.07067644596099854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,1024,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,1536,0.0641546646753947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,8192,0.030841777722040813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,51200,0.64829068713718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,16384,0.08442488643858169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,768,0.01965777741538154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,12288,0.046329776446024575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,1024,0.1600666708416409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,2560,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,1024,0.057095110416412354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,16384,0.0321324434545305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,1024,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,12288,0.03455377618471781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,768,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,12288,4.8293173048231335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,8192,0.015557333827018738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,512,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,3072,0.04298399885495504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,4096,0.03651466634538438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,51200,0.5597013367546929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,16384,0.021928888228204515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,3584,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,1024,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,768,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,2560,0.019855111837387085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,2048,0.021295999487241108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,4096,0.09150044123331706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,5120,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,2560,0.016827555166350473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,12288,0.11162400245666504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,2048,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,10240,0.5573200119866265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,1024,0.007096889118353526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,3072,0.05229155553711785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,7168,0.847358226776123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,8192,0.1800453397962782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,7168,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,512,0.014500444134076437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,65536,0.05052888724539015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,5120,0.018026666508780587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,6144,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,65536,0.09594133165147568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,10240,0.02163999941613939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,3072,0.018929777873886954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,6144,0.03390311201413473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,65536,0.020217777954207528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,768,0.03400977783732944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,6144,0.012828444441159567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,768,0.006704888823959563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,256,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,51200,0.12735377417670354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,4096,0.014892444014549255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,12288,0.036955555280049644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,1536,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,3072,0.016594666573736403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,512,0.02460533380508423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,51200,2.583754645453559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,1536,0.03769866625467936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,16384,0.05730488565233019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,2048,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,128,0.022894221875402663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,12288,0.13550755712721083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,5120,0.03589955634540982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,5120,0.05876266956329346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,7168,0.0192266669538286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,7168,0.03153866529464722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,6144,0.11416444513532852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,3584,1.07895999484592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,768,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,256,0.03844889005025228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,128,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,2048,0.015105777316623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,512,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,256,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,512,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,2560,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,3584,0.007346666521496243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,768,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,6144,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,16384,0.3676924440595839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,1536,0.030371556679407757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,1024,0.008080888953473832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,1024,0.015488889482286243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,5120,0.015239111251301236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,768,0.01165422217713462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,10240,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,10240,0.0522328880098131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,5120,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,12288,0.3139413462744819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,128,0.0023226667609479693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,768,0.006076444354322221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,16384,0.30528444714016384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,10240,0.10105866856045193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,7168,0.12240444289313422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,3584,0.0284826656182607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,16384,0.049591110812293164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,128,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,256,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,3584,0.007932444413503012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,512,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,1536,0.03810311026043362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,2560,0.0070159998204973005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,16384,0.0428711109691196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,12288,0.04710844490263197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,12288,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,51200,0.5008959770202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,256,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,768,0.05307022068235609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,512,0.013800889253616333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,2560,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,7168,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,256,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,768,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,128,0.05087555448214213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,51200,0.4054879877302382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,12288,0.015217777755525378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,3072,0.01800355480776893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,1024,0.01815555493036906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,6144,0.011642666326628791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,4096,0.012504888905419244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,12288,2.584226608276367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,768,0.00790666706032223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,12288,0.06968977716233996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,3072,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,3584,0.016561778055297006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,256,0.018566222654448617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,3584,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,4096,0.01676266723208957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,1024,0.08424444331063165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,3072,0.05303111010127597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,1024,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,3584,0.030992888742023047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,5120,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,3072,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,2048,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,1024,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,10240,0.07047200202941895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,7168,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,768,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,128,0.01458755632241567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,7168,0.09084266424179077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,5120,0.010420444111029306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,2560,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,5120,0.009380444056457942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,4096,0.02294933299223582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,8192,0.04346399837070041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,512,0.011168888873524137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,128,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,256,0.017218665944205392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,1536,0.015829333000712924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,768,0.02864711152182685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,10240,0.03523733218510946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,7168,0.033773332834243774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,10240,0.02607111136118571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,12288,0.0232595553000768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,5120,0.020592888196309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,8192,0.03535822365019057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,1024,0.0063484443558586975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,1536,0.016516443755891588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,12288,0.24053422609965006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,2560,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,8192,0.04220444295141432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,65536,0.7615662150912814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,16384,0.020605333977275424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,6144,0.02219466699494256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,7168,0.04234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,8192,0.022688888841205176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,10240,0.1806008948220147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,5120,0.05373422304789225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,12288,0.015429332852363586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,10240,0.004879111217127906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,2560,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,65536,0.10187200042936538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,2560,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,51200,2.1960347493489585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,16384,0.7310320006476508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,2560,0.020974222156736586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,512,0.01902044481701321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,6144,0.035001777940326266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,1024,0.06374222040176392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,51200,0.5796408653259277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,768,0.028255999088287354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,3072,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,512,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,6144,0.018568888306617737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,4096,0.059092442194620766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,65536,5.865319146050347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,6144,0.010026666853162978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,6144,1.1181315316094291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,128,0.007470221983061896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,7168,0.02630133264594608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,3584,0.025422222084469263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,12288,0.0421635541650984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,7168,0.01273155543539259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,4096,0.04310844341913859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,128,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,512,0.041858666472964816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,12288,0.0259253333012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,12288,5.616407182481553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,512,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,2048,0.0703715549574958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,2048,0.026138666603300307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,64,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,51200,0.059522668520609535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,7168,0.06308533085717095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,512,0.03005689051416185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,128,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,512,0.013741333451535968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,5120,0.05779200130038791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,10240,0.02762311034732395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,768,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,32,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,3584,0.01425866617096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,1536,0.02574488851759169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,4096,0.019119110372331407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,1024,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,65536,0.22163732846577963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,51200,0.0495031111770206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,65536,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,512,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,51200,0.14046843846638998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,1536,0.0038746665749284956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,1024,0.009747555686367882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,4096,0.010775999890433418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,10240,0.25841421551174587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,2048,0.026503110925356548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,3584,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,1024,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,10240,0.406314664416843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,64,0.12052622106340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,3072,0.014635556273990206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,7168,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,64,0.006406222366624408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,12288,0.045055111249287925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,2560,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,6144,0.13713867134518093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,32,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,256,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,8192,0.010267555713653564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,6144,0.060311112138960093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,768,0.007172444628344641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,32,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,3584,0.03470933437347412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,3072,0.08113244507047865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,6144,0.348491562737359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,16384,1.7368400361802843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,128,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,1536,0.017338666650984023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,2560,0.10719555616378784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,4096,0.14011377758449978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,5120,0.01959288948112064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,2560,0.007695111135641734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,2048,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,1024,0.003524444583389494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,1536,0.021391999390390184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,32,0.006353777729802662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,4096,0.022036444809701707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,4096,0.04689066608746847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,128,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,6144,0.02419022222359975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,8192,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,256,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,65536,0.7047084702385796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,768,0.0027831111931138565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,7168,0.03784977727466159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,128,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,65536,0.7584826681349012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,65536,1.5431635114881728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,12288,0.042303999265034996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,51200,0.7395093176099988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,5120,0.329715543323093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,10240,0.18838489055633545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,7168,0.10999111334482829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,3072,0.00666133314371109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,256,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,6144,0.014306666122542487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,5120,0.026160000099076167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,2048,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,65536,1.1326932907104492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,51200,0.07134222322040133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,512,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,8192,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,10240,0.02117777698569828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,3584,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,2560,0.01202222208182017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,128,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,1536,0.01107022249036365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,3072,0.055664002895355225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,12288,0.2935822274949816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,5120,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,5120,0.04803466796875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,6144,0.020276443825827706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,768,0.1475715504752265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,4096,0.026720888084835474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,256,0.016936888297398884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,12288,0.028796444336573284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,4096,0.024127110838890076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,64,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,8192,0.040016889572143555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,4096,0.24618132909138998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,8192,0.013738666971524557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,2048,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,7168,0.03125600020090739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,3584,0.03393155667516921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,128,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,51200,0.6028569009568956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,10240,0.011935110721323224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,2048,0.02596355477968852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,6144,0.017383999294704862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,51200,0.05771644247902764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,5120,0.012847999731699625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,51200,0.12120800548129612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,3072,0.00978488889005449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,7168,0.0836168858740065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,12288,0.028204444381925795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,32,0.0057848890622456866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,10240,0.10071555773417155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,12288,0.7670693397521973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,2560,0.018233777748213876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,4096,0.08720088667339748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,16384,0.023982221881548565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,128,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,16384,0.06329333119922213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,5120,0.04010933306482103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,8192,0.09459288914998372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,64,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,1536,0.05728977918624878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,1024,0.0041911109454101985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,5120,0.08363644282023112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,12288,0.353204435772366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,7168,0.015211555692884656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,5120,0.20470933119455972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,512,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,51200,0.152566221025255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,2560,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,3072,0.027089777919981215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,512,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,51200,0.4391893280877008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,3584,0.030343112018373277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,5120,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,2560,0.014895111322402954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,4096,0.021729777256647747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,16384,0.040279110272725425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,1536,0.013387555877367655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,12288,0.02920888861020406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,512,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,768,0.10074400239520603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,512,0.009520889156394536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,768,0.010703999963071613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,6144,0.01956355571746826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,6144,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,6144,0.07463822099897596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,512,0.022904000348514978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,1536,0.05376266770892673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,5120,0.10684000121222602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,1024,0.017779555585649278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,3072,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,1536,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,5120,0.09449244207806057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,6144,0.03372977839575873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,5120,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,64,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,512,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,32,0.001822222231162919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,4096,0.024579554796218872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,768,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,512,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,1536,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,12288,0.019612444771660697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,512,0.041952000723944775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,768,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,7168,0.029109332296583388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,6144,0.042040887806150645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,16384,0.06562844249937269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,1024,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,12288,0.21065778202480742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,4096,0.01720177796151903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,2560,4.701246049669054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,4096,0.10647377702924941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,1024,0.017695999807781644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,51200,0.2001297738817003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,10240,0.2305075592464871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,10240,0.021166221963034734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,1024,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,7168,0.03237955437766181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,16384,0.045490665568245776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,16384,0.0341635545094808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,51200,0.4995368851555719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,51200,0.1513919962777032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,1536,0.010573333336247338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,8192,0.0471937788857354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,512,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,256,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,51200,0.01734133395883772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,4096,0.04004089037577311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,512,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,2560,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,64,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,5120,0.11018044418758816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,8192,0.010891555911964841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,768,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,7168,0.37610933515760636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,512,0.0029075555503368378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,128,0.007000889215204451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,768,0.004836444639497333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,12288,0.7614577611287435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,8192,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,51200,0.037964443365732826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,256,0.019612444771660697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,768,0.03085511260562473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,6144,0.07384710841708712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,32,0.010279110736317104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,768,0.006048888795905643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,12288,0.06306222412321302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,5120,0.055472890535990395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,2560,0.09306933482487996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,7168,0.02896266513400608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,51200,0.07695377535290189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,2048,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,2560,0.014933332800865173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,65536,0.3092622227138943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,10240,0.11459644635518391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,1536,0.05574399895138211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,256,0.00598933340774642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,65536,0.1576168934504191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,2048,0.006384000182151794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,768,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,51200,0.362191120783488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,256,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,10240,0.01646311084429423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,32,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,7168,0.11623822318183051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,65536,0.062253336111704506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,65536,0.6810115708245171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,16384,0.3889644410875108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,768,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,2048,0.3697644339667426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,4096,0.03526755505137973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,768,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,1024,0.12662222650316027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,7168,0.10611199670367771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,8192,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,128,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,5120,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,10240,0.019327999817000497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,8192,0.013782222237851886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,5120,0.11345244778527154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,5120,0.012356444365448423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,7168,0.01367288910680347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,65536,0.2975084516737196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,256,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,4096,0.010086221827401055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,512,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,128,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,51200,0.07015999820497301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,3584,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,1536,0.007824000385072496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,12288,0.06552888949712117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,1536,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,7168,0.036277333895365395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,3072,0.10069955719841851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,3584,0.04387911160786947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,512,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,5120,0.10321778059005737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,8192,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,16384,0.05245866709285312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,256,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,2560,0.021931555536058214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,6144,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,7168,0.024359110328886244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,4096,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,16384,0.07554577456580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,16384,0.08171021938323975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,65536,2.0989661746554904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,51200,0.05903999672995674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,32,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,3072,0.036630223194758095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,12288,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,6144,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,12288,0.20858400397830537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,12288,0.013629333012633853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,8192,0.322376012802124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,4096,0.35173866483900285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,51200,7.824295891655816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,2048,0.2696177694532606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,10240,0.12407199541727702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,5120,0.015767110718621146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,16384,0.022572444544898138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,2048,0.005535999933878581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,1536,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,10240,0.0192248887485928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,7168,0.059030221568213574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,6144,0.015328889091809591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,4096,0.030694223112530176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,32,0.3181564542982313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,3584,0.009295110901196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,256,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,32,0.008229333493444655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,2048,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,1536,0.003524444583389494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,128,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,32,0.014083556003040738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,51200,0.264119995964898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,3072,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,12288,0.20775732729170057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,1536,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,51200,0.017785777648289997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,256,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,2560,0.08509955803553264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,16384,0.09404799673292373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,512,0.027403554982609216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,4096,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,3584,0.009530666801664565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,5120,0.03721866673893399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,7168,0.02274666726589203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,3584,0.06796533531612821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,10240,0.013090666797426013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,12288,0.23004976908365884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,12288,0.0588702228334215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,1536,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,4096,0.6027759975857204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,8192,0.016103999482260812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,65536,0.13135022587246364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,8192,0.06263111035029094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,3584,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,2048,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,1536,0.011372444530328115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,51200,0.2192151149113973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,4096,0.05866222249137031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,128,0.005758222192525864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,16384,0.4288648764292399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,7168,1.1642853418986003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,3072,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,12288,0.044927110274632774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,51200,0.2896871036953396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,10240,0.43104712168375653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,768,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,1024,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,10240,0.07079733080334134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,4096,0.1945777734120687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,4096,0.010634666515721215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,3072,0.03042222062746684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,32,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,128,0.005396444350481033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,5120,0.009892444643709395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,256,0.03128177920977274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,12288,0.06239289045333862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,8192,0.023757333556811016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,512,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,3584,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,2560,0.038021332687801786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,5120,0.11707288689083523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,2048,0.00756177802880605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,2048,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,16384,0.09105066458384196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,1024,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,8192,0.016955556141005624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,1536,0.02865511178970337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,3072,0.007373332977294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,3584,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,1024,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,65536,1.3477119869656031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,256,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,5120,0.009583999713261923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,512,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,51200,0.1298560036553277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,3584,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,1024,0.011502222054534487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,2048,0.005151111218664381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,3072,0.05522133244408501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,32,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,1024,0.05044177836842007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,2560,0.03934489025010003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,12288,0.14753688706292048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,4096,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,51200,0.10758399963378906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,3584,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,65536,2.319634755452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,2560,0.020447111792034574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,2048,0.04675555560323927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,256,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,65536,0.09107466538747151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,65536,0.13668621910942927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,6144,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,7168,0.15198043982187906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,6144,0.07920622163348727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,256,0.005154666801293691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,256,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,256,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,16384,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,16384,0.3337253464592828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,3072,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,768,0.017143110434214275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,12288,0.027257778578334387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,6144,0.010273777776294285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,2048,0.005821333163314396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,65536,0.3076159954071045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,3584,0.052292442984051175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,1024,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,4096,0.12171822124057347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,1024,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,10240,0.020777778493033517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,16384,0.037823110818862915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,256,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,2560,0.027837332752015855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,5120,0.039686222871144615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,512,0.04894311229387919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,5120,1.154073821173774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,3072,0.2704373465643989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,4096,0.15705955028533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,3584,0.02973777718014187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,5120,0.09341777695549859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,10240,0.018193778064515855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,12288,0.030245333909988403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,12288,0.2740515602959527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,1536,0.028417779339684382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,16384,0.07589688565995958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,256,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,256,0.01460800071557363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,16384,0.01904800037542979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,3584,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,128,0.0700426697731018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,256,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,5120,0.01739644507567088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,51200,0.2715653313530816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,5120,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,32,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,2048,0.06580622328652276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,64,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,1536,0.06382044156392415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,51200,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,2560,0.01648533344268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,16384,0.3073306613498264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,768,0.08406577507654826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,3072,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,32,0.0081395556529363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,512,0.01811555524667104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,768,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,65536,0.08172177606158786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,5120,0.028115553988350764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,51200,0.010850666297806634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,5120,0.032009777095582746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,3072,0.048222220606274076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,6144,0.03526044554180569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,64,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,3072,0.1137946711646186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,4096,0.025719111164410908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,16384,0.0203804439968533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,10240,0.04334311021698845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,128,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,51200,0.050252444214291044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,2048,0.005102222164471944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,10240,0.13167822360992432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,2560,0.022991110881169636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,32,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,3584,0.007675555845101674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,1024,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,2048,0.028951999213960435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,8192,0.04385333259900411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,4096,0.020959110723601446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,65536,0.2730773289998372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,10240,0.12989777988857693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,7168,0.0653031137254503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,12288,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,6144,0.02069155540731218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,2048,0.01640977793269687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,3072,0.05858133236567179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,256,0.014761777387724983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,10240,0.013964444398880005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,4096,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,10240,0.00852711084816191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,6144,0.028347555134031508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,3584,0.017283555534150865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,5120,0.03229866756333245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,10240,0.44470845328436953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,8192,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,8192,0.05597244368659126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,3072,0.02120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,5120,0.012699555191728802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,12288,0.047525333033667676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,4096,0.21820799509684244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,768,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,512,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,65536,0.1399893363316854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,32,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,7168,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,32,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,32,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,7168,0.052616000175476074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,1536,0.01531377765867445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,64,0.006862222320503659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,128,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,256,0.008087111016114553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,8192,0.025037333369255066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,2560,0.02676800058947669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,2048,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,768,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,65536,1.1891618304782443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,3584,0.011358222199810876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,768,0.012039110892348819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,8192,0.01739199956258138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,32,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,3072,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,768,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,64,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,2048,0.01697688963678148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,12288,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,2560,0.057168887721167676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,16384,1.1189262602064345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,4096,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,7168,0.04731377628114489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,3584,0.013669333524174161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,7168,0.3864435619778103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,10240,0.5759475496080186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,1536,0.029650668303171795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,256,0.03665333324008518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,3072,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,1536,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,4096,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,6144,0.039974222580591835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,1024,0.007711110843552484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,16384,0.022700443863868713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,3584,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,1536,0.005075555708673266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,51200,0.21723200215233696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,5120,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,6144,0.12005689409044053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,1536,0.011343999869293638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,8192,0.23219911257425943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,2560,0.006463999963468975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,6144,0.033556444777382746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,7168,0.02332266668478648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,7168,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,10240,0.12411377165052627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,3584,0.033306668202082314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,6144,0.08111200067732069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,12288,0.025547555751270715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,7168,0.06441511048210992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,1536,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,51200,0.05597688754399618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,3072,2.3972621493869357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,2560,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,7168,0.02812444501452976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,51200,0.2651715543535021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,7168,0.09400266408920288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,512,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,16384,1.3960107167561848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,4096,0.034359110726250544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,512,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,128,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,768,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,51200,0.27019021246168345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,4096,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,16384,0.038210666841930814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,512,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,51200,0.967192014058431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,6144,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,1024,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,3584,0.06884533166885376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,768,0.006494222415818109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,4096,0.1090995536910163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,256,0.040377779139412776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,5120,0.04055022199948629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,64,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,512,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,3584,0.02966044346491496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,1024,0.007340444458855524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,8192,0.017975111802419026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,7168,0.1829235553741455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,6144,0.6666035652160645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,7168,0.013100444442696042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,256,0.019573332534896005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,12288,0.052983999252319336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,32,0.003952888978852166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,2560,0.08533422152201335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,12288,0.02714577813943227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,1536,0.009825777676370408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,3072,0.01035644445154402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,6144,0.02812444501452976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,16384,0.281858656141493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,8192,0.010976889067225985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,8192,0.15329689449734157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,6144,0.014231999715169271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,10240,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,2560,0.07292177942064074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,2560,0.012991111311647626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,1536,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,7168,0.05396977729267544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,10240,0.09379021988974677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,1024,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,2048,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,7168,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,16384,0.2702488899230957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,7168,0.05315022336112129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,3072,0.018567999203999836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,4096,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,10240,0.0351528889603085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,512,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,5120,0.02499733368555705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,2048,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,10240,0.16151821613311768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,3072,0.03825511203871833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,1024,0.058871997727288135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,32,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,16384,0.02050933407412635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,2560,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,51200,0.06377244657940336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,16384,0.015608888533380298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,6144,0.03864799936612447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,12288,0.022708444131745234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,2048,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,512,0.011711999773979187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,4096,0.02536533276240031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,12288,0.03010222315788269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,512,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,16384,0.06730310784445868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,5120,0.019128888845443726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,10240,0.01313777764638265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,1024,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,51200,3.8274375067816844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,256,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,2048,0.021689777572949726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,8192,0.020888888173633154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,128,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,768,0.009383111364311641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,8192,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,512,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,2560,0.04213688770929972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,3072,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,3072,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,3584,0.1277679999669393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,5120,0.010126222338941362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,2560,0.028407110108269587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,65536,0.4475884437561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,7168,0.09782133499781291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,6144,0.03901066713862949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,6144,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,6144,0.03596444593535529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,1024,0.02109066645304362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,256,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,5120,0.019280888968043856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,3584,0.019007111589113872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,1536,0.14882844024234348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,10240,0.13318667146894667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,512,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,51200,0.1281377739376492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,2560,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,7168,0.019514666663275827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,2560,0.007477333148320516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,64,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,768,0.017854221993022494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,8192,0.010998222563001843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,128,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,1536,0.012753778033786349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,768,0.01643200053109063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,16384,0.34702311621771914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,2048,0.02107288936773936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,32,0.04045955671204461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,768,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,2560,0.028304000695546467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,8192,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,1536,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,6144,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,10240,0.01923288901646932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,4096,0.07561955849329631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,51200,0.10905244615342881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,1024,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,512,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,512,0.013870221873124441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,7168,0.01772177716096242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,1024,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,3072,0.020413332515292697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,1024,0.3337093459235297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,65536,0.6445697678460015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,4096,0.015829333000712924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,12288,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,3584,0.03395466672049628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,2560,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,6144,0.21219555536905924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,8192,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,12288,0.12463111347622341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,3072,0.0247653325398763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,3584,0.013358222113715278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,3072,0.058661335044436984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,128,0.010054222411579555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,512,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,768,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,2560,0.015871110889646742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,2560,0.43027377128601074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,512,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,10240,0.11098844475216335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,8192,0.04381600022315979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,256,0.0019271110908852683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,3584,0.14221244388156468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,12288,0.5824044545491537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,1024,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,1024,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,5120,0.015841777125994366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,2048,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,8192,0.02589599953757392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,3072,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,7168,0.026233777403831482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,1024,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,3072,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,3072,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,2048,0.014045332868893942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,3072,0.02029333346419864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,4096,0.0096133334769143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,6144,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,7168,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,768,0.0032231110251612137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,128,0.06080000268088447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,51200,0.0946764416164822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,3584,0.06747821966807048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,16384,0.05077422327465481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,4096,0.031904889477623835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,16384,0.6801964441935221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,16384,0.48151376512315536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,1024,0.046966221597459584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,6144,0.00902844468752543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,512,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,1536,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,12288,0.018975110517607797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,512,0.00425866660144594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,1536,0.02853155467245314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,1024,0.024101333485709295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,6144,0.010141332944234213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,16384,0.11929778258005778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,12288,0.03191555539766947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,7168,0.1276951101091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,512,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,256,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,4096,0.03770133190684848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,1024,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,10240,0.04180444280306498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,51200,0.25052176581488717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,512,0.011422221859296164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,3584,0.014362666341993542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,5120,0.02514666650030348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,8192,0.03285600079430474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,3584,0.16314755545722112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,2560,0.02294755478700002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,16384,0.023594665858480666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,3072,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,6144,0.01605955594115787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,65536,3.662077162000868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,5120,0.05659200085533989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,8192,0.06333244509167142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,5120,0.028154667880800035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,16384,0.038217776351504855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,6144,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,4096,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,256,0.01819022165404426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,12288,0.031223111682467993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,3072,0.007723555796676212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,51200,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,512,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,3072,0.02569155560599433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,768,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,4096,0.04325422313478258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,5120,0.04132089018821716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,10240,0.017653332816229928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,10240,0.016350222958458793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,64,0.0018986666368113623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,7168,0.03620533479584588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,8192,0.08943821986516316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,3584,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,512,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,5120,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,7168,0.03733866744571262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,6144,0.012558221817016602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,512,0.013062222136391534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,128,0.016498666670587327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,4096,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,16384,0.04000000158945719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,5120,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,65536,12.664984809027779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,12288,0.05593511131074694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,8192,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,2048,0.008904000123341879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,10240,0.09399111403359307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,512,0.014497778481907315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,12288,0.01720622181892395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,16384,0.27545687887403697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,5120,0.026297777891159058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,512,0.028411553965674505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,3072,0.11154754956563313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,8192,0.3949066797892253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,16384,0.22317422760857475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,10240,0.014864888456132678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,8192,0.23760888311598036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,3072,0.037745777103636004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,51200,0.02642311155796051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,3584,0.018079110317760043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,3584,0.011920889218648275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,3584,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,4096,0.011644444531864591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,6144,0.03990488913324144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,3584,0.08986310826407538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,8192,0.14747467305925158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,2560,0.01938666734430525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,1024,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,8192,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,12288,0.013437333206335703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,5120,0.05379288726382785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,64,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,3584,0.01423288881778717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,51200,0.3395697805616591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,1024,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,64,0.0021342221233579847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,128,0.048766222265031606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,6144,0.008952889177534316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,12288,0.28234844737582737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,5120,0.08853599760267471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,5120,0.010644444160991244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,128,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,16384,0.038831111457612776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,256,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,16384,1.0251600477430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,1024,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,7168,0.050287110937966235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,5120,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,12288,0.021336000826623704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,6144,0.021776000658671062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,3072,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,7168,0.7828968895806206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,2048,0.212954666879442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,256,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,8192,0.10782488849427964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,2048,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,2560,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,51200,0.3597137663099501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,3584,0.024432889289326135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,10240,0.01699199941423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,1536,0.018544889158672757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,1024,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,3584,0.06418133444256253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,3584,0.1149111058976915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,768,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,2048,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,5120,0.009156444834338294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,51200,0.43108354674445254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,4096,0.023451555106374953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,8192,0.20069866710238984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,7168,0.11318933963775635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,5120,0.018945778409639995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,32,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,51200,0.0210844443904029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,512,0.011520889070298938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,51200,0.05111999975310432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,2560,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,6144,0.03983555568589105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,4096,0.0075697774688402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,768,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,1024,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,4096,0.08987644645902847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,5120,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,512,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,5120,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,128,0.019914666811625164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,16384,0.048730668094423085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,768,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,65536,0.2581431070963542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,4096,0.02536711096763611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,7168,0.02922044528855218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,3072,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,1536,0.29276710086398655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,8192,0.01826755536927117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,12288,0.04355644517474704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,5120,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,12288,0.011720889144473605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,64,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,5120,0.016573333077960543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,128,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,128,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,4096,0.18570933077070448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,4096,0.02722399930159251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,2560,0.28483555052015513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,7168,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,3584,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,4096,0.0165955556763543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,768,0.021219554874632094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,3584,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,8192,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,1024,0.020080000162124634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,2560,0.04214577873547872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,12288,0.057773331801096596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,5120,0.015439111325475903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,4096,0.02235377828280131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,3072,0.02055288851261139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,10240,0.022669333550665114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,16384,0.02089777754412757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,512,0.020291555258962844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,3584,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,1536,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,256,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,4096,0.010110222631030613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,65536,0.3288399908277723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,64,0.0019031111150979996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,51200,0.06151022513707479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,5120,0.04933155576388041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,65536,0.36202843983968097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,5120,0.040716442796919085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,128,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,3072,0.027786665492587622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,65536,2.466844346788194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,16384,0.020958221620983545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,2048,0.016240888171725802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,5120,0.16096088621351454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,65536,0.4301146666208903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,512,0.0025600000388092464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,2048,0.056292441156175405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,1024,0.023011555274327595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,6144,0.024116444918844435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,768,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,768,0.011992000043392181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,1536,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,3072,0.06755022207895915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,256,0.017533333765135873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,3584,0.035664889547559954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,6144,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,4096,0.03418933351834615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,6144,0.05164711011780632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,2560,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,51200,0.18261867099338105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,2048,0.012792000340090858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,12288,0.0359866652223799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,51200,0.7531529002719455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,7168,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,12288,0.021020443903075323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,6144,0.18721866607666016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,6144,0.07047110795974731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,3072,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,128,0.02843022346496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,1536,0.028938667641745672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,256,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,512,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,768,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,1536,0.0063991112013657885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,3072,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,16384,0.1502879990471734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,65536,0.08874311049779256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,768,0.07515733109580146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,768,0.013896889156765409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,768,0.035327110025617815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,64,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,7168,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,12288,0.10185422499974568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,4096,0.05535822113355001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,5120,0.9888071484035916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,10240,0.014541332920392355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,16384,0.024382222029897902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,64,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,8192,0.1118124458524916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,256,0.047573334640926786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,6144,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,2560,0.0166302224000295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,3072,0.013898667362001208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,1024,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,1536,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,12288,0.04745599958631727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,12288,0.0357528891828325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,12288,0.050310220983293324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,6144,0.03055466545952691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,4096,0.03682311044798957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,8192,0.11234577496846516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,2048,0.049774222903781466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,512,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,3072,0.02754488918516371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,1024,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,16384,0.09035466776953803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,512,0.01573422220018175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,65536,0.06667288806703356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,5120,0.019311111834314134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,16384,1.0102880265977647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,1536,0.027232888672086928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,512,0.0325528879960378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,4096,0.00927111092540953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,65536,0.15281421608395046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,7168,0.06552355819278292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,5120,0.04941866795221964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,10240,0.017190222938855488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,128,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,3072,0.029302222861184016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,8192,1.2143938276502821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,51200,0.054061333338419594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,16384,0.024013333850436743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,128,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,7168,0.010046222143703038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,2048,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,768,0.017966222431924608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,512,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,51200,0.021927111678653296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,32,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,16384,0.04542666673660278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,2560,0.023573334018389385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,64,0.0028719999310043124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,2560,0.048615111245049365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,16384,0.023342221975326538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,65536,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,12288,0.15455022123124865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,3584,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,12288,0.013783999615245394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,256,0.0219217770629459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,256,0.043004443248113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,128,0.06246933009889391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,2048,0.04170666800604927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,1536,0.011669333610269757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,3072,0.02237777743074629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,65536,0.7506275706821017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,7168,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,65536,0.35279999838935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,3584,0.03825066818131341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,3584,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,65536,0.29556088977389866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,1024,0.017768000562985737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,2560,0.009777777724795872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,128,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,16384,0.022985777921146814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,6144,0.011775111158688864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,16384,0.04509333438343472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,1536,0.02181688944498698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,512,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,3584,0.02435644467671712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,16384,0.24543910556369355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,128,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,12288,0.03741333219740126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,1024,0.728690677218967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,12288,2.3112320370144315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,7168,0.1552168925603231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,64,0.00776088900036282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,65536,0.07527377870347765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,3584,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,16384,0.05086844497256809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,5120,0.010967999696731567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,768,0.061390221118927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,10240,0.021534222695562575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,768,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,6144,0.056447108586629234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,10240,0.1622248888015747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,2560,0.03800355394681295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,51200,0.04271822174390157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,64,0.005728000154097875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,65536,0.058360000451405845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,512,0.013581333061059317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,768,0.0058462222417195635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,7168,0.011749332977665795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,10240,0.0247022228108512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,10240,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,3072,0.06339822212855022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,4096,0.009080000221729279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,2560,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,1536,0.004355555607212914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,32,0.007727111379305522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,4096,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,3584,0.037428445286220975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,512,0.018373333745532565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,1024,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,32,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,10240,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,1536,0.025055110454559326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,10240,0.16819377740224203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,2560,0.03921422362327576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,3072,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,7168,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,1024,0.053194665246539645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,32,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,1536,0.006145777801672618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,1536,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,256,0.01681777834892273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,10240,0.02823288904296027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,3072,0.677985774146186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,1024,0.5850488874647353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,16384,0.025596444805463154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,6144,0.11125066545274521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,128,0.013965333501497904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,8192,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,2560,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,1024,0.011033777561452655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,256,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,128,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,128,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,16384,0.09026844633950128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,1024,0.04981866810056898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,5120,0.03275733192761739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,7168,0.0398231115606096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,512,0.03661333190070258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,12288,0.10803111394246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,128,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,3072,0.048082666264639966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,7168,0.1869279940923055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,12288,0.08528711398442586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,16384,0.04088088870048523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,51200,0.4042186737060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,256,0.04601688848601448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,2560,0.017869333426157635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,7168,0.034853332572513156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,8192,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,32,0.0028880000528362063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,12288,0.2525048785739475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,3072,0.14056532912784153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,3584,0.09329155418607925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,1536,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,32,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,2048,0.013976888524161445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,7168,1.1117590798272026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,5120,0.024683554967244465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,16384,0.023966223001480103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,65536,9.791795518663195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,16384,0.03081688947147793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,65536,0.13998399840460882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,3584,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,4096,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,65536,0.12925777170393202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,3584,0.01866222255759769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,2048,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,10240,0.04936800069279141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,8192,0.07008088959587945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,1536,0.019951999187469482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,3072,0.020983111527231004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,65536,0.16258577505747476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,51200,0.28722045156690806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,2560,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,2560,0.0088755554623074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,12288,0.04369689027468363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,128,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,2048,0.0826444427172343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,3072,0.05057511064741346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,3072,0.031516446007622614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,3072,0.011417778001891242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,768,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,128,0.01072533345884747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,1536,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,10240,0.10527466403113471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,5120,0.14417333073086208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,1536,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,2560,0.04700977934731377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,768,0.015560888581805758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,16384,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,128,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,12288,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,3584,0.059804446167416044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,64,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,16384,0.26182489924960667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,4096,0.00981422182586458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,768,0.04417777723736233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,512,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,16384,0.11424356036716038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,1536,0.0232595553000768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,16384,0.06332177586025663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,512,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,10240,0.21586755911509195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,6144,0.06106488572226631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,512,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,16384,0.6176711188422309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,32,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,5120,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,512,0.010647999743620554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,7168,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,12288,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,4096,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,1536,0.008598222500748104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,10240,0.05636622508366903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,12288,0.09791288773218791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,2048,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,3584,0.0716844465997484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,8192,0.026022222306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,10240,0.0658515559302436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,1024,0.06468622552023993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,7168,0.05803822146521675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,12288,0.483301321665446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,1536,0.0038684445122877755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,8192,0.022329777479171753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,128,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,10240,0.4186275535159641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,256,0.03307200140423245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,8192,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,7168,0.013786666923099093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,4096,0.00811288919713762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,3584,0.03439199924468994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,16384,0.13827999432881674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,2560,0.01755199995305803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,1024,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,6144,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,1536,0.006697777658700943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,51200,0.4863084687127008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,16384,0.05723288986417982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,51200,0.0784462226761712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,6144,0.01990133358372582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,12288,0.02812622321976556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,64,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,16384,0.10317599773406982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,16384,1.304933336046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,10240,0.027079110344250996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,1536,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,10240,0.04743377698792351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,5120,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,512,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,4096,0.031982223192850746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,7168,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,16384,0.05340266558859083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,768,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,7168,0.04530044396718343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,6144,0.028251555230882432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,512,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,12288,2.380449718899197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,6144,0.03569511241383023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,10240,0.15922844409942627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,7168,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,3584,0.025935999221271936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,2560,0.011847111086050669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,512,0.006132444573773279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,768,0.02927910950448778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,6144,0.007737778127193451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,7168,0.6036862267388238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,3072,0.020993777447276648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,128,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,256,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,2048,0.006833777659469181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,768,0.03156977891921997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,768,0.02424977719783783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,1536,0.28603378931681317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,512,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,32,0.0019191110299693213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,768,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,512,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,512,0.003531555748648114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,3072,0.04718044400215149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,7168,0.19329599539438883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,12288,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,6144,0.14060710536109075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,10240,0.022261333134439256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,3072,0.03102933367093404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,65536,0.63991732067532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,3072,0.07833421892589994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,3584,0.01958666741847992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,256,0.021200888686709937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,2560,0.0827253328429328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,12288,0.04131644301944309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,4096,0.10655911101235284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,2048,0.017338666650984023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,2560,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,1536,0.008612444831265343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,7168,0.01936533384852939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,10240,0.019160888261265226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,128,0.006999111009968652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,768,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,1024,0.020833777056799993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,1536,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,512,0.0053688887920644544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,5120,0.09351911147435506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,64,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,128,0.00565244464410676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,128,0.0020026666008763844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,128,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,5120,0.027115555273161993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,8192,0.01571999986966451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,6144,0.010065777434243096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,7168,0.03087644444571601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,128,0.006309333360857434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,512,0.0035111109415690103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,10240,1.2273964352077908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,2048,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,32,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,7168,0.03420889046457078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,10240,0.012432888978057437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,5120,0.0902728901969062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,2048,0.0064693333374129395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,64,0.012119111087587146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,2048,0.02089244458410475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,3072,0.7228364414638944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,10240,0.1604142189025879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,1536,0.02236533330546485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,51200,22.379642062717014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,51200,0.16409244802263048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,3072,0.056403557459513344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,51200,0.12666577763027614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,3584,0.0140346669488483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,3072,1.0792053010728624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,51200,0.22185333569844565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,6144,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,7168,0.012936000194814471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,256,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,65536,0.5825600094265407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,10240,0.008708444734414419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,7168,0.026315554976463318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,2048,0.010793777803579966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,7168,0.15884355703989664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,51200,0.12195732858445908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,10240,0.021024000313546922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,768,0.05833422475390964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,3584,0.04269599914550781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,65536,0.10013777679867214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,5120,0.0807360013326009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,256,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,3072,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,2560,0.009284444153308868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,768,0.0047600000268883174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,1024,0.017969777186711628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,1536,0.033831109603246055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,3072,0.011587555209795633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,1024,0.007722666694058313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,10240,0.11484177907307942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,12288,0.1890151103337606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,1536,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,4096,0.24269866943359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,512,0.01591999994383918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,2560,0.00982844415638182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,768,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,1536,0.008108444511890411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,10240,0.15738756126827663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,16384,0.12001155482398139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,256,0.025458667013380263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,12288,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,512,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,5120,0.012235555383894177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,128,0.0063902222447925145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,2048,0.03326933251486884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,10240,0.09196444352467854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,32,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,16384,0.11486044194963242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,1536,0.0640373362435235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,8192,0.037929776642057635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,1024,0.018210666047202218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,32,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,12288,0.027078222897317674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,16384,0.023592889308929443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,1024,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,64,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,10240,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,10240,0.13298843966590032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,10240,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,256,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,5120,0.27475378248426646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,3072,0.00813066628244188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,128,0.00221422231859631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,8192,0.7892284393310547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,256,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,7168,0.015072888798183866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,1536,0.06964177555508085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,1024,0.02456355591615041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,2560,0.10591733455657959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,3072,0.014185777968830533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,32,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,768,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,128,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,1536,0.0192577772670322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,12288,0.040412445863087974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,6144,0.32575999365912545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,3584,0.01221777747074763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,12288,0.20780711703830293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,65536,56.76441446940104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,3072,0.012425777812798819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,768,0.028880887561374243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,512,0.02265600032276577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,8192,0.03159111075931125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,64,0.004106666478845808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,65536,0.13628089427947998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,1024,0.1050693326526218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,12288,0.02826933397187127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,7168,0.020621332857343886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,51200,0.05745066536797417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,10240,0.016206221448050607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,2048,0.010696888797812991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,7168,0.014594667487674289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,1024,0.08807288938098484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,512,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,16384,0.17295377784305146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,256,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,51200,0.05279199944602119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,8192,0.03405422303411696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,4096,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,3584,0.008352888955010308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,768,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,256,0.028361777464548748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,65536,0.8875404463873969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,4096,0.00814400033818351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,65536,0.6285146607293023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,16384,1.0066044065687392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,2048,0.060431996981302895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,3072,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,768,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,10240,0.10090311368306477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,8192,0.021129777034123737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,128,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,32,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,3584,0.01590577761332194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,10240,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,16384,0.05860533316930135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,3584,0.008114666574531132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,2048,0.03805333375930786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,12288,0.02130844526820713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,5120,0.03444000085194906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,2560,0.21042044957478842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,65536,0.22933689753214517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,65536,0.2553173436058892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,1536,0.015949333707491558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,1536,0.01696622206105126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,768,0.01828622155719333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,5120,0.027493332823117573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,65536,0.10271377695931329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,1536,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,8192,0.02400711178779602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,10240,0.07064000103208753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,1024,0.053582221269607544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,4096,0.14065689510769314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,256,0.029851555824279785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,768,0.019863110449579027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,3072,0.031193776263131037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,1536,0.01018666641579734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,3072,0.009480000370078618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,2048,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,5120,0.02160088883505927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,5120,0.022304889228608873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,2048,0.03409333361519708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,8192,0.15983822610643175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,7168,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,16384,0.36036178800794816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,51200,0.06347999970118205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,65536,0.2947031127081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,1024,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,256,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,4096,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,12288,0.020225778222084045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,4096,0.007833778030342525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,51200,7.4176686604817705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,3584,0.08464888731638591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,8192,0.17334932751125762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,12288,0.02946044339074029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,12288,0.1276408831278483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,128,0.009129777550697327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,32,0.004885333279768626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,7168,0.04144444399409824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,51200,0.09557066361109416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,768,0.020204444726308186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,12288,0.037676443656285606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,2048,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,16384,0.04459822177886963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,64,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,3072,0.03527999917666117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,5120,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,6144,0.017859554953045316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,1024,0.005990222096443176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,64,0.13538755310906306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,1536,0.035573333501815796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,128,0.01126488877667321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,51200,0.05374222331576877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,12288,0.12632977962493896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,51200,0.009642666412724389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,3584,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,6144,0.08942488829294841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,64,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,7168,0.059465779198540576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,768,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,5120,0.05546666516198052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,64,0.015100444356600443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,6144,0.062301331096225314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,5120,0.09170933564503987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,5120,0.01603822244538201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,8192,0.013423999978436364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,12288,0.013421333498424955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,2560,0.06383111079533894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,1536,0.03445955448680454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,128,0.029673778348498877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,3584,0.007537778053018782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,51200,0.15637955400678846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,65536,1.5910071267022028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,768,0.02432800001568264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,1536,0.016685333516862657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,3072,0.038155555725097656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,65536,0.016991110311614144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,65536,1.833187527126736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,768,0.02861422300338745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,32,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,128,0.0038488888078265717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,6144,0.06530400117238362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,2560,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,16384,0.028239998552534316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,4096,0.011199110911952125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,65536,0.10299111074871486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,2048,0.07707822322845459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,8192,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,12288,0.06201510959201389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,1024,0.05446222093370226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,65536,0.03371466530693902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,51200,0.05975377559661865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,2048,0.007240888973077138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,65536,0.0765742195977105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,1024,0.02737422287464142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,8192,0.06066489219665527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,65536,15.478510538736979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,1536,0.03199555476506551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,128,0.02268444498380025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,512,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,16384,0.2945680088467068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,768,0.01832444469134013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,128,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,4096,0.035318222310807966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,512,0.045511109961403735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,3584,0.012006222373909421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,1536,0.11605421702067058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,2560,0.04483377933502197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,128,0.011589333415031433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,6144,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,6144,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,8192,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,512,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,16384,0.02067733307679494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,51200,0.05552355448404948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,10240,4.150680118136935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,2048,0.010320888625250923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,65536,0.07605244716008504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,12288,0.30424711439344615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,10240,0.058337774541642934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,12288,0.022625777456495497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,3584,0.10057689083947076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,7168,0.011329777538776398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,3072,0.03173866536882188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,8192,0.03613511059019301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,768,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,768,0.04038577940728929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,8192,0.014376888672510782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,2560,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,5120,0.03906311260329352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,3584,0.01981511049800449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,2048,0.27031289206610787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,16384,0.031161778502994116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,51200,0.0683786670366923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,10240,0.026551110876931086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,256,0.002959111084540685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,16384,0.016144000821643405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,2560,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,2560,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,3584,0.00812266684240765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,3072,0.009726222190592024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,128,0.015236443943447538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,4096,0.03522133496072557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,6144,0.02198933396074507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,768,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,65536,0.07525777816772461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,2560,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,10240,0.24352622032165527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,512,0.012718222207493253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,7168,0.009876444107956356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,768,0.017509332961506315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,10240,0.030905776553683813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,5120,0.02752888864941067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,128,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,512,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,3584,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,4096,0.0620595547888014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,12288,0.018060444129837882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,3584,0.01164533363448249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,768,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,2048,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,12288,0.014282666974597506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,10240,0.036664889918433294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,12288,0.06438844733768039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,64,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,32,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,4096,0.1727395587497287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,128,0.02271733350223965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,10240,0.04003555576006571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,4096,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,10240,2.160775078667535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,10240,0.03676266802681817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,65536,0.18543467256757948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,16384,0.028419554233551025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,2560,0.04876266585456001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,128,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,4096,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,5120,0.18215733104281953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,1536,0.020032000210550096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,64,0.009757333331637913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,64,0.03359466791152954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,7168,0.03291644321547614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,51200,0.8705680105421277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,1536,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,32,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,1536,0.01091733326514562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,6144,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,3072,0.019362666540675692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,5120,0.022356443934970435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,512,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,1024,0.013381332986884646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,5120,0.010325333310498131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,16384,4.7145610385470915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,3584,0.02160800000031789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,8192,0.030264887544843886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,12288,0.3968044386969672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,6144,0.014745778507656522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,8192,0.007344000041484833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,512,0.012162666353914471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,1024,0.0047271110945277745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,10240,0.01738933391041226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,64,0.003269333392381668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,768,0.005668444352017508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,32,0.01404266721672482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,51200,0.04685066805945503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,1024,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,1536,0.02602844436963399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,5120,0.5764142142401801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,64,0.02671555512481266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,2048,0.13277955849965414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,7168,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,8192,0.03660000032848782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,1024,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,128,0.010687111152542962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,51200,0.16330933570861816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,3072,0.012268444730175866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,51200,0.06008266740375095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,768,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,2048,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,3584,0.6473760075039333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,5120,0.0259253333012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,5120,0.028062221076753404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,3072,0.008380444513426887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,8192,0.02693422304259406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,128,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,128,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,1536,0.011155555645624796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,3072,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,2560,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,10240,0.027664888236257765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,6144,0.11206577883826362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,16384,0.09448799822065566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,12288,0.10332977771759033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,256,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,10240,0.036981334288915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,512,0.0034871109657817413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,512,0.020618667205174763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,512,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,1024,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,5120,0.07430933581458198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,2560,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,51200,0.17241244845920137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,256,0.0018097776919603348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,5120,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,128,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,1024,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,128,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,7168,0.06176355812284681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,5120,0.027701333165168762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,1024,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,256,0.02953688965903388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,1536,0.3057742118835449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,1024,0.05199911197026571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,8192,0.04760711060629951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,768,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,5120,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,3584,0.010611555642551847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,65536,0.23804267247517905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,2560,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,128,0.03318399853176541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,65536,0.2540480030907525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,51200,0.058166219128502734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,6144,0.13099288940429688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,1024,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,8192,5.105061425103082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,6144,0.017480888300471835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,10240,0.05013511247105069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,8192,0.01424888935354021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,51200,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,128,0.002238222294383579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,5120,0.2900515662299262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,1536,0.06521955463621351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,7168,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,7168,0.05983733468585544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,3584,0.16752799352010092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,256,0.025585777229732935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,2048,0.0546231104267968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,65536,1.6857208675808375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,16384,5.702597300211589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,51200,0.19075022803412545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,2560,0.008457777400811514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,128,0.012712000144852532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,6144,0.009937778115272522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,64,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,2560,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,768,0.006764444626039929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,65536,0.014944000376595391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,6144,0.02212000058756934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,4096,0.019003555178642273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,3072,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,16384,0.0410497784614563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,2560,0.022896000080638464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,64,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,7168,0.005160000175237656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,51200,0.048935999472936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,2560,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,65536,0.6626986927456325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,4096,0.05112089051140679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,5120,0.04211022125350105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,128,0.012642666697502136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,16384,0.05212710963355171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,7168,0.01984177695380317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,2560,0.16610933674706355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,6144,0.10259999831517537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,12288,0.020976000361972388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,7168,0.020396444532606337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,5120,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,5120,0.03276000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,8192,0.15393866433037653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,5120,0.014689778288205465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,3072,0.03656355539957682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,7168,0.012448888685968188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,2048,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,768,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,5120,0.04839555422465006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,8192,0.03910222318437364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,2560,0.07902666595247057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,5120,0.020263999700546265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,5120,0.06586044364505343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,32,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,10240,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,768,0.07660622066921659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,2560,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,32,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,2560,0.0883973307079739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,6144,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,6144,0.02681688964366913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,2560,0.004847110973464118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,51200,0.28959377606709796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,5120,0.0292195545302497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,3072,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,32,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,3072,0.01944533321592543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,128,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,32,0.03147911032040914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,1536,0.03410933415095011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,6144,0.01790488925245073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,64,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,3584,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,10240,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,2560,0.03692888882425096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,16384,0.1098160015212165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,32,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,3584,0.19108266300625273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,256,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,12288,0.7143840259975858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,1536,0.24106576707628039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,8192,0.026486222942670185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,51200,0.11042933331595527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,10240,0.0510755545563168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,256,0.023826667004161414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,6144,0.01760444376203749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,5120,0.01752088963985443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,16384,0.03187377916442023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,10240,0.005636444522274866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,3072,0.011295110815101199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,7168,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,256,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,65536,0.021743110484547083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,10240,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,16384,0.019321777754359774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,32,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,512,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,512,0.007532444265153673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,3072,0.015843555331230164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,16384,0.07565155294206408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,7168,0.02202755543920729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,768,0.20799377229478624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,1024,0.023013333479563396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,6144,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,3072,0.03286577926741706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,768,0.12801867061191136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,65536,0.05373422304789225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,4096,0.022932445009549458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,512,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,51200,0.10793333583407932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,3584,0.18771821922726104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,3072,0.043176889419555664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,4096,0.01371022231049008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,3072,0.07788444227642484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,128,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,64,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,2560,0.01367555558681488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,51200,1.8954808976915147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,16384,0.10574933555391099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,12288,0.4139404561784532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,128,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,1024,0.09247822231716579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,10240,0.025102222959200542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,5120,1.3096133338080513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,16384,0.020569778150982328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,4096,0.14738578266567653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,1024,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,2560,0.02991644541422526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,128,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,768,0.005985777825117111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,65536,0.08066844277911715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,51200,0.04903822143872579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,768,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,6144,0.05228177706400553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,128,0.02200444373819563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,6144,0.009128000173303816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,1536,0.02330933345688714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,1536,0.014280888769361707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,1536,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,10240,0.026533333791626826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,4096,0.036291556225882635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,65536,9.282393561469183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,1024,0.007061333292060428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,8192,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,6144,0.6759555604722766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,16384,0.031153778235117596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,65536,1.6421377393934462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,1536,0.15827288892534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,256,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,10240,0.0952684415711297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,65536,0.11954932742648655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,1536,0.006456888798210356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,2048,0.009856888817416297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,16384,0.019500444332758587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,1536,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,6144,0.019253333409627277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,6144,0.12289955880906846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,512,0.02660711109638214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,2048,0.018287110659811232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,768,0.011135111252466837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,1024,0.0617555578549703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,16384,0.01854044364558326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,4096,0.010608000059922537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,2560,0.010606221854686737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,1536,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,1536,0.033470223347345986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,6144,0.023356444305843774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,2048,0.30432621637980145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,6144,0.01980444457795885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,10240,0.024852444728215534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,1024,0.003752888904677497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,16384,0.04464000132348803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,12288,0.019072888625992667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,768,0.011424889167149862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,4096,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,1024,0.0064560001095136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,6144,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,16384,0.0225608895222346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,6144,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,3584,0.03919466667705112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,64,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,6144,0.011132444772455426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,3072,0.007808888951937358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,7168,1.987381405300564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,10240,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,768,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,3072,0.020563556088341605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,16384,0.15515022807651097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,3584,0.02369955513212416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,3584,0.02238311039076911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,65536,0.5215804311964247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,8192,0.011438222395049201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,65536,0.010824888944625854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,128,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,256,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,12288,0.138319108221266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,5120,0.11112977398766412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,8192,0.02528000043498145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,3072,0.011136888629860349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,3072,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,16384,0.055629332860310875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,7168,0.014166222678290473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,10240,0.019926221834288705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,2560,0.004953777624501122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,768,0.011952888634469775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,12288,0.12367910808987087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,12288,0.1341057750913832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,768,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,8192,0.11725244257185195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,3072,0.017589333984586928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,8192,0.019374221563339233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,8192,0.019138667318556044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,10240,1.7010889053344727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,7168,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,2048,0.029320889049106177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,12288,0.5831875801086426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,6144,0.11739467249976264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,256,0.009380444056457942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,2560,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,8192,0.010748444332016839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,768,0.016718222035302054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,5120,0.0627688897980584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,10240,0.07038133011923896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,768,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,8192,0.04782488942146301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,7168,0.02513422237502204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,64,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,51200,1.1165190802680123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,6144,0.072781331009335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,3072,0.028427554501427546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,8192,0.02199733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,6144,0.03411555621359084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,65536,0.08182399802737765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,768,0.020472000042597454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,32,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,3072,0.029138667715920344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,2048,0.5381795565287272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,3072,0.03005511230892605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,128,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,64,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,768,0.01771377854877048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,51200,0.11342667208777533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,6144,0.02977422210905287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,65536,0.16331556108262804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,16384,0.02163999941613939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,16384,0.03752355443106757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,65536,0.40351022614373105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,1536,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,128,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,32,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,512,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,51200,1.3677164713541667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,4096,0.11373066902160645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,8192,0.1375973357094659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,64,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,2048,0.052467554807662964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,65536,0.06457689073350695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,16384,0.14451200432247585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,7168,0.028078221612506445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,1024,0.016472889317406546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,32,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,7168,0.08489600155088638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,65536,0.05876977576149834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,65536,0.44903379016452366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,7168,0.20969155099656847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,256,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,7168,0.011585777832402123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,256,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,256,0.02273511058754391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,1536,0.022786666949590046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,2048,0.03742311067051358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,5120,0.02606933315594991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,256,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,65536,0.21737866931491426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,51200,0.00665155549844106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,65536,0.1722773313522339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,2560,0.11681955390506321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,12288,0.014230221509933472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,12288,0.010760000182522668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,128,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,256,0.0057795556883017225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,128,0.018395556343926322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,12288,0.013770666387346057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,1536,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,10240,0.03977333174811469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,768,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,256,0.017161778277821012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,512,0.01874755488501655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,65536,0.12000355455610488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,32,0.055199109845691256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,5120,0.46482133865356445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,768,0.2590995629628499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,51200,0.10735911130905151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,10240,0.6855440139770508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,7168,0.02461599972512987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,4096,0.3789964516957601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,16384,0.042633778519100614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,16384,0.08916711144977146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,8192,0.04583466715282864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,51200,1.0183288786146376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,12288,0.015792888071801927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,1024,0.007700444095664554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,1536,0.026176000634829204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,2560,0.1360293361875746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,65536,0.5440390904744467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,4096,0.11992711491054958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,1024,0.013264888690577613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,65536,0.5477822091844347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,768,0.02933599882655674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,4096,0.17289066314697266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,16384,0.06552444563971625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,32,0.022104889154434204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,8192,0.2442248927222358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,4096,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,1536,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,8192,0.03224710954560174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,8192,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,16384,0.07708888583713107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,128,0.013737777868906656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,768,0.018908444378111098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,8192,0.03346933258904351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,5120,0.05637866920895047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,65536,2.0418177710639105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,51200,0.15683555603027344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,64,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,2048,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,3584,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,2560,0.014333332578341166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,8192,0.14135021633572048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,256,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,51200,0.15206132994757757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,2048,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,6144,1.1660435994466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,3584,0.026383999321195815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,7168,0.02843466732237074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,128,0.020232889387342665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,5120,0.0105262224872907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,64,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,10240,0.7508426772223579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,3584,0.027170666389995154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,8192,0.03775733378198411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,12288,0.6273182233174642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,8192,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,2560,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,768,0.012410666379663678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,7168,0.015213333898120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,10240,0.048357334401872426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,768,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,8192,0.012597333225939008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,2560,0.009552000297440423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,65536,0.08750222126642863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,51200,1.0213866763644748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,4096,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,12288,0.03693600164519416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,12288,0.02907466557290819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,51200,2.365950266520182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,65536,0.15026044845581055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,10240,0.2812160121070014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,256,0.017503110898865592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,5120,0.2329253355662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,8192,0.09558577669991387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,5120,0.03502488798565335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,6144,0.05378489030732048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,3072,0.007924444145626491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,4096,0.031552000178231135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,128,0.015300444430775113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,3072,0.04990577697753906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,6144,0.007795555724038019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,3584,0.46246666378445095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,10240,0.022471111681726243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,1536,0.030416889323128596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,512,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,1536,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,16384,5.494204627143012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,1536,0.012793777717484368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,3584,0.008243555823961893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,1024,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,2048,0.0203795548942354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,16384,0.04686577783690559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,2560,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,16384,0.05561511384116279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,65536,1.0420986811319988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,8192,0.011001778145631155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,8192,0.036700444089041814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,10240,0.020615110794703167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,51200,0.058033777607811816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,6144,0.05170933405558268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,3584,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,2048,0.2535591125488281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,2560,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,512,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,12288,0.022230222821235657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,768,0.026777777406904433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,768,0.014875556031862894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,51200,0.6830675337049695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,128,0.10991822348700629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,1024,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,32,0.0019582222319311565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,2048,0.020980444219377305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,32,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,3072,0.014470222923490735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,10240,0.29725154240926105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,2048,0.01698933376206292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,12288,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,3584,0.03053599927160475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,16384,0.318320009443495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,512,0.015558222929636637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,51200,0.15369778209262425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,768,0.00998577806684706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,6144,0.029151999288135107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,2560,0.010438222024175854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,4096,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,6144,0.06173688835567898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,6144,0.01920000049802992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,51200,0.103657775455051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,3072,0.02642222245534261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,5120,0.012823111481136747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,8192,0.037223110596338906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,7168,0.03661600086424086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,10240,0.2031573322084215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,1536,0.024140444066789415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,5120,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,8192,0.035671999057133995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,10240,0.012381333443853589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,65536,0.1482737726635403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,1024,0.07413333654403687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,65536,0.0928586655192905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,4096,0.015297777122921415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,12288,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,768,0.01053955571519004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,3584,0.025957332717047796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,16384,0.04349866509437561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,512,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,7168,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,128,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,8192,0.04126400086614821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,2560,0.007221333682537079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,5120,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,2560,0.024726221958796184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,7168,0.208898663520813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,7168,0.08505955669615005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,64,0.11251289314693874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,3584,0.004210666649871402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,65536,0.17877066135406494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,256,0.00960622231165568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,768,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,1024,0.08222311072879367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,128,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,2560,0.07674133115344577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,1536,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,1536,0.07738044526841906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,10240,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,1536,0.02019733356104957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,16384,0.1130488846037123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,8192,0.04519911275969612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,512,0.4221199883355035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,16384,0.02235377828280131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,1536,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,10240,0.15966667069329157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,3072,0.06455377737681071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,51200,0.06450755728615655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,7168,0.02459466622935401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,12288,0.2524204519059923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,32,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,8192,0.016028443972269695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,65536,1.2154169082641602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,512,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,256,0.006067555397748947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,7168,0.7003902329338921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,512,0.051114665137396924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,1024,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,7168,0.07019111182954577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,128,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,2048,0.09938044680489434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,10240,0.1405644416809082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,2560,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,128,0.07288622193866305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,4096,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,1024,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,3584,0.05401688814163208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,128,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,16384,0.04477511180771721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,32,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,4096,0.0387386679649353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,1536,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,16384,0.02084622283776601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,3072,0.060621334446801074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,1024,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,65536,0.17701421843634713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,1024,0.00647288892004225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,65536,0.2894017696380615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,5120,0.04917600088649326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,1024,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,6144,0.020976000361972388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,2048,0.06537510951360066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,7168,0.12270932727389866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,7168,0.030415998564826116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,4096,0.23876621988084581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,16384,0.13750133249494764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,3584,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,1024,0.31101687749226886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,6144,0.05667644739151001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,128,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,12288,0.04637155599064297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,4096,0.007920888562997183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,8192,0.10019022226333618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,1024,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,4096,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,1024,0.006871999965773688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,256,0.027464000715149775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,128,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,128,0.04917777909172905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,51200,0.6088142395019531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,1536,0.021951999929216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,51200,0.8527893490261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,5120,0.04804089003139072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,4096,0.16593510574764675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,16384,0.02892000145382351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,65536,1.1509058210584853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,51200,0.05230666531456841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,1536,0.011339555184046427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,6144,0.010244444840484196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,4096,0.01834222177664439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,3584,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,12288,1.6558817757500544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,8192,0.017960889471901786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,51200,0.5440355406867133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,3584,0.06544266806708442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,1536,0.18049689133961996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,16384,0.36698399649726016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,3584,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,2048,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,768,0.007874666816658443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,6144,0.01611111064751943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,16384,0.0373271107673645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,256,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,8192,0.6837004555596246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,16384,0.12633421685960558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,768,0.006693333387374878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,12288,0.05173066589567396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,1536,0.01928622192806668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,51200,0.05631999837027656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,512,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,51200,0.0689857800801595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,2048,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,65536,0.09774577617645264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,5120,0.008549333446555668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,5120,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,3072,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,7168,0.03370133373472426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,5120,0.1256613334019979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,12288,0.0173999998304579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,8192,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,1024,0.019360888335439894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,512,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,768,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,1536,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,6144,0.10421422455045913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,7168,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,7168,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,51200,0.05943999687830607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,256,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,2560,0.043226665920681424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,64,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,2048,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,3072,0.0450764430893792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,65536,0.07643822166654798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,4096,0.015595555305480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,65536,1.9113538530137804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,256,0.0024871110088295406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,1024,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,6144,0.010079111489984725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,512,0.007199111084143321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,2560,0.012178666889667511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,3072,0.04818933208783468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,32,0.031165334913465712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,7168,0.014041778114106921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,12288,0.03054311209254795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,3584,0.012192000117566852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,4096,0.4477226469251845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,5120,0.060144888030158154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,10240,0.1225493351618449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,5120,0.01463822192615933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,64,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,12288,0.261007997724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,768,0.04768266611629062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,65536,0.015572445260153877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,3584,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,8192,0.023054222265879314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,51200,1.523580445183648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,3072,0.03092977735731337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,10240,0.03393866618474325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,128,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,65536,0.2458497683207194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,256,0.01055555542310079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,1536,0.029269334342744615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,2048,0.004551110996140374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,5120,0.2493893305460612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,10240,0.031172444423039753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,51200,7.914282904730903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,512,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,8192,0.2092195616828071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,12288,4.162476433648004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,16384,0.2214062213897705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,6144,0.31784354315863717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,3584,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,8192,0.01351999988158544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,256,0.037829332881503634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,16384,0.13361510965559217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,16384,0.3315555519527859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,256,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,4096,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,5120,0.05057155423694187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,3584,0.27068977885776097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,3584,0.008022222254011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,256,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,8192,0.07234311103820801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,8192,0.14314844873216417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,256,0.01613066593805949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,128,0.021928000781271193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,1536,0.008647110727098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,65536,0.12512355380588108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,3584,0.007739555504586961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,256,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,3072,0.10361511177486843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,4096,0.03977244430118137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,8192,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,5120,0.05747199720806546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,4096,0.015619556109110514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,4096,0.023560000790490046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,65536,0.22545154889424643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,6144,0.023665777511066858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,768,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,3072,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,10240,0.02737955583466424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,1024,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,2560,0.005688000056478713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,3072,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,3072,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,51200,0.024018666810459558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,6144,0.011672889192899069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,8192,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,8192,0.022375111778577168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,6144,0.024528000089857314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,3072,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,256,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,32,0.024136889312002394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,1024,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,32,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,256,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,5120,0.062036441432105176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,512,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,65536,0.31951289706759983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,1536,0.009883555273214975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,128,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,512,0.038366221719317965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,6144,0.17752799722883436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,256,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,256,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,3584,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,1536,0.0034755555291970572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,32,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,51200,0.7062453163994683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,51200,3.095997280544705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,8192,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,4096,0.026113778352737427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,128,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,2048,0.01256711118751102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,3584,0.02160266704029507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,256,0.14433244864145914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,16384,0.08372888962427776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,1536,0.006890666567616993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,65536,0.05950755543178982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,16384,0.03302399979697333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,512,0.032122668292787336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,256,0.014895111322402954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,7168,0.02090577781200409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,2048,0.05487289031346639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,10240,1.4643377727932398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,10240,0.4657146665785048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,2048,0.011061333119869232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,8192,0.1023155583275689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,512,0.004528000122971005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,1536,0.028552889823913574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,2048,0.07451911105049981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,3072,0.09422577752007379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,768,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,768,0.029291556941138372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,8192,0.014576888746685453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,8192,0.025050666597154405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,256,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,12288,0.0199644449684355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,1536,0.055656890074412026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,2560,0.026716444227430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,51200,0.05163733495606316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,4096,0.02096533278624217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,51200,1.1182017856174045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,768,0.03454577922821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,8192,0.015858666764365304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,16384,0.2613324324289958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,5120,0.016187555260128446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,64,0.026289777623282537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,7168,0.03308533297644721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,1536,0.027455111344655354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,12288,0.013926222920417786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,51200,0.05410844418737623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,10240,0.24454845322502983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,32,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,1024,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,128,0.014070222775141398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,256,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,512,0.01882844501071506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,10240,0.010948444406191507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,10240,0.027544000082545813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,51200,0.05472977624999153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,1536,0.0035582222044467926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,10240,0.012705778082211813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,10240,0.07748622364468045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,1024,0.01108444482088089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,32,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,768,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,512,0.004204444587230682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,3584,0.006544888847404056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,16384,0.020279111133681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,7168,0.20772266387939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,10240,0.04650577902793884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,256,0.006121777825885349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,3072,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,1024,0.039339555634392634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,8192,0.02295911146534814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,3072,0.009140444298585257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,5120,0.017889777819315594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,256,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,512,0.002604444407754474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,65536,1.7608586417304144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,128,0.008590222232871586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,128,0.008263111114501953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,5120,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,51200,0.05520088805092705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,6144,0.024361777636739943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,16384,0.022671111755900916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,10240,0.3060079945458306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,512,0.007694222033023834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,128,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,2560,0.06711910830603705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,8192,0.056016001436445445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,768,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,10240,0.018406222263971966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,16384,0.02574488851759169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,64,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,3584,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,3072,0.004552000098758274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,8192,0.014375110467274984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,3584,0.008795555267069075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,1024,0.03271377748913235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,7168,0.03712177938885159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,2048,0.023818666736284893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,2560,0.018139556050300598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,3072,0.010155555274751451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,512,0.018966222802797954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,2560,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,3584,0.016544888416926067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,10240,0.027172444595230952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,2048,0.09602222177717422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,4096,0.0625928905275133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,12288,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,8192,0.010393777655230628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,5120,0.7000720236036512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,51200,0.12032266457875569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,51200,12.298016018337675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,2560,0.10239022307925755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,7168,0.0630355543560452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,1536,0.03696799940533108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,8192,0.13911733362409803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,1024,0.025063999825053748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,32,0.011975111232863532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,2048,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,3584,0.03806666533152262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,2048,0.014311111635631986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,3584,0.007236444287829929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,6144,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,16384,0.046522667010625206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,2048,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,65536,0.04745955599678887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,256,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,8192,0.012777778009573618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,6144,0.01221155540810691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,3584,0.04657511247528923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,51200,0.10917599995930989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,1536,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,64,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,16384,0.025621333056026038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,2048,0.004287111262480418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,65536,0.16201066970825195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,6144,0.0974000030093723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,12288,0.08393422100279067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,32,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,8192,0.029821332958009508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,1536,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,1536,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,8192,0.010681777364677854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,32,0.07199911276499431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,8192,0.036350223753187395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,32,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,256,0.08113244507047865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,256,0.11850221951802571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,2560,0.07438577546013726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,2048,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,2560,0.14355733659532335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,65536,0.11172355545891656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,51200,0.10877422491709392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,51200,1.5163244671291773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,51200,0.0703351100285848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,10240,0.06826044453514947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,51200,2.118183135986328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,7168,0.013069333301650153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,16384,0.6791520118713379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,10240,0.012697777814335294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,12288,0.30905066596137154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,4096,0.017509332961506315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,128,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,10240,0.016920000314712524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,65536,0.05451022254096138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,32,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,65536,0.2085262272093031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,2048,0.007185777856243982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,10240,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,2048,0.04492977923817105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,16384,0.373016013039483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,65536,0.09247022204928929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,128,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,10240,0.05279644330342611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,2048,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,10240,0.01313777764638265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,3072,0.02126311096880171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,8192,0.013830222189426422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,4096,0.07007377677493624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,16384,0.019536889261669584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,1536,0.02828977836502923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,6144,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,65536,10.951911926269531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,12288,0.020229332976871066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,512,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,1536,0.012381333443853589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,51200,0.7793502277798123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,7168,0.24175643920898438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,16384,0.04944622185495165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,16384,0.10952177974912856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,6144,0.037646220790015325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,5120,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,7168,0.013908444179428948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,2048,0.03145688772201538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,3584,0.02667377723587884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,2048,0.014965333872371249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,256,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,768,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,51200,0.0985413326157464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,2048,0.009291555318567488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,128,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,65536,0.7411706712510852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,3072,0.01734044485621982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,65536,0.127710223197937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,12288,0.2776346736484104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,12288,0.024666666984558105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,8192,0.0411102208826277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,4096,0.09098755651050144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,768,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,768,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,64,0.06962755653593275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,2560,0.044585777653588184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,12288,0.15233777629004583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,10240,0.028138667345046997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,3584,0.016823111308945548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,1024,0.3814106782277425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,2048,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,256,0.013287111288971372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,2560,0.05663022067811754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,128,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,64,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,2048,0.020989333589871723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,32,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,65536,0.17062221633063424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,3072,0.06587911314434476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,6144,0.20287644863128662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,7168,0.343356450398763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,32,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,1536,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,10240,0.02366755571630266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,2560,0.02773155603143904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,4096,0.06899822420544095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,5120,0.05714133050706652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,5120,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,6144,0.030138668086793687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,3584,1.3080168830023873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,8192,0.01685422162214915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,5120,0.058005332946777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,512,0.013507555756303998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,5120,0.027031110392676458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,1024,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,65536,0.14634754922654894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,512,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,1536,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,512,0.006047111004590988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,3072,0.01886488828394148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,16384,0.017715555098321702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,6144,0.027260444230503503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,8192,0.13186844189961752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,10240,0.18158400058746338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,10240,0.05212177832921346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,256,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,3072,0.028047998746236164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,3584,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,64,0.0369795560836792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,2560,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,2048,0.009090666969617208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,32,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,65536,2.3419723510742188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,8192,0.02357599967055851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,64,0.10758311218685573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,7168,0.06663466824425592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,16384,0.17746755811903211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,64,0.0017946666727463405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,3072,0.014941333068741692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,256,0.020378667447302077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,3584,0.05196266704135471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,128,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,3072,0.014294221997261047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,1024,0.10568710830476548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,10240,0.19470132721794975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,5120,0.017993777990341187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,2560,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,7168,0.03882755504714118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,3584,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,32,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,256,0.09179733196894328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,3584,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,2048,0.010070222119490305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,12288,0.016876444220542908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,256,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,16384,0.513338671790229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,768,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,51200,1.4937733544243708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,2560,0.07615022526846991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,12288,0.017400888933075797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,51200,0.05694755580690172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,768,0.005136888888147143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,65536,0.06337510877185397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,768,0.030786666605207656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,512,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,3072,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,512,0.027831110689375136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,1024,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,128,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,5120,0.01186222251918581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,2048,0.021512000097168818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,10240,0.3797813256581624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,768,0.18203288979000518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,1024,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,2048,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,128,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,128,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,2560,0.15860177410973444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,1536,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,32,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,768,0.056773331430223256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,5120,0.03044711218939887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,65536,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,16384,0.030204445123672485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,3584,0.007727111379305522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,2048,0.008628444539176093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,16384,0.021203555994563635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,65536,0.015979554918077257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,768,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,2560,0.06595466534296672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,6144,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,512,0.013615111509958902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,5120,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,7168,0.012820444173283048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,32,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,3584,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,5120,0.033053331904941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,1024,0.019253333409627277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,3072,0.6659430927700466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,7168,0.2849840058220757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,512,0.003966222206751506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,2048,0.07208622164196439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,512,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,128,0.006542222367392645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,3072,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,1536,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,768,0.012412444584899478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,2560,0.022014222211307947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,64,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,3072,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,2560,0.03388799892531501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,2560,0.01220266686545478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,1024,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,8192,0.042928887738121875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,51200,1.1925795873006184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,8192,0.09557066361109416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,8192,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,256,0.08949777815077041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,32,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,2560,0.00886222223440806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,7168,0.06464355521731906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,256,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,65536,0.24061155319213867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,3072,0.10659555594126384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,10240,0.00776622196038564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,768,0.020592888196309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,8192,0.014300444059901767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,3072,0.00958577791849772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,128,0.007178666690985362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,3584,0.06611288918389215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,4096,0.04074044360054864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,1024,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,768,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,2048,0.018749333090252347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,16384,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,128,0.17052177588144937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,1536,0.008098666866620382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,1024,0.03009777930047777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,51200,0.047853334082497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,256,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,128,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,512,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,5120,0.011537777880827585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,3072,0.014364444547229342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,768,0.01220266686545478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,128,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,64,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,7168,0.011039111349317761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,5120,0.01714666684468587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,2560,0.009285333255926767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,768,0.015413333972295126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,6144,0.08957244290245904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,51200,0.05689511034223768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,3072,0.03086044390996297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,128,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,7168,0.009748444788985783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,2560,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,2048,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,256,0.010885333021481832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,256,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,12288,0.13805777496761745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,5120,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,6144,0.10105599959691365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,10240,1.544902271694607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,256,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,10240,2.4444569481743708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,16384,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,8192,0.01408088869518704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,8192,0.023424888650576275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,10240,0.05091822147369385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,4096,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,2560,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,12288,0.18636000156402588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,256,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,6144,0.009098666409651438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,256,0.06163733535342746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,3072,0.10290488931867812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,10240,0.03181333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,2560,0.019077334139082167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,16384,0.05792088641060723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,51200,0.04551555713017782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,10240,0.08753066592746311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,1024,0.012340444657537671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,10240,0.02477688921822442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,1536,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,51200,0.08392888969845241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,51200,0.03344622254371643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,2048,0.10273510879940456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,256,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,1024,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,2048,0.04099911120202806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,64,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,12288,0.054995556672414146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,2048,0.015273777974976433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,256,0.00927111092540953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,1536,0.260279999838935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,1536,0.0847804413901435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,65536,0.12327199512057835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,1024,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,768,0.013145777914259167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,8192,0.02952622373898824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,12288,0.023639111055268183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,4096,0.010122666756312052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,4096,0.015393777026070489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,16384,0.024440889557202656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,10240,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,6144,0.12578311231401232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,6144,0.009915555516878763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,51200,0.9187609354654948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,51200,0.07128177748786078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,512,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,10240,0.013325333595275879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,32,0.028877334462271795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,8192,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,2048,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,3584,0.019485332899623446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,16384,0.9451422161526151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,512,0.013959111438857185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,2048,0.004974222017659081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,12288,0.05344533258014255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,65536,0.06514755884806316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,6144,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,2560,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,2560,0.02711733347839779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,5120,0.18172532982296416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,2560,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,10240,0.010099555883142684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,7168,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,5120,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,128,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,6144,0.018594667315483093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,5120,0.7594461970859103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,512,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,10240,0.12096444765726726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,128,0.0124195557501581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,8192,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,3072,0.005340444544951121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,12288,0.013800889253616333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,7168,0.038373334540261164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,1024,0.06442844205432467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,5120,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,10240,0.2766995429992676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,256,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,32,0.011538666983445486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,12288,0.08220089144176908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,6144,0.034607112407684326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,65536,0.24988977114359537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,4096,0.11298844549391006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,64,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,65536,0.2661840120951335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,5120,1.642004436916775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,768,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,256,0.011146666275130378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,256,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,4096,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,5120,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,1024,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,2560,0.02973066767056783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,1024,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,3072,0.026966222458415564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,12288,0.10896355575985378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,512,0.046296000480651855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,65536,0.13663910494910345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,4096,0.03481600019666884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,64,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,512,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,32,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,5120,0.02366222275627984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,8192,0.13136533896128336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,3584,0.025432889660199482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,65536,0.8653342458936902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,16384,0.24026311768425834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,1536,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,1024,0.006041777630647023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,6144,0.04403733213742574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,3584,0.028543998797734577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,5120,0.1194017728169759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,8192,0.03525866733656989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,1024,0.04062311185730828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,10240,0.023836443821589153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,5120,0.04192622171507942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,2048,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,8192,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,2560,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,16384,0.0959813329908583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,3584,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,6144,0.009136888715955947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,10240,0.08673955334557427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,1024,0.009390222529570261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,768,0.4874595536126031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,3584,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,6144,0.04662933283381992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,65536,0.5545822249518501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,2560,0.04314577910635206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,2048,0.013385777672131857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,3584,0.06219289037916395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,10240,0.7730773289998373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,8192,0.025342222717073228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,8192,0.17979110611809623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,3072,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,2048,0.05900355842378405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,1536,0.03088266650835673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,10240,0.03451733456717597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,5120,0.03668355610635545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,1536,0.07014399766921997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,64,0.0018986666368113623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,512,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,7168,0.13829688231150308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,8192,1.3584924274020727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,65536,0.21993511252933076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,8192,0.10008266899320815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,3584,0.011100444528791638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,2560,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,8192,0.041315555572509766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,64,0.033779554896884494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,7168,0.3331937789916992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,5120,0.04812444580925835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,7168,0.01016622202263938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,512,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,7168,0.18561066521538627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,64,0.0019866666860050624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,65536,1.1133333841959636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,6144,0.06598044766320123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,7168,0.0125946667459276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,3584,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,6144,2.0556763543023004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,12288,0.020869332883093093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,4096,0.241284449895223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,32,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,5120,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,1536,0.006745777610275481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,12288,0.6169315444098579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,10240,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,2560,0.024735111329290602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,2048,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,65536,0.08979822529686822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,3584,0.007357333269384172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,16384,0.05228444602754381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,768,0.018124444617165458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,8192,0.40719466739230686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,1536,0.03676355547375149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,51200,0.06268088685141669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,4096,0.022494221727053326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,7168,0.357935110727946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,768,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,1024,0.008264888491895463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,256,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,65536,0.1204035547044542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,1536,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,8192,0.014016888207859464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,6144,0.018858666221300762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,1536,0.015248888068728976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,65536,0.18021867010328504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,12288,0.2069466643863254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,51200,0.2103253338072035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,2560,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,256,0.016891555653678048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,4096,0.007075555622577667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,8192,0.8037377993265787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,7168,0.13996533552805582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,5120,0.010450666977299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,10240,0.018056000272432964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,3584,0.11232444975111221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,16384,0.2300879955291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,1536,0.07543022102779813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,6144,0.7429199748569064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,32,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,7168,0.04488799969355265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,6144,0.04906400044759115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,10240,0.020590222544140287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,256,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,7168,0.13012533717685276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,6144,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,2048,0.00423200014564726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,65536,1.780463112725152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,8192,0.10833244191275702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,12288,0.14931378099653456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,32,0.009139555195967356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,10240,0.4114391009012858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,51200,12.700339423285591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,51200,0.061837335427602134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,5120,0.2730026774936252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,7168,0.06314133273230658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,8192,0.043783111704720386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,2560,0.08146755562888251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,2560,0.04028355413013034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,2560,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,512,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,32,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,2560,0.01885511146651374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,768,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,5120,0.14429867267608643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,512,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,2048,0.022477333744366963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,2560,0.019253333409627277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,10240,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,1536,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,1024,0.011558222273985544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,10240,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,3072,0.02365866634580824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,5120,0.5012550883822972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,128,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,2560,0.021930666433440313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,16384,0.1413431167602539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,256,0.014198222094111972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,512,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,256,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,1024,0.0466302235921224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,2560,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,65536,0.0621982216835022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,128,0.05714044306013319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,65536,0.6662853558858236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,10240,0.02223644488387638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,51200,0.4334897730085585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,3072,0.008459555606047312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,128,0.013527111046844058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,12288,0.18809155623118082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,12288,0.01829600003030565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,64,0.0018622221218215094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,5120,0.12822755177815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,768,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,2048,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,6144,0.01984355515903897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,1024,0.026127111580636766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,512,0.011648889217111798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,768,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,512,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,128,0.017796445224020217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,512,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,512,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,51200,0.1258257759941949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,7168,0.09960089127222697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,10240,0.06703911225001018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,10240,0.17587289545271131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,2560,0.2355768945482042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,128,0.08402933014763726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,16384,0.025224889318148296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,768,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,3584,0.02481599979930454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,5120,0.012201777762836881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,512,0.009491555392742157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,3072,0.030124445756276447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,51200,0.18782844808366564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,3072,0.013437333206335703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,6144,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,8192,0.25811555650499135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,256,0.018210666047202218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,128,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,10240,0.039042668210135564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,512,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,128,0.014488000008794995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,32,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,3072,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,128,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,51200,0.29046667946709526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,1536,0.02843822373284234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,12288,0.08067911201053195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,51200,0.2561306688520643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,3072,0.03477333320511712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,3584,0.012028444144460889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,8192,0.011424000064531961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,16384,0.05019466744528877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,64,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,2048,0.0058266665372583605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,4096,0.025109334124459162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,10240,0.6166106859842936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,2048,0.040829333994123675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,2560,0.02032355633046892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,51200,0.524993790520562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,512,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,256,0.008447999755541483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,65536,2.0286621517605252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,2560,0.04507911205291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,65536,0.4151146676805284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,10240,0.03010222315788269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,128,0.014950222439236112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,3072,0.10668088992436726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,128,0.011227555572986603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,512,0.015486222174432544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,6144,0.03912711143493652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,512,0.05291111270586649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,8192,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,768,0.003917333152559069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,12288,0.13709155718485513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,65536,0.22864090071784127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,512,0.024534222152498033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,16384,0.022032888399230108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,3072,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,65536,0.3535742229885525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,1536,0.011207111179828644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,16384,0.040714667903052434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,256,0.006544000158707301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,4096,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,3072,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,3584,1.4885120391845703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,6144,0.0240320000383589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,128,0.0083155557513237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,2048,0.011697777443461947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,16384,0.0958062211672465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,3584,0.011119999819331698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,6144,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,768,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,7168,0.007714666426181793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,1536,0.009519110951158736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,12288,0.19673777951134575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,2048,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,51200,0.32309778531392414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,12288,0.08279733525382148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,2048,0.01440355512830946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,8192,0.03936444388495551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,65536,0.33663378821478945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,8192,0.09319288863076104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,512,0.033850666549470686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,16384,2.15555297003852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,768,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,3072,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,768,0.020336889558368258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,3584,0.19194666544596353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,2048,0.021137777302000258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,512,0.0199608885579639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,51200,0.38890666431850857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,12288,0.024246222443050806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,128,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,10240,0.012984889249006907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,64,0.01180622229973475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,6144,0.04328533344798618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,3072,0.01971733404530419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,512,0.013613333304723104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,768,0.030047999487982854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,5120,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,5120,0.10498488611645168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,128,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,256,0.04401155643992954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,16384,0.04619466596179538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,2048,0.018179555733998615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,65536,0.4221182134416368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,1536,0.019813333948453266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,65536,0.07714222537146674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,2560,0.04085510969161987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,2048,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,10240,0.16691911220550537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,12288,0.08391200171576606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,8192,0.021357332666714985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,65536,0.207458668284946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,51200,0.05520533190833198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,3584,0.011185777684052786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,12288,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,3584,0.3923528989156087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,16384,0.01759999990463257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,51200,1.3159199820624459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,6144,1.2632924185858834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,51200,1.4371200137668187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,256,0.017888888716697693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,1024,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,10240,0.06684533092710707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,8192,0.05838488870196872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,6144,0.01631733278433482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,7168,0.010758221977286868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,10240,3.3613404168023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,12288,0.4440462324354384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,768,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,3072,0.011433777709801992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,1536,0.018560889694425795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,51200,0.1683715581893921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,10240,0.056249777475992836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,2560,0.0068711112770769335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,2560,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,256,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,8192,0.07249866591559516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,16384,0.020066666934225295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,64,0.01676177812947167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,1024,0.011694221860832639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,16384,0.04844089017974006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,4096,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,256,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,5120,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,7168,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,3584,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,3584,0.013901333014170328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,2048,0.005327111317051782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,5120,0.024135111106766596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,4096,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,65536,2.740446302625868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,768,0.029898666673236426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,7168,0.017430222696728177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,10240,0.026891556051042344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,16384,0.7987839910719129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,128,0.015003555350833468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,4096,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,128,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,3072,0.06384888622495863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,6144,0.15801689359876844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,3584,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,4096,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,8192,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,1024,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,1024,0.03495555453830295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,6144,0.036421332094404436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,1536,0.06651466422610812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,32,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,51200,0.11292978127797444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,10240,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,65536,0.425181335873074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,512,0.03374400072627597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,10240,0.03082044588194953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,1024,0.20604977342817518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,65536,0.15261956055959067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,1024,0.010149333212110732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,32,0.0359208881855011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,768,0.11574844519297282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,4096,0.049081779188579984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,16384,0.030379556947284277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,6144,0.0717466672261556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,2048,0.004594666676388847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,3072,0.12217599815792507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,768,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,2048,0.025199999411900837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,3584,0.09649599923027886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,1024,0.11782311068640815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,10240,0.17733600404527453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,10240,0.012887111140622033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,768,0.022665777140193518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,8192,0.06353688902325101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,128,0.015088000231319003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,5120,0.016647110382715862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,4096,0.019888889458444383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,128,0.05824266539679634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,8192,0.37644089592827695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,64,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,1024,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,12288,1.3979262246025934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,4096,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,5120,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,8192,0.032623112201690674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,16384,0.03938755393028259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,3072,0.04717955655521817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,7168,0.14470222261216906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,12288,0.03241422110133701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,256,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,7168,0.10283555587132771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,12288,0.07998044623268975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,768,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,1536,0.008449777960777283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,256,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,8192,0.4302977720896403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,5120,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,1536,0.03141244583659702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,2048,0.023631999890009563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,12288,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,128,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,5120,1.1804879506429036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,128,0.02351555559370253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,3584,0.01680177781316969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,768,0.008101333346631791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,3584,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,768,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,10240,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,4096,0.044886221488316856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,128,0.010246222217877706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,51200,0.1728737751642863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,256,0.015038222074508667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,5120,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,768,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,7168,0.07982666624916925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,768,0.1379279957877265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,8192,0.014607111612955729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,512,0.031850665807724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,1536,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,5120,0.06971910927030775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,2048,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,768,0.017972444494565327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,3584,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,6144,0.027629334065649245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,65536,17.199442545572918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,32,0.003411555455790626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,65536,0.08431022034751044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,3072,0.018904889623324077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,16384,0.0389120015833113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,8192,0.0292195545302497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,51200,0.03564000129699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,512,0.02843022346496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,16384,0.0734586649470859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,128,0.017127111554145813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,5120,0.01700444519519806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,6144,0.13201599650912815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,512,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,3584,0.027354665928416785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,4096,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,7168,0.5839733547634548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,6144,0.08556622266769409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,65536,0.07878133323457506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,10240,0.026742221580611333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,10240,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,128,0.04715199934111702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,4096,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,64,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,1024,0.011571555501884885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,51200,0.47057512071397567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,512,0.018592000007629395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,1536,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,7168,0.031109333038330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,10240,0.11476622687445746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,3584,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,2048,0.006809777683681912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,16384,0.24980266888936362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,16384,0.028393778536054824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,6144,0.18911822636922201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,768,0.02972266740269131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,3584,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,5120,0.01500888831085629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,2560,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,6144,0.01092266705301073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,3072,0.02507466740078396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,8192,0.015395555231306287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,65536,0.2879297733306885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,2560,0.027927110592524212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,7168,0.029467556211683486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,12288,0.029431111282772485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,16384,0.03196977906756931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,3072,0.026320889592170715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,2048,0.18101333247290718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,1024,0.01131199962562985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,1024,0.03611822260750665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,128,0.014926221635606555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,65536,0.021187555458810594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,2560,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,10240,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,65536,0.011941333611806234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,512,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,6144,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,32,0.07444444629881117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,512,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,3584,0.022076444493399724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,5120,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,512,0.012111110819710625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,51200,0.03834133346875509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,16384,0.06539822287029691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,2048,0.10567733314302231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,51200,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,1536,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,12288,2.2118977440728083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,768,0.013361777696344586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,1024,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,2560,0.02421866688463423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,32,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,4096,0.04051377707057529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,12288,0.04864533411131965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,768,0.006138666636413998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,128,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,1536,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,128,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,65536,0.10570133394665188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,3584,0.03796799977620443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,7168,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,51200,0.1489057805803087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,7168,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,3584,0.03053955568207635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,3584,0.045659555329216846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,7168,0.1000337799390157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,7168,0.03513955407672458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,256,0.10111911429299249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,6144,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,6144,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,51200,0.14810933007134333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,3584,0.02253688871860504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,65536,0.06478399700588651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,12288,0.0565333366394043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,2560,0.05619644456439548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,10240,0.07313244210349189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,256,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,8192,2.1232666439480252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,1536,0.027836445305082533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,4096,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,7168,0.057223108079698354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,1536,0.01832266648610433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,7168,0.01811911165714264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,3584,0.1511982282002767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,32,0.026360000173250835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,3072,0.018352889352374606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,1024,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,10240,0.006127111199829314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,6144,0.6157626575893825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,512,0.021624000536070928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,768,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,512,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,12288,0.0373946660094791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,768,0.015088000231319003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,3072,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,64,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,5120,0.0477804442246755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,12288,0.021262221866183813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,2560,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,1024,0.006883555402358373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,7168,0.08506133159001668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,768,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,2048,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,256,0.007012444237867991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,2048,0.1689608891805013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,6144,0.08719644281599258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,512,0.01700799994998508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,51200,0.12442489465077718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,12288,0.02349511120054457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,2560,0.02132533325089349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,1536,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,3072,0.051820443736182324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,2048,0.007054222126801808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,128,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,3584,0.008709333009190029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,2048,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,32,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,10240,0.03065511253145006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,7168,0.010440888504187265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,3584,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,6144,0.010306666294733683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,32,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,65536,3.6100730895996094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,3584,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,1536,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,768,0.008745777938101027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,1536,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,256,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,10240,0.16469421651628283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,7168,0.05697421895133125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,6144,0.013174222575293647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,128,0.025041777226659987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,51200,0.10472266541586982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,128,0.015536889433860779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,2560,0.12729599740770128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,256,0.015136000182893542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,12288,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,65536,0.06837599807315402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,2048,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,1024,0.033960888783137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,128,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,3072,0.06486755609512329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,12288,0.05025955703523424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,128,0.03297066688537598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,8192,0.030694223112530176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,8192,0.9560391108194987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,6144,0.04704622096485562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,10240,0.022634666826989915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,2048,0.0380328893661499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,65536,1.1945199966430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,12288,0.8789991272820367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,3072,0.49266666836208767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,2048,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,256,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,1536,0.3404248820410834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,7168,0.03110577662785848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,8192,0.05650666687223646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,65536,0.08534310923682319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,2048,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,6144,0.01700266698996226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,8192,0.11324889130062527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,256,0.012838222086429596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,6144,0.03165333469708761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,8192,0.116321775648329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,1024,1.191626654730903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,3072,0.02997066577275594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,16384,0.04292444388071696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,1024,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,4096,0.059290667374928795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,10240,0.04338666796684265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,768,0.01332800007528729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,768,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,3584,0.04002666804525588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,16384,0.026081777281231348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,5120,0.018816888332366943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,5120,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,8192,0.07671022415161133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,10240,0.06856266657511394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,12288,0.042803555727005005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,65536,0.1900391048855252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,32,0.009463999834325578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,65536,1.907973395453559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,65536,0.07146310806274414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,1536,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,1536,0.09343377749125163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,3072,0.013836444252067141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,8192,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,256,0.0058346668051348785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,2048,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,128,0.017270222306251526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,3584,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,512,0.002789333255754577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,10240,0.02846222122510274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,7168,0.027366222606764898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,51200,0.2265404330359565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,2048,0.030880000856187608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,6144,0.08964088890287612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,6144,0.008765333228641087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,10240,0.20227732923295763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,6144,0.024364444944593642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,2560,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,512,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,3584,0.0188773340649075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,10240,0.06560177935494317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,256,0.020446222689416673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,256,0.018575110369258456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,65536,0.18932978312174478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,3584,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,2560,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,128,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,6144,0.1063235534562005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,51200,0.046563555796941124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,1536,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,1536,0.036652445793151855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,10240,0.13502489195929632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,51200,0.05849866734610664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,10240,0.1515395508872138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,65536,0.4335813257429335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,8192,0.03365688853793674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,1024,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,3072,0.019866666860050626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,1024,0.006042666733264923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,51200,0.06342311037911309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,5120,0.02779200010829502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,1536,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,512,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,6144,0.0326257778538598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,64,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,12288,0.03838222225507101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,64,0.07276266813278198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,7168,0.019036443697081674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,51200,0.19395821624332002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,4096,0.008612444831265343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,12288,0.024534222152498033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,1024,0.03078311019473606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,4096,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,10240,0.22244177924262154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,7168,0.46219555536905926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,8192,0.14066221978929308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,7168,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,5120,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,7168,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,12288,0.1085546678966946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,65536,0.6836462020874023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,64,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,65536,0.06350844436221652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,1536,0.0362755556901296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,6144,0.024813334147135418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,3072,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,5120,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,10240,0.2474693457285563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,51200,2.308880064222548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,65536,1.9505163828531902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,3072,0.010615111225181155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,10240,0.3374355634053548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,65536,0.08340355422761704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,2560,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,2560,0.013912000589900546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,128,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,8192,0.19029600090450713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,5120,0.05827199750476413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,1536,0.018028444714016385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,4096,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,10240,0.014180445008807711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,10240,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,51200,0.06582488616307576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,32,0.07100177473492093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,3584,0.0959635575612386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,8192,0.12223822540707058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,3072,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,8192,0.31357065836588544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,512,0.023041778140597876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,256,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,512,0.011327111058764987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,512,0.0037760001917680106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,64,0.005640888793600931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,16384,0.6548568937513564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,2048,0.01889333294497596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,2560,0.044963555203543774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,2048,0.09971733225716485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,16384,0.23214488559299043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,2560,0.03276266654332479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,10240,0.11505866050720215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,3072,0.02878755662176344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,256,0.018775110443433125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,16384,0.027789334456125896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,32,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,3072,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,128,0.013432000246312885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,2560,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,512,0.017684444785118103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,7168,0.013087111214796701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,10240,0.03367022342152066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,4096,0.020251555575264823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,8192,0.1303608947330051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,10240,0.03603733248180813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,2560,0.01587466730011834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,10240,0.14487111568450928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,2048,0.018555555078718398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,4096,0.07275199890136719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,2048,0.013528889252079858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,128,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,8192,0.030343112018373277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,256,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,51200,7.848815070258246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,64,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,4096,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,16384,0.1772444380654229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,8192,0.019274666905403137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,2560,0.14652089277903238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,2048,0.03179733289612664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,8192,0.020169778002632987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,8192,0.04258844587537977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,7168,0.07953155703014798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,3584,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,2560,0.0661111142900255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,6144,0.060679111215803355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,256,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,8192,0.038698666625552706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,65536,0.060679111215803355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,4096,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,64,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,3584,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,1024,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,3584,0.4792053434583876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,512,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,2560,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,4096,0.015079110860824585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,1024,0.028079999817742243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,768,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,5120,0.11840355396270752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,128,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,1536,0.018608889646000333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,2048,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,3584,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,2048,0.047263112333085805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,3584,0.10752355390124851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,16384,0.029824889368481104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,1536,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,2560,0.08203111092249553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,3584,0.03214933474858602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,12288,0.09712800052430894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,8192,0.014692443940374585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,10240,0.2530951235029432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,128,0.0254711111386617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,65536,0.042340444193945996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,65536,2.690132353040907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,3072,0.05893333090676201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,128,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,8192,0.07198844353357951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,10240,0.9186506271362305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,6144,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,8192,0.027443556321991816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,64,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,4096,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,1024,0.005688888745175467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,5120,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,3072,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,6144,0.019208888212839764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,8192,0.08528266350428264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,1536,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,4096,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,5120,0.0381315549214681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,51200,1.2254355748494465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,16384,0.030266665750079687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,4096,0.04760800136460198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,65536,0.2867066595289442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,256,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,512,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,128,0.013981334037250943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,2560,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,5120,0.019997333486874897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,4096,0.008255999949243333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,1536,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,51200,0.9452853732638888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,65536,1.4673048655192058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,768,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,65536,0.08526755703820123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,3072,0.03221155537499322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,3584,0.02364088926050398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,4096,0.018980445133315194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,64,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,768,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,512,0.0063991112013657885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,512,0.0058551111982928375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,3072,0.4095902178022597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,7168,0.06269155608283149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,6144,0.08538311057620579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,512,0.027588443623648748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,8192,0.02570399973127577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,4096,0.026111111044883728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,2560,0.044680890109803945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,65536,0.06300799714194404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,65536,4.122839185926649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,128,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,51200,0.7943155500623914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,2560,0.008264000217119852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,4096,0.0636319981680976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,8192,0.022485334012243483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,64,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,128,0.12788711653815374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,768,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,65536,0.2752266724904378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,1536,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,10240,0.044043554200066455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,128,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,2560,0.005064888960785336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,65536,0.0481315553188324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,1024,0.006672888994216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,2560,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,768,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,12288,0.23029067781236437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,768,0.018040888839297824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,3584,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,128,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,7168,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,6144,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,3584,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,3584,0.06384088595708211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,768,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,1024,0.09609511163499619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,6144,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,2048,0.0876435571246677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,1024,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,768,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,51200,0.1535680029127333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,16384,0.03015822172164917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,1536,0.010185778141021729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,16384,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,64,0.005971555494599872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,2560,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,6144,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,4096,0.07421955797407362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,2048,0.01296266665061315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,2048,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,64,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,8192,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,3584,0.07389422257741292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,512,0.012117333710193634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,5120,0.06866666343477038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,51200,0.022618666291236877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,5120,0.02818044357829624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,51200,0.15817066033681235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,64,0.06035111347834269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,3584,0.033526221911112465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,65536,0.9735867182413737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,6144,0.04462577899297079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,2048,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,2560,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,5120,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,1536,0.01129066695769628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,2048,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,7168,0.03893688983387417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,768,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,7168,0.01532977819442749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,512,0.039113776551352605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,128,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,4096,0.14250132772657606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,512,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,16384,0.017628444565667045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,2048,0.009866666462686326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,256,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,16384,0.0530008872350057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,256,0.03299288948376974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,6144,0.01752000053723653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,7168,0.01035644445154402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,51200,0.435898674858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,64,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,32,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,256,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,8192,0.5141439967685276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,768,0.003220444545149803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,2048,0.03439911206563314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,2560,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,1024,0.05888711081610786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,5120,0.04899377624193827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,12288,0.08336000310050116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,128,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,768,0.016202666693263583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,768,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,2560,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,10240,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,1536,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,10240,0.08917422427071466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,7168,0.035303112533357404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,1024,0.060462223158942334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,7168,0.052375111314985484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,64,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,3072,0.1128515534930759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,7168,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,6144,0.07023555702633329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,1536,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,2048,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,3584,0.02349511120054457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,768,0.01922044489118788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,7168,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,6144,0.01572177807490031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,10240,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,2048,0.006464889066086874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,3072,0.006679111056857639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,2560,0.4331653383043077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,64,0.019007111589113872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,12288,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,3584,0.1976755592558119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,2560,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,7168,0.5935013559129503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,256,0.006152888966931238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,2560,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,1536,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,3072,0.04878133204248217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,512,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,8192,0.009125333693292405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,32,0.016691555579503376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,512,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,5120,0.01219555570019616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,4096,0.011020444333553314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,12288,0.2988639937506782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,12288,0.0962488916185167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,1536,0.11361155245039199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,65536,0.31345423062642414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,16384,1.0730844073825412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,512,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,3584,0.018889778190188937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,64,0.04401422209209866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,256,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,65536,0.14726666609446207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,256,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,32,0.048511998520957104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,768,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,7168,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,8192,0.013583111266295115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,51200,0.6104604403177897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,256,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,7168,0.04525244567129347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,6144,0.01534755527973175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,8192,0.06614222129185994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,3072,0.3685617711808946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,2560,0.08233510785632663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,16384,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,4096,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,2048,0.020249777370029025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,12288,0.15778577327728271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,64,0.040094223287370466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,512,0.0029155556112527847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,256,0.009420444567998251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,7168,0.1486524475945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,12288,0.5269333521525065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,4096,0.020262221495310467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,768,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,1024,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,3584,0.010267555713653564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,65536,0.5629057884216309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,128,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,2048,0.020226667324701946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,512,0.1839804384443495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,51200,0.10618311166763306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,51200,0.511007997724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,8192,0.0579928888214959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,1536,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,128,0.010944888823562197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,6144,0.25579023361206055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,2048,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,12288,0.06473866436216566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,4096,0.01588444411754608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,2048,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,2048,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,768,0.0032239999208185407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,4096,0.04442222250832451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,51200,0.5008684264289008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,3584,0.023638221952650283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,256,0.023315555519527856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,128,0.0019226666125986313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,2048,0.011895111037625207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,5120,0.009303111169073317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,2560,0.007912889122962952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,2560,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,12288,0.1008453302913242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,64,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,7168,0.02607111136118571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,7168,0.11392088731129964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,1024,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,8192,0.03029866682158576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,1024,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,128,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,65536,0.6923599772983127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,4096,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,12288,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,1024,0.0602817800309923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,65536,0.5558275646633571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,12288,0.0314977765083313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,1536,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,8192,0.011397333608733283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,2560,0.020589333441522386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,768,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,65536,0.17750133408440483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,12288,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,4096,0.016142222616407607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,3584,0.024392889605628118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,2560,0.19344355000389946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,256,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,6144,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,768,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,5120,0.028502222564485338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,65536,0.09391733010609944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,2560,0.006894222150246303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,4096,0.02760355505678389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,7168,0.2680097685919868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,256,0.01630577776167128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,7168,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,10240,0.026512889398468867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,10240,0.09166222148471409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,4096,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,3584,0.03759466608365377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,512,0.002899555489420891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,65536,1.0261795255872939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,64,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,768,0.0035982223020659555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,7168,0.020907556017239887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,4096,0.008560889297061497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,768,0.14706399705674914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,6144,0.02742133372359806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,768,0.1539342270957099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,256,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,1536,0.036512000693215266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,8192,0.03027555677625868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,6144,0.10555022292666966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,2560,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,2560,0.20404799779256186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,2048,0.20297688908047148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,7168,0.014678221609857349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,7168,0.049591110812293164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,768,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,3584,0.015580443872345818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,65536,0.2460070980919732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,1024,0.00886222223440806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,256,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,64,0.005655999812814925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,8192,0.03330399923854404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,3072,0.016143111719025504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,768,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,51200,0.09588533639907837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,65536,0.09059911304050022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,3072,0.08926577700508966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,10240,0.01666933298110962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,128,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,2560,0.07273511091868083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,256,0.011992000043392181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,65536,0.9120871225992838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,3584,0.033527112669414945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,64,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,5120,0.10564088821411133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,65536,0.39412355422973633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,1024,0.02493511140346527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,5120,0.617577764723036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,3584,0.007826666865083907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,512,0.03122933374510871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,768,0.008849778109126622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,256,0.0064284445510970215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,12288,0.033434665865368314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,16384,0.023973332511054143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,6144,0.1018071108394199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,1536,0.01889866590499878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,3584,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,4096,0.871214230855306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,2560,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,10240,0.5013404422336155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,768,0.2745777765909831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,65536,0.1659662193722195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,1536,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,64,0.008053333395057255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,1536,0.011704000333944956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,5120,0.2241120073530409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,1024,0.015841777125994366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,12288,0.12878399425082737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,256,0.02036088870631324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,1536,0.13846222559611002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,32,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,6144,0.03057510985268487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,5120,0.03514488869243198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,3072,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,6144,0.0984755555788676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,12288,0.10844533310996161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,5120,0.01646577815214793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,8192,0.03296355406443278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,3584,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,12288,0.3041786617702908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,1024,0.23795554372999403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,65536,0.0790604419178433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,768,0.0075866662793689305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,256,0.005716444303592046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,2560,0.021759111020300124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,128,0.03572444452179803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,768,0.04842844605445862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,4096,0.020800888538360596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,256,0.014182221558358936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,256,0.14003200001186794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,1536,0.008377778033415476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,16384,0.1138568851682875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,128,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,2048,0.047673778401480786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,32,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,6144,0.031785776217778526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,16384,0.03493688835038079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,2048,0.02353422178162469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,1024,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,512,0.015096000499195524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,3072,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,10240,0.0166320006052653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,512,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,4096,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,32,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,7168,0.03444177905718485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,65536,0.1809662183125814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,65536,0.1405306657155355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,6144,0.010209778116808997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,1536,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,16384,0.057711111174689404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,768,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,65536,0.19559910562303331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,256,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,3584,0.0474275549252828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,512,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,65536,0.07793511284722222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,1024,0.04777955677774218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,3072,0.020398222737842135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,2560,0.007241778075695038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,65536,5.0461040072970915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,3072,0.04646755589379204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,51200,0.12956533167097303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,10240,0.0165928883685006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,1536,0.01791555517249637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,6144,0.05891911188761393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,5120,0.02815199891726176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,5120,0.009287999735938178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,768,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,512,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,10240,0.01759288873937395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,2560,0.24040177133348253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,2560,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,3584,0.06462844212849934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,2560,0.10484177536434597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,32,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,768,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,16384,0.04334400097529093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,10240,0.08322133620580037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,3072,0.07695377535290189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,5120,0.19485867023468018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,16384,0.05198311143451267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,65536,0.07528889179229736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,256,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,51200,0.11292178101009792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,8192,0.051445331838395864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,1024,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,1536,0.023002665903833177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,256,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,65536,0.13379200299580893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,768,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,6144,0.026054221722814772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,256,0.004933333231343163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,3584,0.30271734131707084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,51200,0.09617688920762803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,2560,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,51200,0.5061742464701334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,2560,0.012701333396964602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,16384,0.04098133246103922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,4096,0.08791110912958781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,8192,0.023494222097926672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,51200,2.7869351704915366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,2048,0.010997333460383944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,7168,3.070625729031033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,64,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,128,0.026699556244744196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,3584,0.02237244447072347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,64,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,4096,0.08368177546395196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,51200,0.9517919752332898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,128,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,1536,0.02789155642191569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,1536,0.008150222400824228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,1536,0.029891557163662378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,256,0.03727466530270047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,4096,0.12224355008867051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,512,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,3072,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,12288,0.019859555694792006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,8192,0.022987556126382615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,3072,0.010120000276300643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,32,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,2560,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,10240,0.017605332864655387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,256,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,768,0.02712533374627431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,768,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,1536,0.02866222129927741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,6144,0.030971556901931763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,1536,0.04295911060439216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,3584,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,10240,0.2209599945280287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,4096,0.009201777478059133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,1024,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,51200,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,65536,0.1381066640218099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,512,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,1536,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,3072,0.01684711045689053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,64,0.005990222096443176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,3072,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,512,0.002198222196764416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,1024,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,256,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,512,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,6144,0.011304000185595619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,3072,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,8192,0.029137776957617864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,1536,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,8192,0.01643022232585483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,1536,0.029101332028706867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,51200,0.051559110482533775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,7168,0.02050755586889055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,10240,0.03863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,3584,0.010822222464614444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,2048,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,6144,0.10536622338824803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,5120,0.08832266595628525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,1024,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,51200,0.873744911617703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,12288,0.062450667222340904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,768,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,256,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,10240,0.08936622407701279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,3584,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,5120,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,7168,0.043418665726979576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,10240,0.1752284500333998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,5120,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,12288,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,2048,0.028684443897671167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,1024,0.10815111133787368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,128,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,128,0.07027466429604425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,16384,0.012799999780125089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,512,0.005392000079154968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,16384,0.027801778581407335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,4096,0.04572355416085985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,1024,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,512,0.004925333377387789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,3584,0.00848444468445248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,3584,0.24046399858262804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,7168,0.013473778135246701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,3072,0.05683466460969713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,10240,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,512,0.01808888879087236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,1024,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,4096,0.19686933358510336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,128,0.004983110974232356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,7168,0.031742221779293485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,768,0.03804355528619554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,128,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,6144,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,1536,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,3584,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,3072,0.016228444046444364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,10240,0.027228444814682007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,65536,0.3185715675354004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,1024,0.05987377961476644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,128,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,768,0.005728000154097875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,7168,0.011810666984981961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,256,0.009328888522254096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,12288,0.02751111156410641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,128,0.11774310800764297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,8192,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,7168,0.025260445144441392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,16384,0.2738426791297065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,7168,0.008599110775523717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,6144,0.04420888755056593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,256,0.005624889085690181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,128,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,6144,0.11329421732160781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,2048,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,65536,5.713437398274739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,7168,0.18056089348263213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,512,0.02974133359061347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,4096,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,1536,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,1536,0.010986666712496014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,51200,0.07459021939171685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,51200,0.8656426535712348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,7168,0.13681422339545354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,2560,0.008382221890820397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,1536,0.026742221580611333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,5120,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,65536,0.08076444599363539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,3072,0.3142533302307129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,512,0.002993777808215883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,256,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,3072,0.04531466629770067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,4096,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,65536,0.15451378292507595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,4096,0.015416888727082146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,128,0.0351875556839837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,5120,0.06149511204825508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,51200,0.06228622463014391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,2560,0.019689778486887615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,1024,0.005992888990375731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,6144,0.029750221305423315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,7168,0.04525688952869839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,12288,0.01984444426165687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,3072,5.846138848198785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,32,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,16384,0.04366222355100843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,10240,0.013561777770519257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,3072,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,16384,0.06865421930948894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,4096,0.6758097542656792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,128,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,2560,0.015998222761683993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,12288,0.03806133402718438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,256,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,768,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,64,0.0019804444164037704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,3072,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,5120,0.06440977917777167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,10240,0.07111822234259711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,512,0.013680000272062091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,2560,0.016308445069524977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,1536,0.025960889127519395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,512,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,6144,0.011805333197116852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,16384,0.25107910897996694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,8192,0.8954800499810113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,512,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,12288,0.04186933239301046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,2048,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,2048,0.03709866603215536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,1024,0.020119110743204754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,65536,0.3498400052388509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,3584,0.09249422285291885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,7168,0.015829333000712924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,128,0.009882666998439366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,512,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,2560,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,3072,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,8192,0.037167999479505755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,2560,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,5120,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,2048,0.011838222543398539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,51200,0.056840889983707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,10240,0.03667733404371474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,6144,0.11270044909583198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,3584,0.03693866729736328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,8192,0.08770310878753662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,6144,0.010935111178292168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,64,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,3584,0.009210666848553551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,2048,0.020941333638297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,10240,0.3085626761118571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,8192,0.01054222219520145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,256,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,1536,0.029691557089487713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,6144,0.37934666209750706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,1024,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,6144,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,32,0.001979555520746443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,128,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,5120,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,2048,0.031331555710898504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,64,0.00610133343272739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,1024,0.01701866587003072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,16384,0.038352890147103205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,2048,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,2048,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,3072,0.006784889019197888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,6144,0.010865777730941772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,4096,0.01239466667175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,6144,0.011635555161370171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,2560,0.057447108957502574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,7168,0.009901333186361525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,2560,0.004850666556093427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,1536,0.029816001653671265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,512,0.007501333124107785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,768,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,8192,0.03292266527811686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,6144,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,512,0.015862221519152325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,51200,0.025203555822372437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,51200,0.1071431106991238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,64,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,6144,0.0136817776494556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,3072,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,65536,0.13065866629282633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,1536,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,256,0.02499377727508545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,10240,0.03480799992879232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,4096,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,12288,0.05900355842378405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,128,0.001968888979819086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,256,0.03860177927547031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,768,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,65536,0.07331466674804688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,6144,0.15219467216067845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,1536,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,16384,2.3877813551161027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,6144,1.3153137630886502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,3584,0.054423110352622144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,768,0.036108444134394325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,65536,0.2508968777126736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,3584,0.05639288822809855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,1024,0.05087822344568041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,2048,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,5120,0.43723466661241317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,1024,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,768,0.1224577824274699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,3584,0.11428711149427627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,768,0.35249866379631895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,16384,0.017870222528775532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,32,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,2560,0.044630222850375705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,2048,0.023885332875781592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,6144,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,16384,0.07981777853435941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,1024,0.005375999957323074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,5120,0.020850666695170935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,5120,0.014549333188268872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,3584,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,12288,0.057842665248446994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,12288,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,3584,0.08836533625920613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,12288,0.011397333608733283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,7168,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,3584,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,65536,0.06293155749638875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,5120,0.018160000443458557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,10240,0.3782533274756537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,8192,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,1536,0.033406221204333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,1536,0.0402124457889133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,10240,0.10764266384972467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,128,0.20604888598124185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,3072,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,7168,0.015113777584499784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,8192,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,10240,0.13579644097222224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,6144,0.7376640107896594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,51200,0.061378664440578885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,256,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,7168,0.026732444763183594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,10240,0.06920444303088717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,65536,0.9763075510660807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,1536,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,1024,0.06758311059739855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,64,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,2560,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,4096,0.009584888815879822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,16384,0.023048889305856492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,3072,0.007690666450394525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,256,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,2048,0.04555288950602213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,256,0.03064889046880934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,3072,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,768,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,32,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,3072,0.02827022141880459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,256,0.14575289355383977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,16384,0.2787840101453993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,16384,0.04264266623391045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,1024,0.009635555247465769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,768,0.023544889357354905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,3072,0.044622222582499184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,2048,0.02397777802414364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,512,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,2560,0.02093511157565647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,512,0.018014222383499146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,5120,0.012611555556456247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,256,0.013381332986884646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,2560,0.024139554964171514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,2560,0.029637333419587877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,1024,0.016751110553741455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,2048,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,5120,0.022229333718617756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,2048,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,3072,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,10240,0.16331199804941812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,4096,0.1339226696226332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,4096,0.10136266549428304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,128,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,3584,0.011031111081441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,5120,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,7168,0.09824888573752509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,12288,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,12288,0.03532355692651536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,512,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,16384,0.0728693339559767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,16384,0.022230222821235657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,2560,0.04348799917432997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,10240,0.020485333270496793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,51200,1.883819580078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,16384,0.022749332918061152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,3584,0.03635822070969476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,12288,0.13583466741773817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,3584,0.021846221552954778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,16384,0.047296888298458524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,32,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,12288,0.02281155520015293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,768,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,6144,0.03033866816096836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,512,0.0053004444473319584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,2560,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,12288,0.5209555625915527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,1536,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,768,0.016170667277442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,64,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,6144,0.013432000246312885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,65536,0.07328444719314575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,768,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,32,0.11509154902564155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,7168,0.1265217728084988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,128,0.030639999442630347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,51200,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,768,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,5120,0.02518400053183238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,128,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,6144,0.020236444142129686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,4096,0.032454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,2560,0.03061422374513414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,7168,0.08187733093897502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,3584,0.0584462218814426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,2048,0.046663112110561795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,6144,0.009561777942710454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,16384,1.1519386503431532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,8192,0.561896006266276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,64,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,6144,0.05271466573079427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,1536,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,6144,0.012960889273219638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,6144,0.02197511163022783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,5120,0.00869688888390859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,768,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,64,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,51200,0.03752266698413425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,3072,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,64,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,5120,0.021159110797776118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,6144,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,2560,0.007592000067234039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,1024,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,7168,0.013569778038395775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,6144,0.029001779026455347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,2048,0.024340444140964087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,65536,0.2579724523756239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,128,0.08804266982608372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,6144,0.027057778504159715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,65536,0.06637866629494561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,1536,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,6144,0.12843644618988037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,12288,0.021645334031846788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,51200,0.2319777806599935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,6144,0.032390223609076604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,32,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,51200,0.07027022043863933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,10240,0.02980888883272807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,1536,0.32527467939588756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,16384,0.12973511219024658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,3584,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,768,0.028920888900756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,2560,0.056003557311164014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,16384,0.06677333513895671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,512,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,64,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,512,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,65536,0.29442667961120605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,256,0.029912001556820337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,128,0.006205333189831839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,1024,0.029841777351167467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,12288,0.07451111078262329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,10240,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,32,0.0029093333416514923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,51200,0.04236888885498047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,8192,0.020302222834693063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,6144,0.02200355629126231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,2560,0.010479110810491772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,8192,0.14703199598524305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,2560,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,12288,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,8192,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,8192,0.0532897777027554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,8192,0.031814220878813006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,512,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,512,0.0236124445994695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,8192,0.04022577736112807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,7168,0.03123377760251363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,1536,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,256,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,32,0.021055110626750525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,51200,0.0649813347392612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,5120,0.008425777984990014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,3072,0.016920000314712524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,256,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,2560,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,2048,0.005528000079923206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,16384,0.0834337804052565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,2048,0.036269333627488874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,256,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,64,0.11344356007046169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,256,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,10240,0.016024889217482675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,32,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,8192,0.013139555851618448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,4096,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,512,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,2560,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,8192,0.09360000160005356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,65536,0.4033342202504476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,1024,0.08211821979946561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,2048,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,3584,0.05658666955100166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,128,0.005420444326268301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,512,0.0028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,256,0.538224909040663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,12288,0.018397332893477548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,51200,0.07528799772262573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,2560,0.007002666592597961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,6144,0.03297244509061178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,51200,0.07364444600211249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,10240,0.016748444901572335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,8192,0.0509422222773234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,3072,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,4096,0.012625777886973487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,6144,1.126303990681966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,16384,0.020942222740915086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,1536,0.009881777895821465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,5120,0.02719288898838891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,2560,0.03763644562827216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,12288,0.04605688982539707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,4096,0.025815111067559984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,768,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,65536,0.06057688925001356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,128,0.021027555068333943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,7168,0.056477334764268666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,12288,0.06189599964353773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,512,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,1024,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,3584,0.0321742229991489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,6144,0.04313688808017307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,1024,0.02546577817863888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,6144,0.11741866005791558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,1536,0.030647112263573542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,512,0.0075706665714581805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,512,0.01683999929163191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,51200,0.05843822161356608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,3072,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,128,0.0063902222447925145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,2560,0.007888889147175683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,7168,0.04657599992222256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,12288,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,8192,0.05085688829421997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,6144,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,512,0.02221333318286472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,2560,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,10240,0.02959644463327196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,256,0.009726222190592024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,51200,0.04506044586499532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,1024,0.017235555582576327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,16384,0.023744889431529578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,1024,0.003641777982314428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,51200,0.10709244675106472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,51200,2.284464094373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,4096,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,12288,0.09768088658650716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,128,0.06088888645172119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,10240,0.03161688976817661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,1024,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,2560,0.03455111053254869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,65536,0.06690133280224271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,51200,0.22544354862636992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,64,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,3072,0.006174222048785951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,128,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,4096,0.338351117240058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,65536,0.031354665756225586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,1024,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,768,0.11908622582753499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,51200,0.10647644599278767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,3584,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,128,0.0214720004134708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,6144,0.017859554953045316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,5120,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,4096,0.049950222174326576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,6144,0.05020622081226773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,7168,0.012825777961148156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,2048,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,5120,0.01759111053413815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,2048,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,256,0.002552888873550627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,2048,0.01789866718981001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,8192,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,8192,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,512,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,8192,0.08147555589675903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,3584,0.010273777776294285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,7168,0.04076000054677328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,2048,0.009194666312800514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,6144,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,7168,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,4096,0.0140684445699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,10240,0.08061155345704821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,1536,2.6311963399251304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,65536,0.08567466338475545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,5120,0.009494221872753566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,7168,0.17436533504062227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,256,0.03866844375928243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,768,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,16384,2.3862968020968967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,7168,0.03860800133811103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,6144,0.7217306560940213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,7168,0.0701306660970052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,3584,0.008649778034951951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,16384,0.03952266772588094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,64,0.0074177781740824384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,256,0.03638133406639099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,16384,0.05721866422229343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,5120,0.217994663450453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,512,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,7168,0.02420622275935279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,2560,0.07455022467507257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,1536,0.026701332794295415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,4096,0.03954044315550063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,12288,0.334500445259942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,12288,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,51200,0.2650204499562581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,12288,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,1536,0.02606577840116289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,512,0.02624888883696662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,2560,0.023984000086784363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,1024,0.025064888927671645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,7168,0.032812443044450544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,5120,0.06316710842980279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,1536,0.021319111188252766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,5120,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,512,0.13760711087120905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,3584,0.05645777781804403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,65536,0.5628799862331814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,16384,0.08894489208857219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,8192,0.01163644426398807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,16384,0.035932444863849215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,128,0.018983999888102215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,8192,0.1773351165983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,2048,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,4096,0.009926222264766693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,3072,0.034838222795062594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,10240,0.05894666910171509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,65536,12.659982469346788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,32,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,512,0.01845066746075948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,2048,0.021645334031846788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,12288,0.21430133448706734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,5120,0.008192000289758047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,2048,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,32,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,6144,0.1111031108432346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,10240,0.017712000343534682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,51200,0.14497955640157065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,3072,0.023896000451511804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,3072,0.12614666091071233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,128,0.0064284445510970215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,512,0.020296888218985665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,16384,0.04598044355710348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,2560,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,5120,0.030064887470669214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,128,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,6144,0.010078222387366826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,10240,0.020592000749376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,3584,0.03430044319894578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,6144,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,2048,0.44835646947224933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,768,0.050127999650107495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,512,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,768,0.046181334389580615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,4096,0.009924444059530893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,16384,0.025234666135576036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,256,0.019540444016456604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,64,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,5120,0.36305689811706543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,3072,0.007761778102980719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,51200,0.05069777700636122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,65536,0.6694995562235514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,6144,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,128,0.029493331909179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,3584,0.309808890024821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,10240,0.14933066897922093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,512,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,128,0.011141333315107556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,5120,0.2698586781819661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,2560,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,5120,0.011078221930397881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,32,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,64,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,2048,0.06697155369652642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,65536,0.06205422348446316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,51200,0.04924355612860786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,3584,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,5120,0.03763644562827216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,128,0.014597333139843412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,6144,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,1024,0.07709778017467923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,1024,0.09917333390977646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,3072,0.032030221488740705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,5120,0.013815999858909182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,2048,0.007010666860474481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,12288,0.029221332735485498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,2048,0.027928001350826685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,12288,0.03455555438995361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,65536,0.08117244640986125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,128,0.023689778314696416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,768,0.025285333395004272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,2560,0.057040890057881675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,2560,0.013497778111033969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,32,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,3072,0.6321866777208116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,512,0.0282008879714542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,7168,0.009935110807418823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,65536,0.6202284495035807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,6144,0.08006755510965984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,12288,0.13481956058078343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,32,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,128,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,8192,0.1991200049718221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,65536,2.257650587293837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,2560,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,512,0.004268444246715969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,8192,0.0320995549360911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,64,0.002625777696569761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,65536,0.3292551040649414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,51200,0.10304266876644558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,256,0.07058488660388522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,256,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,256,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,3072,0.031222220924165513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,64,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,8192,0.17273333337571886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,5120,0.0578133331404792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,2560,0.06531733274459839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,4096,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,1536,0.06925955745908949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,768,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,6144,0.15723644362555608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,10240,0.09187910954157512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,16384,0.23250399695502388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,3072,0.040878223048316106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,3072,0.00977422214216656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,512,0.007104888558387756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,512,0.018400889303949144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,128,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,2048,0.8364275296529134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,51200,1.6683777703179254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,256,0.002607999990383784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,32,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,128,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,4096,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,1024,0.004851555658711327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,51200,0.05989155504438612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,3072,0.021091555555661518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,2560,0.015459555718633862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,2560,0.00674311113026407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,51200,0.2093120018641154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,1536,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,12288,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,512,0.17400177319844565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,6144,0.6103635364108616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,5120,0.0092604441775216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,3072,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,64,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,16384,0.034112887250052557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,6144,0.01982133256064521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,128,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,8192,0.826408863067627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,1024,0.03189955486191644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,16384,0.26844000816345215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,3584,0.02398666739463806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,768,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,7168,0.019336000084877014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,5120,0.04674311147795784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,3584,0.6576888826158311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,512,0.011020444333553314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,5120,0.5231902334425185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,256,0.017880888448821176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,768,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,65536,0.06839199860890706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,2048,0.00849333322710461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,4096,0.02051822178893619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,3072,0.04687822196218702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,16384,0.18243911531236437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,256,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,5120,0.1961902247534858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,12288,0.02036088870631324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,768,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,5120,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,128,0.005320000151793162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,256,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,65536,0.39893778165181476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,16384,0.022509333160188463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,768,0.0041964443193541635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,128,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,256,0.02772533396879832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,1024,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,4096,0.07672533061769274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,768,0.09372978077994452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,51200,0.11679466565450032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,4096,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,51200,0.07485510905583699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,1536,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,32,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,8192,0.05240622162818909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,12288,0.03689422210057577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,2048,0.0646666685740153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,6144,0.026591999663247004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,3072,0.47229957580566406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,51200,0.037745777103636004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,2048,0.08386844396591187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,4096,0.10753422313266331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,10240,0.04543022314707438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,3584,0.051121777958340116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,5120,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,512,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,12288,0.018191999859280057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,5120,0.1728666623433431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,4096,0.017436444759368896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,512,0.00563466673096021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,12288,0.02993511160214742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,8192,0.050015111764272056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,65536,0.05716177490022448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,12288,0.02966488732231988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,7168,0.010520888699425591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,8192,0.03612355391184489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,65536,0.04655199911859301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,1536,0.03969511058595445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,6144,0.040694223509894475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,10240,0.22749243842230904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,6144,0.255237340927124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,5120,0.017264000243610807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,3072,0.036248889234330915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,4096,0.03886577818128798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,1024,0.010294222169452244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,16384,0.025936888323889837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,2048,0.012111110819710625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,32,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,8192,0.15408266915215388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,51200,1.430047141181098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,256,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,10240,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,1536,0.009621333744790819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,256,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,1024,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,1024,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,32,0.0052888890107472735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,4096,0.13048711087968615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,51200,0.0522346662150489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,7168,0.02905244297451443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,12288,0.07271733548906115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,2560,0.014106666048367819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,3072,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,10240,0.028251555230882432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,1536,0.019163555569118924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,3584,0.01995377739270528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,128,0.01126933346192042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,5120,0.04075377848413255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,768,0.007357333269384172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,3584,0.06782133049435086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,5120,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,8192,0.35018134117126465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,4096,0.0406311121251848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,768,0.016784888174798753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,3584,0.016215999921162922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,4096,0.05678400066163805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,256,0.023680888944202002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,51200,0.06120889054404365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,10240,0.7331226666768392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,7168,0.07486844725079007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,64,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,6144,0.03390222125583225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,51200,3.9887483384874134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,1024,0.022623111804326374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,2560,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,10240,0.17429599497053358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,3072,0.007781333393520779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,6144,0.8125466770595975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,10240,0.057279109954833984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,8192,0.04244800077544319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,1536,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,1024,0.009903999666372934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,768,0.0041866666740841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,1536,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,10240,0.041329777903027005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,128,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,3072,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,3584,0.008244444098737504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,16384,0.08814489179187351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,65536,0.010382222632567087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,5120,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,4096,0.014928888943460254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,7168,0.0068959999415609576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,768,0.024316444993019104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,512,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,12288,0.062149332629309766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,1024,0.06896089182959662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,4096,0.11041422022713555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,2560,0.006015111174848344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,2048,0.04241155584653219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,12288,0.018567111757066514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,6144,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,7168,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,2048,0.07786666684680514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,1024,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,8192,0.010656000011497073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,3584,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,2048,0.007101332975758447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,512,0.00543200017677413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,12288,0.05156800150871277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,4096,0.034359110726250544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,10240,0.06247111161549886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,5120,0.24142222934299043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,3584,0.013678222066826291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,6144,0.022304000125990972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,256,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,2560,0.06990133391486274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,1536,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,1024,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,1024,0.008380444513426887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,51200,1.7479377322726781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,10240,1.80970213148329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,256,0.006186666587988536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,16384,0.02348622183005015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,512,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,128,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,16384,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,7168,0.010958222051461538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,65536,0.08816978004243638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,16384,0.04831111099984911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,256,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,3584,0.016160888804329764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,3584,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,256,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,51200,0.056165330939822726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,2560,0.04584799872504341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,3584,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,1536,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,768,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,51200,0.10571377807193333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,1024,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,8192,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,1024,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,12288,0.24821689393785265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,7168,0.032200889454947576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,5120,0.02861600120862325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,2560,0.004022222426202563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,8192,0.08326133092244466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,16384,0.17330666383107504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,5120,0.03255733185344272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,1536,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,1024,0.0035075553589397003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,5120,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,10240,0.1758399936887953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,512,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,2560,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,4096,0.04059200154410468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,64,0.10724711418151855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,2048,0.018925334016482036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,3584,0.02530577778816223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,128,0.02109866672092014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,51200,0.06473333305782743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,12288,0.3662453227572971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,10240,0.04194577866130405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,64,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,8192,0.10281866788864136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,51200,0.04378222094641792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,3584,0.03983022107018365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,1024,0.015278221832381355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,8192,0.05007288853327433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,5120,0.0262737770875295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,32,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,4096,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,51200,0.059227552678849965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,51200,0.15512177679273817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,32,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,6144,0.019186667270130582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,16384,0.031167109807332356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,128,0.018979556030697294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,8192,0.03349599904484219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,51200,0.3776711093054877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,2048,0.048997332652409874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,768,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,65536,1.0728123982747395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,512,0.015045333239767285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,12288,0.30651288562350804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,3584,0.024375999967257183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,16384,0.019291554888089497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,12288,0.03173955612712436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,2560,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,10240,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,512,0.013757333159446716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,2048,0.023318222827381555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,16384,0.02106844385464986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,16384,0.17029244369930693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,51200,3.7445021735297312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,1536,0.047114666965272695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,3584,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,3584,0.0722915530204773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,2048,0.005389333185222413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,10240,0.019888889458444383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,2048,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,3072,0.017753778232468497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,5120,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,1536,0.028195556667115953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,10240,0.017116443978415597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,64,0.039176887936062284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,768,0.011223110887739392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,5120,1.067339579264323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,10240,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,12288,0.08406222528881496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,6144,0.012749333348539142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,12288,0.08050400018692017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,7168,0.02407466702991062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,6144,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,256,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,3072,0.026219555073314246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,12288,0.055982225471072726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,4096,0.7451919979519315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,4096,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,256,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,32,0.006508444332414203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,6144,0.014696000350846184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,1536,0.01000711073478063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,3584,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,64,0.01625066664483812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,10240,0.058839109208848744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,2560,0.02029333346419864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,6144,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,4096,0.17245511213938394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,65536,0.02477333280775282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,65536,0.4779777526855469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,3584,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,512,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,7168,0.021385777327749465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,3584,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,16384,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,2048,0.020620443754725985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,768,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,128,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,1536,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,512,0.0832586685816447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,256,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,3584,0.08342844247817993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,6144,0.09752621915605332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,128,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,1536,0.02639466689692603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,3584,0.008604444563388824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,2048,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,1024,0.015211555692884656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,2048,0.01942133406798045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,8192,0.037224001354641385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,6144,0.509119987487793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,6144,0.02847111225128174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,2560,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,768,0.0038168889780839286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,2048,0.12881510787540013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,65536,8.36542002360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,1024,0.018191999859280057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,65536,0.10702488819758098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,16384,0.16140000025431314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,10240,0.024492444263564214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,12288,0.05258311165703667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,768,0.09413777457343207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,512,0.04828177889188131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,10240,0.06469155682457818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,128,0.017166222135225933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,16384,0.08415910932752822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,256,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,65536,2.205289840698242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,1536,0.008575110799736446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,16384,1.8084498511420355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,32,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,1024,0.016563556260532804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,2048,0.03716977768474155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,4096,0.05636355612013075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,2048,0.00684533350997501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,1024,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,1536,0.012879999975363413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,65536,0.5291377703348795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,16384,0.018972444865438674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,16384,0.028252445989184912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,51200,0.05747110976113213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,256,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,512,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,16384,0.06492800182766385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,12288,0.004980444494220945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,512,0.0025617778301239014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,8192,0.031148443619410198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,2048,0.03548177745607164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,2560,0.11410133043924968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,5120,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,2560,0.004969777746333016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,3072,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,10240,1.9713448418511286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,5120,0.029959112405776978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,8192,0.02681422233581543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,2048,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,64,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,2048,0.03644444545110067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,2560,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,3584,0.1320942242940267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,2560,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,5120,0.045596443944507174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,2560,0.03254577848646376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,1536,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,1024,0.03343288766013251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,16384,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,1024,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,128,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,768,0.012684444586435953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,4096,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,65536,0.06723644336064656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,64,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,10240,0.044658667511410184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,8192,0.11003022061453925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,1024,0.012515555653307172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,16384,0.15453067090776232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,16384,0.10646400186750625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,7168,0.03597866826587253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,512,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,4096,0.04571911030345493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,7168,0.059188445409139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,768,0.011600889265537262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,256,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,65536,0.428546667098999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,12288,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,64,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,2560,1.0796800189548068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,2560,0.011619555453459421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,2560,0.022315555148654517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,1024,0.01603822244538201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,10240,0.03410311208830939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,8192,0.013371555341614617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,7168,0.14633066124386257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,64,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,128,0.015576889117558798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,6144,0.14530489179823133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,10240,0.06938488615883721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,8192,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,16384,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,51200,0.07060711251364814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,3584,0.019653333557976615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,128,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,12288,0.031763556930753924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,64,0.005501333210203383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,2560,0.018360000517633226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,1024,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,65536,0.09308089150322808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,12288,0.03418311145570543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,1024,0.03700000047683716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,2048,0.004237333519591225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,12288,0.016917333006858826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,51200,0.06392888890372382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,16384,0.06220000320010715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,3072,0.0384568903181288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,128,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,3072,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,8192,0.028866665230857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,256,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,51200,1.680761761135525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,51200,0.3668417665693495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,512,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,8192,0.7031031184726291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,7168,0.093414220545027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,12288,0.04112711217668321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,1536,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,3072,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,12288,0.46418131722344297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,7168,0.15633955266740587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,1536,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,1024,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,64,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,6144,0.025683555338117812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,256,0.0029146667155954572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,512,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,128,0.0403422216574351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,768,0.01886666648917728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,8192,0.02453244394726223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,768,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,3584,0.036673777633243136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,2048,0.007129777636792924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,65536,0.0902257760365804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,10240,0.44605689578586155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,128,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,10240,0.05428533421622383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,1536,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,8192,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,128,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,1536,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,10240,0.025637333591779072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,1536,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,32,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,16384,0.09633066919114854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,10240,0.03387555480003357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,32,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,65536,0.5758906470404731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,768,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,6144,0.3783066537645128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,1536,0.04569066564242045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,32,0.0017831110292010838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,4096,0.016255999604860943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,8192,0.02625688910484314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,32,0.001783999924858411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,128,0.015836444165971544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,8192,0.07183644506666395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,65536,0.5611884329054091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,6144,0.11442222860124375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,5120,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,256,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,8192,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,256,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,128,0.009633777870072259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,10240,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,1024,0.005495111147562663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,2048,0.012688889271683164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,12288,0.0707119968202379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,65536,1.0401200188530815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,3584,0.15127821763356528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,2048,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,6144,0.05395199855168661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,512,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,64,0.004971555537647671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,8192,0.17360266049702963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,4096,0.06016177601284451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,128,0.021874666213989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,3072,0.009191999832789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,16384,0.024840000602934096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,1536,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,65536,1.0734835730658638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,512,0.023761777414215937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,7168,0.015319999721315173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,6144,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,2560,0.05681955814361572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,1024,0.02386133372783661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,3584,0.029120001527998183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,16384,0.07225333319769965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,7168,0.010998222563001843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,768,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,512,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,12288,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,8192,0.015772443678643968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,3584,0.02809422214825948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,8192,0.04214488797717624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,2560,0.02221155497762892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,256,0.028008888165156048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,8192,0.0195448895295461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,8192,0.059055997265709766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,5120,0.0329751107427809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,16384,0.04443200098143684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,1536,0.02441155579355028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,2560,0.04192177785767449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,16384,0.06148622433344523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,65536,0.0588231086730957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,2048,0.09037599960962932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,3584,0.07284177674187554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,10240,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,512,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,5120,0.04555288950602213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,65536,0.08082133531570435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,16384,0.11328532960679795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,2560,0.028431998358832464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,5120,0.05925155348247952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,6144,0.05217511124081082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,4096,0.040256887674331665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,2560,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,3072,0.018935999936527677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,2560,0.017154667112562392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,128,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,16384,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,10240,0.06924800078074138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,768,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,4096,0.03077600068516201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,3072,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,2048,0.012468444804350534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,10240,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,768,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,5120,0.04427999920315213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,10240,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,512,0.032167110178205706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,1024,0.017952000101407368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,2560,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,256,0.02235200007756551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,4096,0.019651555352740817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,256,0.07361244493060641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,10240,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,4096,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,5120,0.01904711127281189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,2048,0.0075662218862109715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,5120,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,51200,0.061837335427602134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,128,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,1536,0.006527110934257507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,2048,0.05741688940260145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,32,0.0018151110659042995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,5120,0.008848889006508721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,2048,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,2048,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,768,0.0158897770775689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,7168,2.0260658264160156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,10240,1.7523075739542644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,51200,0.6032968627081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,16384,0.49518664677937824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,768,0.0038408889538711975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,10240,0.02702399922741784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,768,0.024383111132515803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,32,0.007761778102980719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,2560,0.11509511205885146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,10240,0.034683555364608765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,128,0.018193778064515855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,16384,0.1777226659986708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,512,0.002789333255754577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,256,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,64,0.014134221606784396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,7168,0.02188088827662998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,32,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,6144,0.06012711260053846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,2048,0.02201866606871287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,10240,0.017424000634087455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,3584,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,1536,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,32,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,128,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,10240,3.5197618272569446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,7168,0.062076442771487765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,2560,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,10240,0.013521778086821238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,2048,0.013083555632167392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,5120,0.022408000297016565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,768,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,7168,0.020744888318909537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,32,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,512,0.038324445486068726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,8192,0.012180444267061023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,12288,0.2790995438893636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,6144,0.020571556356218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,1536,0.006527110934257507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,16384,0.3011315663655599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,12288,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,4096,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,512,0.027679110566775005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,768,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,12288,2.092016855875651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,5120,0.2757768895890978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,1536,0.026153778036435444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,1024,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,4096,0.03431466552946303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,1536,0.04088800152142843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,3584,0.07553688685099284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,6144,0.058196445306142174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,4096,0.07131733496983846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,7168,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,768,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,512,0.0072266666425599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,256,0.03250933355755276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,768,0.015772443678643968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,128,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,128,0.02439555525779724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,1536,0.03584711088074578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,16384,0.04036800066630045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,1024,0.020204444726308186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,256,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,7168,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,2560,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,768,0.03150577677620782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,1024,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,7168,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,3072,0.7867040104336209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,7168,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,65536,0.19164533085293242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,4096,0.030193779203626845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,8192,0.6659475432501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,3584,0.047183109654320605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,1536,0.010320888625250923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,128,0.0388995541466607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,4096,0.06492355797025892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,128,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,6144,0.11186577214135064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,51200,0.031034668286641438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,1024,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,5120,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,64,0.0018959999498393799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,16384,0.05499200026194254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,12288,3.067428376939562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,2048,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,512,0.006461333483457565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,65536,0.0872231125831604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,12288,0.06940088669459026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,51200,0.06390488810009427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,1536,0.01278133359220293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,128,0.014497778481907315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,4096,0.015746666325463187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,5120,0.009186666872766282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,1536,0.022799111074871484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,5120,0.02751911183198293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,7168,0.027394667267799377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,12288,0.05256711112128364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,768,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,16384,0.6341342396206325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,10240,0.13906310664282903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,128,0.014523555835088095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,16384,1.5226204130384657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,12288,0.07150844732920329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,128,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,6144,0.02990666694111294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,51200,0.055366221401426524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,512,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,8192,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,4096,0.04957688848177592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,3072,0.011161777708265515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,51200,0.06459644767973158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,768,0.05081955591837565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,128,0.0018995555324686898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,1536,0.03317866722742716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,51200,6.534353044297959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,128,0.6292879846360948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,64,0.006671999891599019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,6144,0.021951110826598272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,16384,0.512481795416938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,64,0.06328533093134563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,51200,0.1572000053193834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,32,0.003129777808984121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,1024,0.005483555710977978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,128,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,51200,0.5902382002936469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,3584,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,10240,0.017154667112562392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,3584,0.1753822167714437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,256,0.01309333327743742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,768,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,6144,0.3684097660912408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,7168,0.035347554418775774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,768,0.007408888803588019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,65536,0.1481777826944987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,1024,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,2560,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,6144,0.06797511047787137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,1024,0.1913671096165975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,8192,0.09684977928797404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,3584,0.05165688859091865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,7168,0.0374257763226827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,4096,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,128,0.020566221740510728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,10240,0.05261511272854275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,8192,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,6144,0.020949333906173706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,1536,0.008819555242856344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,12288,0.10778488715489705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,4096,0.3066942161983914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,2560,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,1536,0.023379556006855432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,7168,0.027299554811583623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,6144,0.1501164436340332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,256,0.020615999897321064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,51200,0.08260266648398505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,10240,0.2822471194797092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,128,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,128,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,51200,0.06934577888912626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,16384,1.1183431413438585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,32,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,1024,0.02164177762137519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,5120,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,64,0.008348444269763099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,65536,0.2902764479319255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,3072,0.027987556325064764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,51200,0.0682471129629347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,4096,0.060327112674713135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,1024,0.01700355609258016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,12288,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,8192,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,8192,0.10759111245473225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,2048,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,768,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,128,0.013991999957296582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,7168,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,3072,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,10240,0.08215022087097168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,3072,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,512,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,5120,0.3986559973822699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,64,0.03709955679045783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,7168,0.027074666486846075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,5120,0.020437333318922255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,8192,0.6159182124667698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,2560,0.028331554598278467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,7168,0.030392888519499037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,256,0.028314666615592107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,128,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,768,0.024559999505678814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,1024,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,65536,0.06609066989686754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,6144,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,4096,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,10240,0.1721866660647922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,64,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,3584,0.03949688871701559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,51200,0.0947306685977512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,256,0.024650666448805068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,8192,0.0362044440375434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,8192,0.05815911293029785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,768,0.03035199973318312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,12288,0.05015644431114197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,7168,0.07365244626998901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,768,0.024442666106753882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,8192,0.03840000099605984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,2048,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,10240,0.06931377781762017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,768,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,51200,0.13103200329674616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,65536,10.414290534125433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,4096,0.00868711123863856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,5120,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,5120,0.05647377835379707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,128,0.027437332603666518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,1024,0.029558221499125164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,16384,0.07774577538172404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,768,0.0341306659910414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,7168,0.022449778185950384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,3584,0.013212444053755866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,1536,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,4096,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,768,0.02756888833310869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,12288,0.031197332673602637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,6144,0.05514488948716057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,512,0.00721866637468338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,768,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,1024,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,32,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,3072,0.04026666614744399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,3584,0.028780443800820246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,65536,0.1762533320320977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,128,0.024547555380397375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,10240,0.012014221813943652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,32,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,2560,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,256,0.025270221961869135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,3072,0.017881777551439073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,7168,0.020568889048364427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,10240,0.1592995590633816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,256,0.01348888874053955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,512,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,16384,0.016545777519543965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,3072,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,4096,0.2059182193544176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,1024,0.04841244551870558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,51200,0.06172088781992594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,1536,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,256,0.03136444422933791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,1024,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,4096,0.5502408875359429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,3072,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,4096,0.04503555430306328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,65536,1.1767341825697157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,1024,0.010634666515721215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,7168,0.11634222666422527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,12288,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,1024,0.01148444414138794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,4096,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,5120,0.01955911186006334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,65536,0.17234222094217935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,256,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,51200,1.5754906336466472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,3072,0.024332443873087566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,1536,0.06944355699751112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,1024,0.003536000019974179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,32,0.0019297777778572505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,6144,0.013281777501106262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,16384,0.26187290085686576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,3072,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,768,0.03873422079616123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,2560,0.05007377929157681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,2560,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,256,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,7168,0.040277332067489624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,512,0.09067200289832221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,3584,0.007684444387753804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,65536,0.24760799937778047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,1024,0.01013777818944719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,10240,0.014808000789748298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,512,0.013090666797426013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,768,0.01625688870747884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,65536,2.18323728773329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,32,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,10240,0.027056000298923914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,768,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,65536,1.3996995290120442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,3584,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,768,0.022224000758594934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,768,0.018294221825069852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,12288,1.3964924282497828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,12288,0.15244978004031712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,6144,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,65536,4.140453338623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,1024,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,6144,0.0440631111462911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,51200,0.051903999514049955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,64,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,4096,0.020167110694779288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,4096,0.16735733879937065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,256,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,16384,0.1590462260776096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,128,0.0022479999396536085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,512,0.025378665990299646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,2048,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,1536,0.03850400116708543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,1024,0.005803555664088991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,6144,0.024363555841975745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,10240,0.05610311031341553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,64,0.020423110988405015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,4096,0.06500533554289076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,8192,0.006132444573773279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,3072,0.532900439368354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,5120,0.027295110954178706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,10240,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,5120,0.06008533636728922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,128,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,1024,0.016422222057978313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,768,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,1536,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,51200,1.7864390479193792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,1536,0.0310791101720598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,65536,0.2524622281392415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,128,0.010283555421564314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,3072,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,256,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,3072,0.8368319935268826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,3584,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,8192,0.09176888730790879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,2560,0.03859022259712219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,256,0.051889777183532715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,51200,0.2699182298448351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,2560,0.3035111162397596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,4096,0.014292443792025248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,7168,0.04661600126160515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,10240,0.05487289031346639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,128,0.01790488925245073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,2048,0.03664711117744446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,3072,0.013723555538389417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,12288,14.795516967773438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,7168,0.0188711103465822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,7168,0.02527911133236355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,7168,0.028453333510292902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,7168,0.023353776997990076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,1536,0.10188177559110854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,51200,0.3539777861701117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,5120,0.036009778579076133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,256,0.020247111717859905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,5120,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,6144,0.035212443934546575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,8192,0.03813599877887302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,1536,0.03731822305255466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,768,0.1092453334066603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,4096,0.01849600010448032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,4096,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,10240,0.03661244445376926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,4096,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,65536,1.4098338021172419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,65536,0.26225778791639537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,2560,0.041089776489469744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,51200,0.13753510846032038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,3072,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,512,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,2560,0.012940444052219391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,6144,0.012133333418104382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,64,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,512,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,10240,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,4096,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,256,0.009386666946940953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,1024,0.021262221866183813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,512,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,2560,0.028821332587136164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,10240,0.0639431079228719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,65536,0.25939467218187123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,1024,0.023338667220539514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,3072,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,3072,0.023933332827356126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,7168,0.010088889135254754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,256,0.002387555523051156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,5120,0.03197688857714335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,10240,0.03687644335958693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,2560,0.015343111422326831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,3584,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,65536,0.08094488912158541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,3584,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,768,0.37434132893880206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,1536,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,12288,0.26351555188496906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,1536,0.03962488969167074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,51200,0.2663022147284614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,64,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,4096,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,512,0.0034888887570963967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,51200,0.9470738304985894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,1536,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,16384,0.025387555360794067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,7168,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,2048,0.029430223835839167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,512,0.05367288986841837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,1024,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,128,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,12288,0.026054221722814772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,3584,0.020616888999938965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,5120,0.14975466993119982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,12288,0.1411457724041409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,10240,0.027112889620992873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,7168,0.0806808869043986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,768,0.007039111521508958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,3072,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,51200,0.19097422228919136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,128,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,1536,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,128,0.0018648888087934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,6144,0.26050933202107746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,2048,0.01828888886504703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,51200,0.31509245766533744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,1024,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,4096,0.015273777974976433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,32,0.0068400001360310455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,2048,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,6144,0.06061955293019613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,3072,4.816636403401692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,65536,0.4291848871443007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,2048,0.007158222297827403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,1024,0.013549333645237817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,4096,0.03740088807211982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,2560,0.007913777397738563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,2560,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,10240,0.03575199842453003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,3072,0.022093334131770666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,64,0.0019315555691719055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,2560,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,32,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,768,0.024136000209384497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,1536,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,16384,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,12288,0.29597332742479115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,10240,0.5843644671969944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,768,0.01755644381046295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,512,0.017416889468828838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,512,0.02197422252760993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,64,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,2048,0.029878222280078467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,10240,0.025493333737055462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,4096,0.09945333003997803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,16384,0.14557777510748968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,65536,0.06092444393369886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,1024,0.013762666947311826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,8192,0.04483911063936022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,10240,0.3251822259691027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,3072,0.01093155559566286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,1536,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,768,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,256,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,65536,0.12603021992577448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,128,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,2048,0.043884442912207715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,128,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,4096,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,2560,0.019136889113320243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,3584,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,16384,0.0420959989229838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,16384,0.020911999874644812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,6144,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,3072,0.014211555322011312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,2048,0.44863197538587785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,7168,0.1690657801098294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,5120,0.14242045084635416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,7168,0.3410800033145481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,2048,0.05019199848175049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,16384,0.06916444169150458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,4096,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,2560,0.015523556205961438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,5120,0.04125422239303589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,65536,0.06782311201095581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,256,0.044116444057888456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,128,0.015842666228612263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,128,0.013863110707865821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,32,0.010557333628336588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,16384,0.04439377784729004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,5120,0.008951110972298516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,51200,1.1281876034206815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,2048,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,5120,0.004025777594910728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,768,0.027264000640975103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,2048,0.005495999836259418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,3072,0.013713777893119387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,6144,0.02571644385655721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,8192,0.01963911122745938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,2560,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,16384,0.010967999696731567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,12288,1.2778275807698567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,4096,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,768,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,1536,0.07618400123384264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,1024,0.01424711114830441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,32,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,51200,0.290791114171346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,256,0.015113777584499784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,1536,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,12288,0.020405333903100755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,2048,0.008884444832801819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,768,0.019534221953815885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,10240,0.24863645765516493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,2048,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,6144,0.7765110863579644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,4096,0.3348977830674913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,7168,0.024892444411913555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,1024,0.014512888259357877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,10240,0.03995733459790548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,8192,0.11018133163452148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,2560,0.021309332715140447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,10240,0.02014666630162133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,3072,0.007348444726732042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,51200,0.1232995589574178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,4096,0.009124444590674506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,5120,0.7328684594896105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,768,0.016916443904240925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,3072,0.07943999767303467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,16384,0.058218664593166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,8192,0.019516444868511625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,2560,0.011178666518794166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,8192,0.0707999997668796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,8192,0.010394666757848527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,5120,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,2560,0.026877333720525105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,1024,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,16384,0.7855066723293729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,8192,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,1024,0.012527111503813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,768,0.11814133326212566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,256,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,65536,0.9228550593058268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,51200,0.3210844463772244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,3072,0.012825777961148156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,51200,0.41718488269382054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,128,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,128,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,4096,0.02029511166943444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,7168,0.10118310981326634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,256,0.014879999889267815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,5120,0.009076444639099969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,1536,0.007551111280918121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,2048,0.040927112102508545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,128,0.01794755624400245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,4096,0.02442666722668542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,5120,0.11254844400617812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,2048,0.019909333851602342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,51200,0.4870266914367676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,512,0.0041697778635554844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,6144,0.03626577721701728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,7168,0.017816889617178176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,65536,0.08520622385872735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,128,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,4096,0.0079297779334916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,16384,0.021165332860416834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,2560,0.013770666387346057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,4096,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,16384,0.17913244830237496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,2560,0.050613333781560264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,768,0.06016088856591118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,7168,0.029493331909179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,8192,0.025303999582926433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,16384,0.26313334041171604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,1024,0.007083555890454187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,7168,0.024540444215138752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,4096,0.010235555469989777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,256,0.005697777701748743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,10240,0.026936888694763184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,256,0.006720888945791457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,16384,0.06714489062627156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,51200,0.14736533164978027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,3584,0.0063635553750726916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,12288,0.011902222202883827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,512,0.0029120000286234748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,2048,0.0376853346824646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,3584,0.023172444767422144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,4096,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,4096,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,128,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,768,0.15189066198137072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,5120,0.1205786731508043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,3072,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,256,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,768,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,7168,0.014865777558750577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,256,0.015419556034935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,2560,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,5120,0.12878399425082737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,4096,0.04803377721044752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,5120,0.16186400254567465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,2560,0.31515643331739635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,2560,0.03478755553563436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,65536,0.6658044391208225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,256,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,128,0.0021013333979580137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,3072,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,256,0.003478222423129611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,10240,0.03156088789304098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,3072,0.007028444773621029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,65536,0.028760890165964764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,512,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,16384,0.03107644451989068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,8192,0.018332444959216647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,12288,0.0370071099864112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,512,0.018896889355447557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,768,0.04627289043532478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,2560,0.30528532134162056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,32,0.015523556205961438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,3072,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,4096,0.053852445549435086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,2560,0.17854755454593232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,4096,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,512,0.002226666650838322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,1024,0.013859555953078799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,512,0.013280889226330651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,8192,0.0277493331167433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,1024,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,512,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,65536,0.07819200224346586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,2560,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,768,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,3072,0.030844443374209937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,4096,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,8192,0.041393776734670006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,10240,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,256,0.005463111317820019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,128,0.00774133370982276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,512,0.05111555589569939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,768,0.018186666899257235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,8192,0.04283110962973701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,8192,0.29759110344780815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,7168,0.010399999717871347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,256,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,128,0.021740444832377966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,7168,0.13297955195109049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,5120,0.029153777493370905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,768,0.009080000221729279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,7168,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,51200,0.07719733317693074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,5120,0.020432000358899433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,1536,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,256,0.011633777783976661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,512,0.10132533311843872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,12288,0.035967999034457736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,512,0.030088000827365454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,32,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,256,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,768,0.0057884446448749965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,51200,0.4521919886271159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,5120,0.008451555338170793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,768,0.023396443989541795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,12288,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,2048,0.01625777781009674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,51200,0.06359466579225329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,512,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,8192,0.034713778230879046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,10240,0.24981066915724012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,8192,0.023038221730126276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,32,0.00600355573826366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,4096,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,768,0.04338399900330437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,4096,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,2048,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,1024,0.04947288831075033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,6144,0.009101333717505137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,128,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,5120,0.01442399952146742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,3584,0.013948443863126965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,65536,0.1570862266752455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,4096,0.015402666396564908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,64,0.011208000282446543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,16384,0.07016977998945448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,768,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,5120,0.15908800231085882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,6144,0.02922044528855218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,16384,0.07759822077221341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,3584,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,10240,0.012641777594884237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,2560,0.05711288584603203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,512,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,2560,0.012317332956526013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,2560,0.08031555679109362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,65536,0.048584000931845776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,1536,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,3584,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,768,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,128,0.001855111163523462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,4096,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,128,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,768,0.043046222792731396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,12288,0.0159599996275372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,512,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,5120,0.008910222185982598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,32,0.0032799999333090256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,5120,0.017160000072585214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,512,0.01407999959256914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,12288,0.014877332581414117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,3584,0.1251306666268243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,5120,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,32,0.0024462222225136226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,6144,0.05917155742645264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,7168,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,4096,0.3566799958546956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,128,0.001963555605875121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,4096,0.03035555614365472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,16384,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,6144,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,51200,0.09462133381101821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,2048,0.008838222258620793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,128,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,8192,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,128,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,768,0.012411555482281579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,7168,0.017535999417304993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,1024,0.023152000374264185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,51200,0.043567998541726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,2048,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,2560,0.016398222910033334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,51200,0.050704889827304415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,3584,0.040496001640955605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,7168,0.11194666226704915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,3584,0.028459555572933618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,256,0.015587555037604438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,4096,0.03537155522240533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,4096,0.011257777611414591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,2560,0.6700915760464139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,2048,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,64,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,65536,0.19199911753336588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,3072,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,8192,0.013463111387358772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,6144,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,256,0.024056888288921777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,6144,0.03632355398601956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,6144,0.22046400441063774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,32,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,5120,0.03202755583657159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,51200,0.11625422371758355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,32,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,3584,0.041504889726638794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,256,0.005036444299750858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,4096,0.03417066733042399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,2560,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,51200,0.02264444364441766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,65536,0.2049057748582628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,10240,0.017761778500345018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,8192,0.14652621746063232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,16384,0.07235555516348945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,256,0.005412444472312927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,6144,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,768,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,768,0.015609777635998197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,10240,0.023214222656355962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,3072,0.00638133328821924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,4096,0.01924800044960446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,128,0.0022684445397721394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,7168,0.020735111501481798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,2560,0.004513777792453766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,512,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,6144,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,51200,0.19818666246202257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,1024,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,256,0.027101332942644756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,4096,0.09752000040478176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,1024,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,8192,0.16199555661943224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,4096,0.09911288817723592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,512,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,3072,0.07078666819466485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,1536,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,65536,0.018826666805479262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,768,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,256,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,1536,0.004453333301676644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,65536,0.1311111052831014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,5120,0.09132089217503865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,3072,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,16384,0.02568000058333079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,32,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,6144,0.017849778135617573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,16384,0.043486220969094165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,65536,0.5936773088243272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,768,0.017286222842004564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,768,0.013076444466908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,256,0.07326133383644952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,65536,0.6432115766737196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,65536,0.16599999533759222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,3072,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,6144,0.02858933475282457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,65536,0.7592550913492838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,1024,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,3072,0.009945777555306753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,5120,0.02664888898531596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,51200,0.25410577985975474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,16384,0.021721776988771226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,8192,0.6329004499647353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,1024,0.021632888250880774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,512,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,3072,0.18813599480523002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,256,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,3072,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,4096,0.0489671097861396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,16384,0.06966488891177706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,51200,0.055753780735863574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,4096,0.016764443781640794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,256,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,8192,0.12971644931369358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,2560,0.021077333225144282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,768,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,5120,0.021988444858127173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,12288,0.03013066781891717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,1024,0.018128888474570382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,768,0.5793306562635634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,5120,0.03893866803910997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,10240,0.020058666666348774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,3072,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,256,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,128,0.005377777748637729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,65536,23.195444742838543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,16384,1.076155556572808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,128,0.0032213332338465583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,2048,0.01148444414138794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,2560,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,16384,0.021148444877730474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,1536,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,7168,0.2002035511864556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,65536,0.012591111163298288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,512,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,768,0.022455111145973206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,128,0.013656888571050433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,128,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,7168,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,51200,0.31668978267245823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,2560,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,2048,0.03783377673890855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,16384,0.04954044355286492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,65536,0.0785964462492201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,512,0.012871111432711283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,10240,0.015653333730167814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,5120,0.07316711213853624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,4096,1.1668240229288738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,6144,0.127710223197937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,12288,0.33454132080078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,6144,0.5085200203789605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,8192,0.04494488901562161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,6144,0.052397333913379245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,12288,0.03049066662788391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,3584,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,10240,0.03260355525546604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,128,0.0022115556316243275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,16384,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,1024,0.005132444616821078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,32,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,512,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,2560,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,1024,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,51200,0.12139999866485596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,3072,0.2260355684492323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,2560,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,10240,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,4096,0.12605599562327066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,6144,0.010749333434634738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,5120,0.0322017769018809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,3584,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,12288,0.032131556007597185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,3584,0.020639111598332722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,2560,0.012617777619096967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,1024,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,8192,0.044759111271964185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,7168,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,51200,0.06592888964547051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,128,0.014525334040323893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,1536,0.022662222385406494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,3584,0.02366133365366194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,12288,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,8192,0.008541333178679148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,2560,0.020573332905769348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,128,0.002228444442152977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,1536,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,3072,0.05182933476236132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,51200,0.1168782181209988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,5120,0.0064382221963670515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,8192,0.04218666752179464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,10240,0.027479999595218237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,16384,0.19762489530775282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,65536,0.07685777876112196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,8192,0.021968000464969214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,6144,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,3072,0.05749333567089505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,4096,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,3584,0.04128888911671109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,256,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,2560,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,32,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,2048,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,2048,0.03759200043148465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,1024,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,3584,0.040228443013297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,2048,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,6144,0.09518222014109294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,64,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,7168,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,65536,0.3142906559838189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,65536,0.09450399875640869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,8192,0.015192000402344598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,64,0.017204445269372728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,16384,0.2791457706027561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,3584,0.010764444039927589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,32,0.0029599999801980127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,768,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,7168,0.02436711059676276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,2048,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,8192,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,10240,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,10240,0.6346657541063097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,8192,0.054901334974500865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,5120,0.0884560015466478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,768,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,768,0.015321777926550971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,8192,0.2664408948686388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,5120,0.017229333519935608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,7168,0.02130222154988183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,1536,0.006797333144479328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,8192,0.4616488880581326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,768,0.04637599984804789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,128,0.09649066792594062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,128,0.019890666007995605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,2048,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,51200,0.6137768957349989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,5120,0.025959110922283594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,5120,0.02168888847033183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,32,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,256,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,3584,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,256,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,5120,0.15004621611701116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,2048,0.02778488728735182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,2048,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,2560,0.12468443976508246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,2048,0.05161066518889534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,8192,1.557596418592665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,2048,0.015763555963834126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,768,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,8192,0.02714577813943227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,10240,0.017430222696728177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,1024,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,3072,0.021967111362351313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,16384,0.018554665976100497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,3072,0.014887111054526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,2560,0.015289778510729471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,2560,0.03300266795688205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,768,0.038450668255488075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,12288,0.24551643265618217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,3072,0.010852444503042432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,512,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,3584,0.052629331747690834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,1024,0.022543110781245764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,2048,0.058987557888031006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,16384,0.3649093310038249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,5120,0.034625778595606484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,6144,0.022691556149058874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,3072,0.024784000383483037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,128,0.020988444487253826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,2560,0.006480888773997624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,51200,0.5433271196153429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,1024,0.0038444445365005066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,16384,0.02812444501452976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,768,0.04489244355095757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,2560,0.0539919998910692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,512,0.07905155420303345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,7168,0.030093332131703694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,12288,1.2220524681939018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,64,0.02609866691960229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,12288,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,768,0.14239911238352457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,3584,0.028471999698215063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,2048,0.018394667241308425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,768,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,12288,0.1948986715740628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,12288,0.04983555608325534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,3584,0.02016799979739719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,3072,0.01459022197458479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,3584,0.014549333188268872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,8192,0.010327111515733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,8192,0.2932719919416639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,8192,0.037984887758890785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,2048,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,768,0.026872000760502283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,64,0.006894222150246303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,2048,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,3072,0.05832888682683309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,768,0.04468533396720886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,12288,0.021242666575643752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,768,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,12288,0.10350933339860703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,1024,0.08075377676222059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,1536,0.004903111192915174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,4096,0.07617244455549452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,16384,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,16384,0.021726222501860723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,128,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,6144,0.010969777901967367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,10240,0.02311466634273529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,10240,0.11471822526719834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,10240,0.10952533615960015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,16384,0.1673804389105903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,64,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,5120,0.023548444112141926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,5120,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,51200,0.3621635437011719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,4096,0.008417777717113495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,16384,0.236154662238227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,1536,0.041872001356548734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,2560,0.013588444226317935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,128,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,1536,0.004548444516128964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,512,0.01925155520439148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,2560,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,65536,0.6791315608554416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,8192,0.02200711104604933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,32,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,128,0.05818844503826565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,256,0.06092622545030382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,32,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,51200,8.398380703396267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,65536,0.08582666847440933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,3584,0.034866667456097074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,64,0.013821333646774292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,16384,0.0452497767077552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,16384,0.04887466629346212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,8192,0.1265688869688246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,65536,0.5532062318589952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,6144,0.0975688894589742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,3072,0.030936890178256567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,3072,0.013625777430004544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,64,0.024995555480321247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,8192,0.017820444371965196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,512,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,65536,0.13003110885620117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,1536,0.015372445185979208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,1024,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,64,0.0022408889813555614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,32,0.018144889010323417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,64,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,12288,0.014468444718254937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,2560,0.015493333339691162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,128,0.0021111110432280437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,128,0.005508444375462002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,5120,0.0916515522532993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,10240,0.013919111755159167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,1536,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,768,0.014879110786649915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,10240,0.014259555273585849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,5120,0.024216888679398432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,16384,0.14667377207014295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,1024,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,2048,0.014767999450365702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,32,0.01753777762254079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,65536,0.2675582302941216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,128,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,768,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,7168,0.030813333061006334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,128,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,5120,0.017308443784713745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,512,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,5120,0.02776177724202474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,3584,0.10175466537475586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,1536,0.11150399843851726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,128,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,7168,0.0995066695743137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,65536,0.057130667898390025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,64,0.0018097776919603348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,768,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,7168,0.2919893264770508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,4096,0.8211741977267795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,3072,0.03712711069318984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,12288,0.08129333125220405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,128,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,8192,0.061462223529815674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,8192,0.021819555097156104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,6144,0.18780267238616943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,10240,0.023751111494170293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,51200,0.20043110847473145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,6144,0.01553066737122006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,3072,0.03813955518934462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,768,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,512,0.012478222449620565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,3584,0.0309840010272132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,2048,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,32,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,7168,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,1024,0.013774221969975365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,10240,0.09023111396365696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,128,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,65536,0.48640802171495223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,10240,0.1309386624230279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,16384,0.02296000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,7168,0.012991111311647626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,16384,0.018106665876176622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,2560,0.004579555657174853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,51200,0.6174257596333822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,16384,0.020632000433074105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,3584,0.02518133322397868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,12288,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,12288,0.01996888882584042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,4096,0.010302222437328762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,4096,0.04119111100832621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,51200,0.05333955420388115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,1536,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,6144,0.013472889032628802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,10240,0.06948533323076037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,65536,3.093231201171875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,3072,0.01349244432316886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,3072,0.13674222098456487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,512,0.026536888546413843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,3072,0.07038399908277723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,12288,0.19075911574893525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,16384,0.021916444102923077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,2048,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,5120,0.26433155271742076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,12288,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,3072,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,3072,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,51200,0.05397866831885444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,1024,0.06207911173502604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,7168,0.1148195531633165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,3584,0.016957332690556843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,128,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,3584,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,51200,32.08930460611979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,128,0.01142133358452055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,10240,0.04478311207559374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,5120,0.05771644247902764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,6144,0.012438221938080259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,6144,0.05247644583384196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,7168,0.027408889598316614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,128,0.04532266656557719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,7168,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,2048,0.0830942193667094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,3072,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,51200,0.07221688826878865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,256,0.0046728890803125174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,65536,0.06787200106514825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,3072,0.015135111080275642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,5120,0.015449777245521545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,128,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,128,0.06575822167926364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,64,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,32,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,128,0.023741333021057978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,3072,0.004592000196377437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,4096,0.1458453337351481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,256,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,2048,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,256,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,51200,0.12084799342685276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,1024,0.0038311113086011675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,64,0.002028444367978308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,1024,0.01880088945229848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,128,0.0019511110666725372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,3072,0.03576533330811395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,4096,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,12288,0.12416356139712864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,1024,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,1536,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,12288,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,3584,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,768,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,128,0.00498844434817632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,128,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,1024,0.003912000192536248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,3072,0.014255111416180929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,3584,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,5120,0.08262044191360474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,2560,0.07886488570107354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,512,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,6144,0.010002666877375709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,51200,2.2826417287190757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,2560,0.05081244309743246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,1536,0.008057778080304464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,3584,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,65536,15.716330634223091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,1536,0.02604355580276913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,3072,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,1536,0.0035031110876136353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,16384,0.24153243170844185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,3584,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,512,0.02256177696916792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,128,0.0899439983897739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,7168,0.7307306925455729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,7168,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,2560,0.01850311126973894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,7168,0.9758212831285266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,32,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,256,0.019949333535300363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,8192,0.0562373333507114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,7168,0.04344088832537333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,12288,0.1528844436009725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,3584,0.03003644280963474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,2048,0.01998044384850396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,2560,0.03340889016787211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,64,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,5120,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,8192,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,2560,0.01994044416480594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,3584,0.11046844720840454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,10240,0.05754755602942573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,65536,0.05113600028885735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,5120,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,1536,0.020489777127901714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,10240,0.37344000074598527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,32,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,8192,0.03864799936612447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,51200,0.05479377839300367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,256,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,51200,0.27139822642008465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,32,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,10240,0.01536799967288971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,3072,0.07016266716851129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,3072,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,2560,0.011417778001891242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,2048,0.046776887443330556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,7168,0.06151911285188463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,6144,0.012486221889654795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,10240,0.020555555820465088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,512,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,1536,0.024620445238219366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,3072,0.3312133418189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,12288,0.22200711568196616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,10240,0.46026934517754453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,256,0.01973155637582143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,51200,13.10252465142144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,10240,0.06304622358746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,1536,0.02875822120242649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,1024,0.012655110822783576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,4096,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,1024,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,12288,0.04236711064974467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,2560,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,7168,0.013241777817408243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,512,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,128,0.012373333175977072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,4096,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,8192,0.024634665913052026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,6144,0.6670506795247396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,3584,0.03638311227162679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,5120,0.02015022271209293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,5120,0.02108000053299798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,12288,0.13372445106506348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,2560,0.03895644346872965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,1024,0.027066666218969557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,32,0.006517333288987477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,2048,0.013144888811641268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,7168,0.02032444377740224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,7168,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,1536,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,51200,2.2598302629258895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,51200,14.212070888943144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,512,0.04280799958440992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,12288,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,512,0.015612444943851896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,65536,0.08166222439871894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,8192,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,3584,0.011703111231327057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,128,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,6144,0.11192533704969619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,1024,0.003903999924659729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,2560,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,8192,0.17643466260698107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,256,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,1024,0.006199999815887875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,12288,0.03693600164519416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,128,0.004952888935804367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,8192,0.0648791127734714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,2560,0.009409777820110321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,1024,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,128,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,3072,0.04392177859942118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,1024,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,3072,0.09682933489481609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,768,0.008261333737108443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,1536,0.17938755618201363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,12288,0.027727110518349543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,1024,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,2560,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,12288,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,32,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,768,0.006106666806671355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,5120,0.04168088899718391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,7168,0.2231360011630588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,51200,0.021937777598698933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,5120,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,128,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,1536,0.026133333643277485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,768,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,32,0.002562666725781229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,12288,0.0347653329372406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,768,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,8192,0.028856890069113836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,3072,0.054697778489854604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,4096,1.7427635192871094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,5120,0.14450311660766602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,16384,0.03088888857099745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,3072,0.015234667393896313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,10240,0.05556178092956543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,8192,0.17502221796247694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,3584,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,8192,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,3072,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,5120,0.021235555410385132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,4096,0.02513066596455044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,1536,0.005766222046481238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,4096,0.02643288837538825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,6144,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,3584,0.011094222466150919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,4096,0.03148977624045478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,65536,0.07089333401785956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,51200,0.07633599970075819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,16384,0.2827128834194607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,4096,0.03066844410366482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,2560,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,51200,0.35769422849019367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,5120,0.023315555519527856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,3072,0.07153333557976617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,2560,0.02200266718864441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,12288,0.018945778409639995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,16384,0.08269244432449341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,65536,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,10240,0.039460443788104586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,65536,0.05676266882154676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,51200,0.44860177569919163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,128,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,64,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,768,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,768,0.02071466710832384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,3072,0.010527999864684211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,2560,0.007378666765160031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,1024,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,1536,0.03598489032851325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,8192,0.02871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,1024,0.011384000380833944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,4096,0.04121689001719157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,1024,0.013240888714790344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,5120,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,16384,0.8281200197007922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,64,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,8192,0.0405982236067454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,4096,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,16384,0.04583644535806444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,8192,0.04333600070741442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,7168,0.05052977800369263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,7168,0.031168000565634832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,2560,0.06816533539030287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,51200,0.1319040060043335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,768,0.11433955033620198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,2560,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,64,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,3584,0.05873066849178738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,6144,0.01924711134698656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,256,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,7168,0.6208835707770454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,7168,0.031884445084465876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,1536,0.0061928890645504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,256,0.005048888838953442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,7168,0.0517537792523702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,1024,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,16384,0.02643999954064687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,4096,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,8192,0.12008978260887994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,64,0.015072888798183866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,7168,0.03609866566128201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,2560,0.605297777387831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,64,0.008230222596062554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,32,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,51200,0.7876275380452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,7168,0.10698044300079346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,8192,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,7168,0.6464684274461534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,6144,0.025285333395004272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,64,0.02514311174551646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,6144,0.02578488820128971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,6144,0.03186755710177951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,51200,0.31670933299594456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,10240,0.08250755733913846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,64,0.011979555918110741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,51200,0.20058933893839517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,10240,0.01606399979856279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,768,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,2560,0.027219555444187585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,12288,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,2048,0.004592888885074192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,2048,0.04019822345839606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,2048,0.032272001107533775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,65536,6.14737786187066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,65536,0.08216444651285808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,256,0.017495110630989075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,12288,0.10060089164310032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,2048,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,512,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,5120,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,128,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,5120,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,6144,0.11475022633870442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,256,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,3072,0.01794488893614875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,256,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,2560,0.008895999855465358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,128,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,7168,0.023843554986847773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,768,0.03408711155255636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,1024,0.023286221755875483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,12288,0.04615733358595106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,2560,0.023287110858493384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,2560,0.33355021476745605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,2560,0.045815110206604004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,3584,0.0158933334880405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,2560,0.0439644455909729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,4096,0.03398844599723816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,10240,0.03741066654523214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,16384,0.022314666046036616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,51200,1.0227973726060655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,256,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,5120,0.021690666675567627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,51200,0.05870044231414795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,3584,0.06178400251600477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,512,0.018575999471876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,768,0.04102755586306254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,1024,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,2560,0.02271644439962175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,64,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,3584,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,8192,0.3557768927680121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,12288,0.037680887513690524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,512,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,16384,0.09420088926951091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,1024,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,6144,0.019319999549123976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,8192,0.09141155746248032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,6144,0.0291955570379893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,128,0.0155404441886478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,2560,0.027980443504121568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,12288,0.04571288824081421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,51200,0.040728890233569674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,51200,0.21635733710394967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,2048,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,2048,0.02873777680926853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,8192,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,256,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,768,0.010808889236715106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,16384,0.10028177499771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,65536,0.1745742162068685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,2560,0.027748444014125403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,32,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,64,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,7168,0.13233689467112222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,1024,0.014204444156752693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,6144,0.028559999333487615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,8192,0.013001778059535556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,51200,0.0518871115313636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,12288,0.028794666131337483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,3584,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,65536,0.09895822074678208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,3584,0.034416000048319496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,32,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,3072,0.02986933456526862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,65536,0.15249421861436632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,3584,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,2560,0.17881866296132407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,5120,0.02327644493844774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,512,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,4096,0.12294310993618435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,8192,0.25796977678934735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,7168,0.010411555568377176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,8192,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,7168,0.06307111183802287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,3584,0.4296862284342448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,16384,0.06695466571384005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,768,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,768,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,1536,0.01754666699303521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,768,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,7168,0.00979911122057173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,4096,0.01441333360142178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,4096,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,51200,0.37414132224188906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,12288,1.9110959370930989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,1024,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,3072,0.009286222358544668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,7168,0.02414488957987891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,512,0.008646222452322641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,65536,0.6644906467861599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,6144,0.09046844641367595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,128,0.017849778135617573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,3584,0.01867111192809211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,3584,0.04820266697141859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,3584,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,32,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,32,0.0284862220287323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,16384,0.1596506701575385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,4096,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,51200,0.04749422272046407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,3072,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,12288,0.07273955477608575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,12288,0.137353777885437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,16384,0.013697778185208639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,7168,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,7168,0.005344888816277186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,16384,0.040748443868425153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,128,0.021593777669800654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,32,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,256,0.032167110178205706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,256,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,16384,0.6432622273763021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,5120,0.016792888442675274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,4096,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,64,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,8192,0.03991466760635376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,10240,0.14435199896494547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,5120,0.011062222222487131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,12288,0.05197777681880527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,3584,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,2560,0.23113510343763563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,5120,0.010423111418883005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,4096,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,128,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,16384,0.08494666549894546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,4096,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,768,0.005154666801293691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,65536,0.05689422289530436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,1536,0.027105778455734253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,7168,0.0118568887313207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,7168,0.030966222286224365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,8192,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,5120,0.07545955313576592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,2048,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,8192,0.09266933467653062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,16384,0.03360622127850851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,12288,0.16942044099171957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,65536,2.738167021009657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,6144,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,64,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,512,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,32,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,12288,0.019036443697081674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,6144,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,16384,0.11284177833133274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,128,0.006325333482689328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,1536,0.010317333042621613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,1536,0.020455999506844413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,512,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,65536,0.08393066459231907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,128,0.016925333274735343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,128,0.00981866651111179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,65536,0.05579555696911282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,128,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,512,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,1024,0.01276000009642707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,128,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,1024,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,512,0.010031110710567897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,4096,0.0417191121313307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,7168,0.6173431078592936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,128,0.021278222401936848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,7168,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,10240,0.33383554882473415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,16384,0.037800888220469155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,1024,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,4096,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,1536,0.013821333646774292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,2560,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,1536,0.013570666313171387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,51200,0.24390933248731825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,12288,2.707299550374349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,8192,0.016341333587964375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,512,0.011289777855078379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,5120,0.019052444232834708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,10240,0.19173067145877412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,5120,0.00959822204377916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,5120,0.1686862177318997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,6144,0.02020355562369029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,12288,0.03455200129085117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,10240,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,128,0.022992889086405437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,128,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,3072,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,2048,0.023383999864260357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,1536,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,512,0.1256586710611979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,6144,0.015474667151769003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,768,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,65536,0.14891289340125188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,32,0.0019413334214025075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,10240,0.014388443695174323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,3584,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,8192,0.014044443766276041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,8192,0.04468355576197306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,65536,0.08510666423373753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,64,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,6144,0.04265333215395609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,6144,0.04474755459361606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,12288,0.023356444305843774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,1536,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,32,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,10240,0.019959110352728102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,128,0.016918222109476726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,2560,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,6144,0.012134222520722283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,128,0.016247999336984422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,1024,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,65536,0.1524533298280504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,10240,0.07720622089174059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,256,0.005780444376998478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,12288,0.025454221500290766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,6144,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,2560,0.02267644471592373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,2560,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,4096,0.07846666706932916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,51200,6.646870930989583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,12288,0.19354755348629424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,768,0.023521777656343248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,6144,0.011683555940786997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,4096,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,65536,0.796421315934923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,768,0.007877333296669854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,6144,0.013092444174819522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,3584,0.02267733381854163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,2048,0.027142221728960674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,4096,0.02915466825167338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,1024,0.026383999321195815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,3072,0.10411910878287421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,5120,0.024596444434589807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,3584,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,16384,0.016439111696349252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,1536,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,5120,0.023345778385798138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,65536,0.15096710787879095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,16384,10.950898912217882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,8192,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,1536,0.03382488754060533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,16384,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,2560,0.05451022254096138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,7168,0.803152879079183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,16384,0.17685333887736002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,6144,0.009380444056457942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,3584,0.01984177695380317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,3072,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,8192,0.15368799368540445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,12288,0.11451999346415202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,6144,0.009977777798970541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,51200,0.017595556047227647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,3072,0.023341332872708637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,8192,0.09191911088095771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,256,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,5120,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,768,0.017916444275114272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,1024,0.004018666843573253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,1536,0.25914043850368923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,12288,0.014242667290899487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,3072,0.024600888291994732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,1024,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,512,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,3584,0.02590933276547326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,4096,0.04786755641301473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,6144,0.03191377719243368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,8192,0.011134222149848938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,51200,0.1392151117324829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,8192,0.02882222334543864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,2048,0.14791644944085017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,768,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,3584,0.016798221402698092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,128,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,3584,0.07984799808926053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,4096,0.015451555450757345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,512,0.007560888926188151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,3072,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,6144,0.008019555773999956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,1536,0.10582666926913792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,128,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,6144,0.0661440028084649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,10240,0.025643555654419795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,3072,0.037800888220469155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,2048,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,5120,0.018216888109842937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,128,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,10240,0.03449511196878221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,1536,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,65536,0.13142132759094238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,256,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,256,0.018124444617165458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,2048,0.07002933157814874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,512,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,8192,0.02245866590076023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,256,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,10240,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,65536,6.938737657335069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,1024,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,6144,0.015711999601787992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,64,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,3072,0.015200000670221118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,16384,1.3130373424953885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,5120,0.15379733509487575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,12288,0.04831733306248983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,1536,0.04731288883421156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,10240,0.047968000173568726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,256,0.006779555645253923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,5120,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,4096,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,51200,0.1432373391257392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,51200,0.3951839870876736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,8192,0.0321999986966451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,512,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,2560,0.011693333586057028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,1024,0.021287999219364587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,2560,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,6144,0.026763555076387193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,16384,0.7045520146687826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,2048,0.02144355575243632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,3584,0.027679999669392902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,2048,0.028203556934992473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,65536,2.0605547163221574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,8192,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,2560,0.009728888670603434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,768,0.014388443695174323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,3072,0.033879998657438494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,3072,0.04049511088265313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,512,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,128,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,256,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,12288,0.019154666198624503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,1024,0.023983110984166462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,1536,0.031574222776624895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,5120,0.15983911355336508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,1536,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,10240,0.065138664510515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,5120,0.0520613325966729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,1536,0.008839999636014303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,64,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,10240,0.021943999661339655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,6144,0.065010666847229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,12288,0.16390844186147055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,10240,0.03945422172546387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,3584,0.5842782126532661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,128,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,12288,0.020083554916911654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,512,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,2560,0.010068444742096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,3584,0.09581600295172797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,2048,0.08720444308386908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,7168,0.024064888556798298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,7168,0.018262222409248352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,1024,0.0288355549176534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,5120,0.03208177619510227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,10240,0.06401333543989393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,10240,0.37676088015238446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,1536,0.025967111190160114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,512,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,768,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,8192,0.00980711148844825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,32,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,64,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,10240,0.03589422172970242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,8192,0.14498310618930393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,16384,0.055504000849194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,512,0.007479999628331926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,256,0.0029404444826973807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,1536,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,7168,0.32507022221883136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,10240,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,65536,0.5850746896531847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,1536,0.011468444433477191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,12288,0.020246222615242004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,64,0.0131217779384719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,3072,0.044328000810411244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,65536,0.06276088953018188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,128,0.00885511106914944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,5120,0.029667556285858154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,512,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,4096,0.10177955362531875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,12288,0.030398223135206435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,512,0.018897778458065458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,2048,0.018400889303949144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,32,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,128,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,5120,0.7904862297905816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,7168,0.017439999514155917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,10240,0.06920266813702054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,4096,0.013728000223636627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,64,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,3584,0.10451377762688531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,65536,0.145087997118632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,3584,0.2939608891805013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,3072,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,1024,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,128,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,512,0.006540444576077991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,3584,0.011551111108726926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,768,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,12288,0.01687733332316081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,512,0.02717599935001797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,2560,0.02092355489730835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,1024,0.008382221890820397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,6144,0.029639111624823675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,51200,0.19312089019351533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,2560,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,1536,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,1536,0.01975288821591271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,1536,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,3584,0.04205244448449877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,16384,0.03564711080657111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,128,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,64,0.005655999812814925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,768,0.015936889582210116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,65536,0.10577244228786892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,16384,0.04135555691189236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,51200,0.13113689422607422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,1024,0.04092177748680115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,1536,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,1536,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,5120,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,7168,0.471727106306288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,32,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,5120,0.0517759985393948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,256,0.02498666610982683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,32,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,512,0.011462222370836469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,2048,0.08277866575453016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,768,0.04279822111129761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,12288,0.11561155319213867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,6144,0.11485689216189915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,768,0.03650310966703627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,3072,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,1536,0.02555911077393426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,128,0.045384890503353544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,7168,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,128,0.0027013334135214486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,51200,0.05468711256980896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,3584,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,16384,0.559716436598036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,512,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,768,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,1024,0.007032000356250339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,65536,0.08937422434488933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,16384,0.24549243185255262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,65536,0.07932888799243502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,3072,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,6144,0.01891999940077464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,64,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,256,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,7168,0.022800889280107286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,16384,0.019667555888493855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,768,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,5120,0.08676266670227051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,2048,0.12115022871229385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,2048,0.01629955569903056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,1024,0.0125591109196345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,768,0.018464000688658822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,4096,0.020815999971495736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,1024,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,10240,0.03429422113630507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,512,0.019706666469573975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,256,0.008720888859695857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,12288,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,3584,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,4096,0.3255964385138618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,128,0.0568373335732354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,3072,0.021895110607147217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,2048,0.01955111159218682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,512,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,7168,0.3086266782548693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,128,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,6144,0.03455466694302029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,2048,0.04263822237650553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,2560,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,10240,0.013283555706342062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,3072,0.014910222755538093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,12288,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,4096,0.010440888504187265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,256,0.0052977779673205475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,128,0.020503999458418954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,6144,0.17412977748446992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,16384,0.18292000558641222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,16384,0.05609244770473904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,16384,0.01887999971707662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,1536,0.013811555173661975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,768,0.08411022027333577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,16384,0.0547617773214976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,32,0.07223111391067505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,6144,0.08565511306126912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,7168,0.014456889695591398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,256,0.11653155750698513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,6144,0.012113778127564324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,768,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,8192,0.02641244398223029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,51200,0.04881511131922404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,4096,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,1024,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,16384,0.04491377870241801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,3584,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,4096,0.007919111185603673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,256,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,16384,0.025451555848121643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,10240,0.06746221913231744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,16384,0.19035289022657606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,16384,1.143789291381836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,3072,0.03844711184501648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,768,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,4096,0.011495999991893768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,12288,0.0729457802242703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,2048,0.015947555502255756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,3072,0.01332444449265798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,256,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,2048,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,8192,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,7168,0.01760088900725047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,3072,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,12288,0.018108444081412423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,512,0.010470222267839642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,51200,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,512,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,128,0.017851556340853374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,64,0.1769040028254191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,4096,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,51200,0.763252470228407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,4096,0.04208977686034309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,1024,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,128,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,6144,0.238916450076633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,256,0.0295466681321462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,128,0.0824142230881585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,5120,0.059644447432623975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,6144,0.10493244727452596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,2560,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,256,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,256,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,7168,0.021613332960340712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,3072,0.01661066710948944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,128,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,2560,0.29828090137905544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,32,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,128,0.07878400219811334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,2048,0.011558222273985544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,16384,0.02991644541422526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,128,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,12288,0.18264355924394396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,16384,0.08199110958311293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,6144,0.15582133664025202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,1536,0.0667475528187222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,8192,0.01349511080318027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,5120,0.04928266670968798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,6144,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,7168,0.02405066622628106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,65536,0.06495555241902669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,51200,1.901665793524848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,51200,0.05321244398752848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,51200,0.10997066895167033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,6144,0.2998239994049072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,8192,0.036144001616372004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,8192,0.03178755442301432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,8192,0.12624800205230713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,4096,0.016031111280123394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,65536,2.3089671664767795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,12288,0.15185511112213135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,10240,0.03537333342764113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,768,0.011434666812419891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,12288,0.985362688700358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,10240,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,3072,0.009467555416954888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,3072,0.01962577799956004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,65536,0.947120878431532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,16384,0.05282666616969638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,12288,0.03634044528007507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,1536,0.2522515455881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,1024,0.0318248901102278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,51200,0.8502186669243706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,5120,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,5120,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,4096,0.011620444556077322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,12288,0.2589297824435764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,12288,0.02962933315171136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,32,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,256,0.02847822176085578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,8192,0.1597199969821506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,6144,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,768,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,51200,0.4316906664106581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,2560,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,7168,0.0546248886320326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,7168,0.012554667062229581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,16384,0.09586222304238214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,16384,0.03472444415092468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,10240,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,512,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,32,0.019165333774354722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,6144,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,512,0.08154222038057116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,65536,0.5846844249301487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,16384,0.04485422372817993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,3584,0.07995822032292683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,512,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,51200,0.15420267317030165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,10240,0.016474665866957772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,6144,0.04839377933078342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,16384,0.04605155520968967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,7168,0.30668799082438153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,512,0.0027991111079851785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,16384,0.0814488861295912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,7168,0.028024888700909082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,2048,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,65536,2.1490372551812067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,512,0.016138666205936007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,10240,0.011848889291286469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,1536,0.0038204445607132385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,51200,0.2949039936065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,8192,0.14266222053103977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,7168,0.024522667129834492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,6144,0.07212977939181857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,16384,0.598907576666938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,6144,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,768,0.0031537777847713898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,1024,0.009890666438473595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,768,0.01144977741771274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,8192,0.5152533319261339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,16384,0.10142044226328532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,32,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,1536,0.025084444218211707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,1024,0.00960888879166709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,8192,0.014305777019924588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,6144,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,4096,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,3584,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,256,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,768,0.021684444612926904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,2560,0.040101332796944514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,2560,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,5120,0.1264302200741238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,128,0.06809777683681913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,5120,0.0122097780307134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,8192,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,128,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,1024,0.025729777084456548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,65536,0.1131333377626207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,8192,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,6144,0.0376506679587894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,6144,0.009276444713274637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,3584,0.1013093325826857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,6144,0.03089511063363817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,1024,0.029101332028706867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,5120,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,512,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,768,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,10240,0.434879117541843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,512,0.007442666424645319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,3584,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,1024,0.012678222523795234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,51200,0.2573244571685791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,8192,0.03706577751371596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,8192,0.04144710964626736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,512,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,3584,0.02198222279548645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,10240,0.05029866761631436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,7168,0.023696889479955036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,12288,0.05168177684148153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,8192,0.4996239874098036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,3072,0.11232354905870225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,256,0.026748443643252056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,768,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,768,0.018962666392326355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,64,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,10240,0.03893333342340257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,128,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,10240,0.025316443708207872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,1024,0.009310222334331935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,7168,0.06115288866890801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,32,0.07869422435760498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,51200,0.04946666624810961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,3072,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,512,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,1536,0.03373422225316366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,32,0.040992889139387347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,512,0.012810666528013019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,4096,0.01147377739350001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,1024,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,7168,0.01965688831276364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,7168,0.0337084432442983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,65536,0.060460441642337374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,10240,0.02700622214211358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,12288,0.01676177812947167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,3072,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,4096,0.1118675602806939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,2560,0.027506666051016912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,2048,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,10240,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,1024,0.006340444501903322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,51200,3.272728814019097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,256,0.013021333350075616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,8192,0.020968889196713764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,12288,0.06128444274266561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,3072,0.013316444224781461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,256,0.022317333353890315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,64,0.024732444021436904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,16384,0.6626204384697808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,7168,0.05385422375467089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,51200,0.14087288909488255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,5120,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,6144,0.05738133192062378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,4096,0.014318222800890604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,6144,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,8192,0.05668089124891493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,1024,0.03176177872551812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,51200,0.050599998897976346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,5120,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,1024,0.00812711069981257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,3072,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,10240,0.0638399985101488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,65536,0.06105244159698486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,2560,0.026331555512216356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,51200,0.023036445180575054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,1536,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,10240,0.33021333482530385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,10240,0.11137244436475967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,4096,0.02401777770784166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,65536,0.5155875417921278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,768,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,32,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,768,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,51200,6.482343885633681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,1536,0.209824005762736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,128,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,7168,0.05366222063700358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,65536,0.15762489371829563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,3584,0.15505422486199275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,10240,0.1340293354458279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,2048,0.036201778385374285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,12288,0.031579554080963135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,7168,0.20865955617692736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,3584,0.01421688828203413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,7168,0.03247822324434916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,3584,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,2560,0.030010667112138536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,32,0.011760888828171624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,10240,1.5227368672688801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,51200,0.46123112572564023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,4096,0.8737475607130262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,2560,0.010311110979980892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,65536,0.8802515665690104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,51200,0.4397431214650472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,4096,0.0513555539978875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,128,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,10240,0.08557244141896565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,8192,0.33826756477355957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,65536,0.1876008907953898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,1024,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,1024,0.01440533333354526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,512,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,7168,0.12182044982910156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,10240,0.033193777004877724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,65536,0.5060710906982422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,4096,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,3584,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,51200,0.07230577866236369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,2560,0.0068853331936730286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,3584,0.008052444292439355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,512,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,51200,0.06202666627036201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,12288,0.9555688434176974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,7168,0.08426755666732788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,2048,0.042302221059799194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,12288,0.11019111341900295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,3584,0.02012355625629425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,2560,0.012695111334323883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,16384,0.376909335454305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,3072,0.39526666535271543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,1536,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,3584,0.011689778003427716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,4096,0.02737688852681054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,65536,1.1267031563652885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,51200,2.0461288028293185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,1536,0.019271110494931538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,12288,0.1267813311682807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,512,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,256,0.0022293333378103045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,6144,0.07483822107315063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,2048,0.11212621794806586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,8192,0.37599910630120176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,5120,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,256,0.0022479999396536085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,1024,0.013839999834696451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,8192,1.3001101811726887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,32,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,10240,0.10957244369718765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,4096,0.11686933040618896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,1536,0.004855111241340637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,6144,0.009422221945391761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,2048,0.02387111054526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,51200,0.3472897741529677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,3072,0.05051999953058031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,65536,1.6528231302897136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,32,0.012260444462299347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,8192,2.1148622300889754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,10240,0.6068319744533962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,3072,0.07337155607011583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,512,0.004544888933499654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,768,0.012646222280131446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,1536,0.038351111941867404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,1024,0.018594667315483093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,5120,0.13099022706349692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,16384,13.016695658365885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,6144,0.030717333157857258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,768,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,128,0.08867733346091376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,3072,0.025754666990704004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,12288,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,3584,0.10539111163881089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,65536,1.4113403956095378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,2560,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,10240,0.1091466678513421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,10240,0.016221332881185744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,6144,0.01183644433816274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,64,0.14476799964904785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,3072,0.024825778272416856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,256,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,10240,0.08675822284486558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,5120,0.012892444100644855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,2048,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,10240,0.042929778496424355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,12288,0.08842221895853679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,5120,0.5042853355407715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,51200,0.05978577666812473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,4096,0.010647999743620554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,128,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,64,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,512,0.0705359975496928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,1536,0.01463822192615933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,51200,0.07095377975040011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,16384,0.32045067681206596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,256,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,32,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,1536,0.01685333251953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,3584,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,2048,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,4096,0.016120889120631747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,16384,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,128,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,5120,0.035087111923429705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,3072,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,2560,0.019275556008021038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,3072,0.009547555612193214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,2560,0.018280888597170513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,3584,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,768,0.004528000122971005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,6144,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,4096,0.023649776975313824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,10240,0.08466488785213894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,5120,0.27891021304660374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,12288,0.019309333629078336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,1024,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,16384,0.03441333439615037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,10240,0.021547555923461914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,1024,0.012028444144460889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,12288,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,65536,0.5845564206441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,6144,0.017867555220921833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,4096,0.01771022213829888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,768,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,1024,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,51200,0.4178195529513889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,65536,0.11078311337365045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,10240,0.03530933459599813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,64,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,1536,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,65536,4.0998658074273004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,16384,0.23046043184068468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,6144,0.09349066681332058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,1024,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,512,0.016933333542611863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,12288,0.008749333520730337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,6144,0.055184000068240695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,2560,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,1536,0.09913155767652725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,512,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,6144,0.04811288913091024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,3072,0.36118221282958984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,51200,0.09542489051818848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,256,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,128,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,6144,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,2048,0.0373582210805681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,256,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,8192,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,3584,0.016137777103318107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,65536,0.07875555753707886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,32,0.010523555179437002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,7168,0.010120888551076254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,512,0.008365333080291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,256,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,4096,0.03791911072201199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,2560,0.04602755440606011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,2560,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,7168,0.015289778510729471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,3072,0.021588444709777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,128,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,8192,0.023337778117921617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,1536,0.21219110488891602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,2560,0.024018666810459558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,64,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,6144,0.030553778012593586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,3072,0.011184000306659274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,2560,0.02488800055450863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,6144,0.03912355502446493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,3072,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,8192,0.14021244313981798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,7168,0.038093331787321306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,1024,0.01648888819747501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,5120,0.1913155582216051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,128,0.004179555508825514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,7168,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,16384,0.020870221985710993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,10240,0.02701955537001292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,32,0.0017893332988023758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,512,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,128,0.0019191110299693213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,3584,0.061583108372158475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,1536,0.03727111220359802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,6144,0.06587022542953491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,512,0.017733333839310538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,2560,0.0914471083217197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,16384,0.048060443666246205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,512,0.0031537777847713898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,3072,0.018888889087571036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,10240,0.018062222335073683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,64,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,256,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,768,0.002672000063790215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,2560,0.0722915530204773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,256,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,16384,0.0221377776728736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,5120,0.06668000088797675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,7168,2.300932354397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,16384,0.047206223011016846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,3584,0.012558221817016602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,10240,0.08129422532187568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,2048,0.057017776701185435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,10240,0.04835199978616503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,8192,0.021352888809310064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,2048,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,16384,0.022295110755496558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,256,0.024077332682079736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,12288,0.21187021997239855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,3072,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,2560,0.04961511161592272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,512,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,4096,0.01752622259987725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,512,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,51200,0.14420444435543484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,7168,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,12288,0.2509857813517253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,12288,0.3373013337453206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,51200,0.213774217499627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,65536,0.2627680036756727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,8192,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,8192,0.009188444250159794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,5120,0.005378666851255629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,2048,0.02604533400800493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,16384,0.08637955453660752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,5120,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,32,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,10240,0.3332551055484348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,768,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,3584,0.020709334148301017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,3072,0.1338675551944309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,3584,0.009406222237481011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,7168,0.36652623282538516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,768,0.017873777283562552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,64,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,65536,0.08857155508465236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,16384,0.08464888731638591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,4096,0.03162489003605313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,6144,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,5120,1.827194637722439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,4096,0.09960799747043186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,512,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,128,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,7168,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,512,0.018798222144444782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,10240,0.11236088805728489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,3072,0.007196444604131911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,16384,0.016585777203241985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,2560,0.007481777833567725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,10240,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,128,0.01979200045267741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,6144,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,1536,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,6144,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,6144,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,16384,0.038189331690470375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,768,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,1536,0.024863999750879075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,7168,0.11381155914730495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,32,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,256,0.006657777975002925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,512,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,8192,0.013187555803192986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,4096,0.014904000692897372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,6144,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,65536,0.2604728804694282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,1536,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,1536,0.0038737778862317405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,32,0.00453955555955569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,16384,0.05070666803254021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,3584,0.02473066747188568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,7168,0.013735999663670858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,7168,0.024322667055659827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,128,0.04004089037577311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,256,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,3584,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,128,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,10240,0.01640888883007897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,8192,0.07647377914852567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,65536,1.2382213804456923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,4096,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,2048,0.06380355358123779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,51200,1.1672275331285265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,512,0.06894489129384358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,2048,0.013174222575293647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,10240,0.3787093427446153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,10240,0.03239466746648153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,3072,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,512,0.014497778481907315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,10240,0.01610933409796821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,51200,0.13246222337086996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,1024,0.029775109555986192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,3072,0.08533155255847508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,768,0.0329440004295773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,1536,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,4096,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,1024,0.09968711270226373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,8192,0.037323554356892906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,8192,0.023356444305843774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,256,0.008529778156015608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,512,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,256,0.029164443413416546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,5120,0.03087377879354689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,3584,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,2560,0.10025599930021499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,10240,0.02700800034734938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,3584,0.34703734185960555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,64,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,1024,0.08338044749365912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,1024,0.02854222059249878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,65536,5.251841651068793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,128,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,6144,0.01941066649225023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,51200,20.065543280707466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,4096,0.04808888832728068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,128,0.010778666370444827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,10240,0.020264888803164165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,2560,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,16384,0.9398782518174914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,2560,0.03015466531117757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,6144,0.40225866105821395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,4096,0.018276444739765592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,128,0.03962755534383986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,128,0.010424888796276517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,1536,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,65536,0.17815732955932617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,1536,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,3584,0.07229244709014893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,65536,0.04666755596796671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,7168,0.04842755529615614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,8192,0.03251466817326016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,768,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,12288,0.028549333413441975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,3072,0.01555999947918786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,768,0.13470221890343562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,51200,0.05135644475618998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,1024,0.010064000056849586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,6144,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,1536,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,6144,0.019018666611777413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,768,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,64,0.009288888838556077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,5120,1.053510242038303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,768,0.036836445331573486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,2048,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,2048,0.009736888938479954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,16384,0.05097511079576281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,32,0.010017777482668558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,3072,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,1536,0.04888088835610283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,2560,0.04481688804096646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,1536,0.337768899069892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,1024,0.021054221524132624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,7168,0.012852444416946836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,3072,0.032560888263914324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,64,0.6057573424445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,256,0.047188444270028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,2048,0.037328001525666975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,8192,0.021910222040282354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,1024,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,3584,0.01447111037042406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,1024,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,51200,0.1514399978849623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,256,0.030790223015679255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,512,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,512,0.0041288890772395665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,1536,0.008928889201747047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,12288,0.10302755567762588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,2560,0.008384889198674096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,3072,0.019736889335844252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,2560,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,768,0.035586665074030556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,768,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,2560,0.02679288884003957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,512,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,4096,0.3129173384772407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,5120,0.024550222688251074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,6144,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,2560,0.10011021958457099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,12288,0.034436444441477455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,6144,0.09516355726453994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,51200,0.04835022158092923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,768,0.015211555692884656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,2048,0.004502222355869082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,7168,0.036307556761635676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,4096,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,65536,0.5487333403693305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,51200,0.1045759982532925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,1024,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,256,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,256,0.011136000355084738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,51200,8.722300211588541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,4096,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,2560,0.10626666413413154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,10240,0.6570737626817491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,7168,0.01824977828396691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,6144,0.12450133429633246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,5120,0.02293155590693156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,3072,0.03186310993300544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,8192,0.027849776877297297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,7168,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,1536,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,128,0.03193599979082743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,3072,0.013777777552604675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,10240,0.02640266716480255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,10240,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,5120,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,3584,0.014949333336618213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,8192,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,8192,1.4485715230305989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,3584,0.08408800098631118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,1024,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,10240,0.07207289006974962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,3584,0.012711111042234631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,256,0.0255422227912479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,7168,0.04234844446182251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,65536,0.062401778168148465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,3584,0.02424888809521993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,64,0.01070577816830741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,64,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,2560,0.034520887666278414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,65536,0.0575795571009318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,7168,0.0961715579032898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,12288,0.029699554045995075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,6144,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,2560,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,256,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,512,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,1024,0.48414044910007054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,1024,0.007928888830873702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,5120,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,128,0.008364444805516137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,256,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,32,0.006545777950021956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,2560,0.009610666996902889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,3584,0.022797332869635686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,10240,0.016932444439993966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,2560,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,1024,0.02880711025661892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,8192,0.012757333616415659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,6144,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,4096,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,3072,0.04033599959479438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,7168,0.0351893338892195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,6144,0.044063998593224414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,256,0.016872889465755887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,10240,0.07912800047132704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,10240,0.16333777374691433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,768,0.017255999975734286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,3584,0.010141332944234213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,128,0.0019404445257451797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,16384,0.025010666913456384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,512,0.0421048899491628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,6144,0.00942755573325687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,4096,0.03231111168861389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,10240,0.07340088817808363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,1536,0.029487109846538965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,32,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,3584,0.02134311033619775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,4096,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,3584,0.020615999897321064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,512,0.022103110949198406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,16384,0.04742133286264208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,4096,1.3202151192559135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,5120,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,1024,0.054633776346842446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,5120,0.04285600119166904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,1536,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,3584,0.979582256740994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,65536,14.880311754014757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,16384,0.049976888630125255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,1024,0.006684444430801604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,256,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,768,0.006795555353164673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,6144,0.03456799851523505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,1536,0.07990666892793444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,6144,0.0763182242711385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,64,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,128,0.002459555657373534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,7168,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,3584,0.016358221570650738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,2048,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,51200,0.05596533086564806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,128,0.00638488887084855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,7168,0.012813333008024426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,10240,0.022835556003782485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,4096,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,7168,0.10736088620291816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,5120,0.015023110641373528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,128,0.030093332131703694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,10240,0.04391022192107307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,65536,1.2285049226548936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,1024,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,3072,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,6144,0.007496889266702864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,12288,0.021219554874632094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,65536,0.05569510989718967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,6144,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,3584,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,5120,0.025432889660199482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,7168,0.012691555751694573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,2560,0.021723555194007024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,8192,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,4096,0.053809778557883374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,128,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,8192,0.01570933394961887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,128,0.05483999848365784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,12288,0.06477333439721002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,1536,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,128,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,12288,0.1908106671439277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,3072,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,3072,0.007435555259386699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,51200,0.10013333294126724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,65536,0.31394489606221515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,2048,0.022908444205919903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,6144,0.052312887377209134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,51200,0.27313600646124947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,10240,0.18741599718729654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,128,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,1024,0.02239911092652215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,2560,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,512,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,3584,0.018013333280881245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,2560,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,768,0.01889600025282966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,6144,1.278982268439399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,1024,0.023839111129442852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,6144,0.011672000090281168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,12288,0.04712444543838501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,51200,0.05570399761199951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,10240,0.8909644020928277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,8192,0.03382666574584113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,1024,0.015691555208630033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,16384,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,256,0.02400622268517812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,512,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,2048,0.05817777580685086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,1024,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,16384,0.02885155545340644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,3584,0.006321777900060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,3072,0.026912000444200303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,4096,0.018363555272420246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,3584,0.061273780133989125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,5120,0.027261333333121404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,5120,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,6144,0.017010667257838778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,12288,0.09454222520192464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,512,0.004439110971159405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,10240,0.03630222214592828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,7168,0.02720977697107527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,1536,0.014988443917698331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,1536,0.012607110871209038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,51200,0.07305510838826497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,8192,0.01832355558872223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,3072,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,256,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,7168,0.013391999734772576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,128,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,16384,0.044385777579413525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,8192,0.023818666736284893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,16384,0.04823644293679131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,1024,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,3072,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,3584,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,768,0.2269831233554416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,768,0.01793244481086731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,768,0.022678222921159532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,65536,8.816211276584202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,3072,0.13453955120510525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,1536,0.02912799848450555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,7168,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,5120,0.11556000179714626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,2560,0.0935351120101081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,2560,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,512,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,1024,0.015531554818153381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,4096,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,7168,0.14474133650461832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,2560,0.013813333378897773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,12288,0.016582222448454965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,128,0.005428444594144821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,51200,3.4958676232231984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,768,0.26239111688401967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,10240,0.17814932929144967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,256,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,2560,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,6144,0.013213333156373767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,2048,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,768,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,4096,0.04899555444717407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,51200,0.04577777783075968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,1024,0.09180444478988647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,8192,0.02643022272321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,768,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,128,0.014953777194023132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,4096,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,16384,0.05196088883611891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,64,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,4096,0.014109333356221518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,64,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,12288,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,768,0.036559998989105225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,10240,1.9635431501600478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,65536,0.05135822296142578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,2560,0.08831822209888035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,12288,0.21010133955213758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,16384,0.011005333728260465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,1024,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,512,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,12288,0.06864533159467909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,10240,0.023269333773189124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,7168,0.036312000619040594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,768,0.029261334074868098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,2048,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,512,0.04079466727044847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,8192,0.5011733373006185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,3072,0.052022222015592784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,3072,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,4096,0.01443911095460256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,3584,0.015159999330838522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,3072,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,3072,0.019588443968031142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,12288,0.018206222189797294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,256,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,6144,0.006334222025341458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,3072,0.11756443977355957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,512,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,65536,0.017128888103697035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,64,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,2560,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,768,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,7168,0.020371556282043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,7168,0.027235555979940627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,768,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,12288,0.061836441357930504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,2048,0.01791733337773217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,12288,0.059317330519358315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,3584,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,128,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,8192,0.05707910988065931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,6144,0.11115554968516032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,65536,0.08340266678068374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,65536,1.0995022455851238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,1024,0.012771555946932899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,2560,0.11680800384945339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,768,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,8192,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,1024,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,128,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,512,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,768,0.00305244450767835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,7168,0.03510311245918274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,256,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,3072,0.035573333501815796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,10240,0.2574168841044108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,1024,0.02219466699494256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,3584,0.07526399691899617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,3072,0.015407111909654407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,6144,0.02295644415749444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,16384,0.013768889009952545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,512,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,6144,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,4096,0.021202666891945735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,51200,0.05482133229573568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,51200,0.2118613322575887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,256,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,64,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,1024,0.0038515557017591265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,10240,0.027294221851560805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,5120,0.014955555399258932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,256,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,128,0.01090133355723487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,5120,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,10240,4.392164442274305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,6144,0.34677778349982363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,7168,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,3072,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,5120,0.034811556339263916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,51200,0.3918595578935411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,3072,0.15361066659291586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,512,0.030191110240088567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,3072,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,2048,0.010433778166770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,51200,0.02828177809715271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,6144,0.09220888879564072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,256,0.01957155598534478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,512,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,256,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,32,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,4096,0.024255111813545227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,6144,0.05046933227115207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,3584,0.01421688828203413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,3584,0.008753777378135258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,12288,0.024540444215138752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,6144,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,768,0.014839111102951897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,16384,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,3584,0.0649048884709676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,3072,0.3258559968736437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,3072,0.04275288846757677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,51200,0.39564355214436847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,256,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,3584,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,256,0.011669333610269757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,51200,0.06742578082614474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,2560,0.020452444752057392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,7168,0.01440711153878106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,64,0.00546488869521353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,3072,0.010581333604123859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,8192,0.045531554354561694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,5120,0.12172355916765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,2048,0.006773333168692059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,3584,0.07074666685528226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,64,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,16384,0.19707466496361625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,512,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,1536,0.009720000127951304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,12288,0.13981600602467856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,7168,0.04076888826158311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,3072,0.09040177530712551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,51200,0.3584249019622803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,4096,0.009688888986905416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,16384,0.061144888401031494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,7168,0.03949600127008226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,2560,0.006776000062624614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,256,0.03832088907559713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,5120,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,2048,0.17558666070302328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,1024,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,8192,0.0355679988861084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,5120,0.009782222410043081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,10240,0.012944888737466602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,16384,0.07704622215694852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,12288,0.15557866626315647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,3072,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,2560,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,6144,0.05575288666619194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,12288,0.041909333732393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,512,0.013838222457302941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,4096,0.016237333416938782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,1536,0.02905422117975023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,51200,0.07730133003658719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,4096,0.019578667150603402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,65536,1.2813164393107097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,2560,1.626829359266493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,512,0.028897778855429754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,10240,0.1529946724573771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,16384,0.15202666653527155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,4096,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,51200,1.9567351871066625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,51200,0.12294399738311768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,51200,0.18932622008853486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,1024,0.03605866763326857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,64,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,51200,0.14741510815090603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,7168,0.2477386792500814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,65536,0.2610151237911648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,10240,0.06027911106745402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,2048,0.012757333616415659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,16384,0.6048862139383951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,768,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,51200,0.07222044467926025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,2560,0.02885955572128296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,512,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,65536,1.624934196472168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,5120,0.17558222346835664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,10240,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,1536,0.09558400180604722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,5120,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,512,0.17485333813561332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,65536,0.15499644809299046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,3584,0.2163635492324829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,51200,0.12796266873677573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,65536,0.2540515528784858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,4096,0.009249777429633671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,51200,0.08059466547436185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,6144,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,2560,0.008237332933478886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,128,0.0035031110876136353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,3072,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,8192,0.21520888805389404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,2560,0.028195556667115953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,7168,0.019567999574873183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,1536,0.015996444556448195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,3072,0.016551110479566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,1024,0.006832888970772426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,64,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,51200,0.0594008896085951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,8192,0.45888619952731663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,5120,0.06350488795174493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,7168,0.8364124298095703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,4096,0.034104890293545194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,4096,0.022408000297016565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,768,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,10240,0.19414577219221327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,2560,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,32,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,3584,0.08193155792024401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,2048,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,1536,0.005998222364319696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,5120,1.3139973746405709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,768,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,6144,0.07437422540452746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,64,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,5120,0.1451546616024441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,10240,0.27512089411417645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,65536,0.09894044531716241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,7168,0.10210755798551772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,7168,0.8294284608629016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,10240,0.01828977796766493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,4096,0.02574933403068119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,16384,0.04376088910632663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,32,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,512,0.006491555521885554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,4096,0.010624888870451184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,3072,0.017865777015686035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,2048,0.006031999985376994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,1024,0.02641155487961239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,128,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,64,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,65536,108.74864366319444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,1536,0.020742222666740417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,128,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,4096,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,51200,0.02606133288807339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,6144,0.03345422281159295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,256,0.02198933396074507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,4096,0.02072355482313368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,12288,0.12719733185238308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,5120,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,8192,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,2048,0.010689777632554373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,768,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,64,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,7168,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,2048,0.02948533495267232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,65536,0.06357955270343356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,3584,0.025145777397685583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,7168,0.0488337775071462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,4096,0.02624711063173082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,51200,0.33201244142320424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,256,0.010436444646782346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,16384,0.04309244288338555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,2560,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,32,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,1024,0.026890666948424444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,12288,0.02898755669593811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,51200,0.43937333424886066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,12288,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,64,0.005715555614895291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,12288,0.1937626732720269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,7168,0.04301688737339444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,6144,0.03633066680696275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,65536,0.13533333937327066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,768,0.028674665424558852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,1536,0.01866311166021559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,6144,0.017218665944205392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,256,0.015450666348139444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,5120,0.04790133237838745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,16384,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,16384,0.08301511075761583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,1024,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,256,0.006653333289755716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,768,0.2650115489959717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,16384,0.07448266612158881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,5120,0.0351582235760159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,2048,0.04199555516242981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,2560,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,10240,0.03885689046647813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,6144,0.03306844499376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,256,0.005312888986534542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,51200,0.061556445227728955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,512,0.053138666682773165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,2560,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,1536,0.10695377985636394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,10240,0.04457599918047587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,51200,7.2855682373046875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,64,0.018974221414989896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,5120,0.03147733211517334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,64,0.005772444523043103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,7168,0.24988444646199545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,10240,0.046563555796941124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,7168,0.12445778316921657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,3072,0.031903111272388034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,12288,0.05420355664359199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,1536,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,3584,0.007452444069915348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,3584,0.019685332973798115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,2048,0.06006133556365967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,51200,2.8279031117757163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,7168,0.03052622079849243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,5120,0.01865688959757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,4096,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,4096,0.016619554824299283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,3584,0.02409422232045068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,4096,0.020920889245139226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,1024,0.003607999947335985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,128,0.015225778023401896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,128,0.014704888065656027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,7168,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,512,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,65536,0.1501057810253567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,2560,0.018138666947682697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,6144,0.09536533223258124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,2560,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,65536,0.06400799751281738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,2048,0.02535466684235467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,32,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,512,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,7168,2.8003724416097007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,6144,0.22094755702548555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,1024,0.066484444671207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,16384,0.16115021705627441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,12288,0.0232924438185162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,4096,0.0603333314259847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,2048,0.029896888468000624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,5120,0.010207110808955299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,8192,0.013391111459996967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,128,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,1536,0.016703110602166917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,7168,0.05179199907514784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,8192,0.02440711193614536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,5120,0.03542666633923849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,51200,0.08837333652708267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,12288,0.19845333364274767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,64,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,768,0.015283554792404175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,2048,0.03619822197490268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,7168,0.06443111101786296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,512,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,4096,0.018423999349276226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,10240,0.09081777599122788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,51200,0.050980445411470204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,4096,0.10228444470299615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,128,0.017955554856194388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,2048,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,256,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,3072,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,256,0.015954666667514376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,65536,0.08763911326726277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,7168,0.3276382287343343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,12288,0.05183644427193535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,12288,0.5599199930826823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,16384,0.09099466933144464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,768,0.023061333431137934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,12288,0.011483555866612328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,3072,0.017324444320466783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,512,0.007051555646790399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,2048,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,65536,11.456782023111979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,6144,0.044884443283081055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,3584,0.010402667025725046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,5120,0.009588444398509132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,10240,0.1564799944559733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,128,0.00998933364947637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,128,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,8192,0.03462488783730401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,12288,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,12288,0.1349617772632175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,6144,0.022980444961124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,128,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,2048,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,10240,0.10364888774024116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,3072,0.014023999373118082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,768,0.004824000100294749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,1024,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,2560,0.01902844508488973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,64,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,51200,0.14981244670020208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,6144,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,12288,0.33727288246154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,5120,0.042098667886522084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,768,0.024753777517212763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,5120,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,2048,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,1024,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,256,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,65536,0.06402222315470378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,16384,0.02269866731431749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,1536,0.0123511114054256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,51200,1.8226212395562067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,128,0.02590311070283254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,8192,1.1749600304497612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,6144,0.02668711046377818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,128,0.0041866666740841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,4096,0.007900444169839224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,51200,0.12327288256751166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,128,0.005708444449636672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,512,0.07433333661821154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,12288,0.15711732705434164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,10240,0.05100533366203308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,128,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,4096,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,3072,0.021871111459202234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,12288,0.037520888778898455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,3584,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,12288,0.027058665951093037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,2048,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,2560,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,3584,0.021513778302404616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,1024,0.022542221678627863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,512,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,5120,0.02306933369901445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,2048,0.029489778810077246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,5120,0.018631110588709515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,2048,0.07495022482342191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,2048,0.05643288956748115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,3584,0.026521777113278706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,2560,0.07946399847666423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,64,0.0017991111510329777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,2048,0.0127368892232577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,5120,0.017684444785118103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,256,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,3072,0.30414043532477486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,8192,0.6756275494893392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,7168,0.5725137922498915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,8192,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,10240,0.03622488843070136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,768,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,2560,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,12288,0.05077777637375725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,12288,0.02084355552991231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,512,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,1024,0.008919110728634728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,10240,0.29836710294087726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,12288,0.04376799861590067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,16384,0.27973421414693195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,3072,0.020104888412687514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,5120,0.02792177928818597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,4096,0.006676444576846228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,512,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,6144,0.028565333949195013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,5120,0.09228710995780097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,3584,0.04575733343760172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,6144,0.02735022207101186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,7168,0.12895822525024414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,8192,0.029785778787400987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,2048,0.020091555184788175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,12288,0.044109334548314415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,7168,0.08128088712692261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,8192,0.02369511127471924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,51200,0.4980586899651422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,256,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,3072,0.26030132505628795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,8192,0.03379999929004245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,1024,0.06043733490837944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,512,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,16384,0.029063112205929224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,512,0.043783111704720386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,16384,0.017140444782045152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,8192,0.07246844636069404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,768,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,65536,0.1928337812423706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,3584,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,5120,0.03057066599527995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,2048,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,8192,0.04119022356139289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,3072,0.03210044569439358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,768,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,51200,0.07336178090837267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,2048,0.016143111719025504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,128,0.01576622161600325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,1024,0.006017777654859755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,10240,0.050055109792285495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,16384,0.12391555309295654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,3072,0.06929777728186713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,1536,0.007928888830873702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,65536,0.12132267157236735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,12288,0.0843048890431722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,51200,0.07112799750434028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,8192,0.015090665883488126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,16384,0.4663093354966905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,3072,0.13422132862938776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,768,0.027629334065649245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,8192,0.19659554958343506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,16384,0.35900176895989316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,4096,1.0246071285671658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,10240,0.04265955421659681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,7168,0.06206399864620633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,32,0.14496621820661756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,5120,0.02167555524243249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,5120,0.02234222160445319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,3584,0.022249778111775715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,2048,0.10263644324408637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,1536,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,256,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,3072,0.6789031028747559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,1024,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,8192,0.11369600560930039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,16384,0.0163093325164583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,2560,0.023716444770495098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,16384,0.42309776941935223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,10240,0.05601244502597385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,65536,0.08660444286134507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,7168,0.01183999992079205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,128,0.0022337778160969415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,16384,1.3861849043104384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,16384,0.2177333301968045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,768,0.015964445140626695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,2560,0.3185057904985216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,1536,0.02437688906987508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,128,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,4096,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,4096,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,128,0.014693333042992486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,6144,0.057032889790005155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,3072,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,1536,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,16384,1.007079972161187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,256,0.016933333542611863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,64,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,2560,0.008762666748629676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,6144,0.13359110885196263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,16384,0.032235556178622775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,16384,0.09386310974756877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,1024,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,64,0.03637333379851447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,16384,0.15575733449723986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,768,0.2524622281392415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,51200,1.7020204332139757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,4096,0.012690666649076672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,2560,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,128,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,3072,0.7923537890116373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,7168,0.08394844664467706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,1024,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,6144,0.013632000320487551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,2560,0.064337776766883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,3584,0.010873777998818291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,2048,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,128,0.03843644592497084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,2048,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,10240,0.04907466636763679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,768,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,7168,0.019689778486887615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,16384,0.062317331631978355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,10240,0.012422222230169507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,7168,0.028082665469911363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,3072,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,1024,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,16384,1.1118586858113606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,3072,0.019732443822754752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,128,0.020607110526826646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,8192,1.260911093817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,2048,0.17645955085754395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,1536,0.023888889286253188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,7168,0.034155554241604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,1536,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,32,0.00552977787123786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,65536,0.7306097878350152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,2560,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,2048,0.008610666626029545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,6144,0.13309066825442845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,51200,0.0751404431131151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,256,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,256,0.14694310559166804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,12288,0.03677333394686381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,10240,0.880518224504259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,3072,0.01293688929743237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,7168,0.03362133436732822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,65536,0.7378666665818957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,256,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,6144,0.04007555709944831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,2560,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,3584,0.2895653247833252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,3072,0.02570488883389367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,10240,0.4516426722208659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,7168,0.017596445149845548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,32,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,1536,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,3584,0.01756533318095737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,4096,0.046610666645897754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,65536,0.8404373592800565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,2560,0.017224889662530687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,64,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,1024,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,2560,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,4096,0.021903110875023737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,128,0.00903200027015474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,1536,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,4096,0.023709333605236475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,3584,0.6358675426907009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,3584,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,2048,0.012320888539155325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,51200,0.19071643882327607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,768,0.003642666671011183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,5120,0.029618667231665716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,5120,0.01351377781894472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,51200,0.30967821015252006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,6144,0.02882844540807936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,5120,0.051071110698911876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,2560,0.03242844343185425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,8192,0.30308622784084743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,768,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,6144,0.04462933209207323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,51200,0.016720000240537856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,7168,0.02425066630045573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,32,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,1536,0.145724442270067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,16384,0.018904000520706177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,768,0.006895111252864202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,256,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,1536,0.004966222163703707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,3072,0.0306640002462599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,16384,0.9786471260918511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,51200,0.28983910878499347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,8192,0.049455109569761485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,2048,0.006342222293217977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,768,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,64,0.003730666720204883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,768,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,4096,0.08751289049784343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,2560,0.010307555397351583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,5120,0.0335822237862481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,512,0.04360533422893948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,2048,0.0303146673573388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,10240,0.9386480119493273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,2560,0.05169066786766052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,4096,0.2601546711391873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,3072,0.18863911098904082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,65536,0.5530844264560276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,51200,4.491362677680121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,768,0.02295111119747162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,128,0.10605777634514703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,16384,0.021914665897687275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,256,0.010414222048388587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,10240,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,8192,0.1149937841627333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,32,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,12288,0.10460622443093194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,3072,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,6144,0.07373511128955416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,5120,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,8192,0.052988443109724254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,10240,0.05961333380805122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,51200,0.08355644676420425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,6144,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,5120,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,1536,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,128,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,1024,0.04049777653482225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,6144,0.015059555570284525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,64,0.01498933302031623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,256,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,1536,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,1536,0.015595555305480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,16384,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,12288,0.04082399937841628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,32,0.013927110367351107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,51200,0.48571109771728516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,7168,0.055323554409874805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,512,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,12288,0.03149688906139798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,4096,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,1536,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,5120,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,4096,0.013842666314707862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,256,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,512,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,4096,0.03577422102292379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,1024,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,6144,0.1243893305460612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,4096,0.12576711177825928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,65536,0.38906489478217227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,10240,0.02127200033929613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,512,0.007086222370465596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,8192,0.013364444176355997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,2560,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,7168,0.019282667173279654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,128,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,1024,0.02369333306948344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,2048,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,64,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,256,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,3584,0.01682577696111467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,65536,0.21999822722540963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,16384,0.024205333656734888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,8192,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,3072,0.02091911103990343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,51200,0.056440002388424344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,2048,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,8192,0.0602088901731703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,10240,0.04532444477081299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,2560,0.022742221752802532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,2048,0.020976000361972388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,7168,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,12288,0.030064000023735896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,6144,0.017702221870422363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,4096,0.07713244358698527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,2048,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,6144,0.017646221650971305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,7168,0.18016355567508271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,4096,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,6144,0.034937779108683266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,32,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,512,0.08423644304275513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,64,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,10240,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,6144,0.03358577688535055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,16384,20.371612548828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,1536,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,5120,0.013759999639458127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,3072,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,768,0.03456444541613261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,8192,0.007596444752481248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,768,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,3584,0.17726843886905244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,4096,0.03134844369358487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,5120,0.010491555763615502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,51200,0.5674275292290581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,3072,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,32,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,65536,0.09639200237062241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,8192,0.15260622236463758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,32,0.03086755673090617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,8192,0.017859554953045316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,2048,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,6144,0.009487110707494948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,512,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,12288,0.0391244457827674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,6144,3.7999369303385415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,128,0.0155395550860299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,256,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,65536,2.05403561062283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,51200,0.1942417754067315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,10240,0.15647021929423013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,3072,0.0417866673734453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,32,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,3072,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,512,0.013804444836245643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,7168,0.04515733321507772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,64,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,128,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,512,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,8192,0.025288888149791296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,8192,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,3584,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,3072,0.0217884447839525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,16384,0.3454151153564453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,512,0.022163555026054382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,1024,0.031326221095191106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,128,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,4096,0.04390311241149902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,1536,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,1024,0.07815288835101657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,5120,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,7168,0.09900889131757949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,7168,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,3072,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,7168,0.016720889343155753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,65536,0.07043555710050795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,10240,0.011600000162919363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,6144,0.16164799531300864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,7168,0.08042844136555989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,8192,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,1024,0.010026666853162978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,512,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,4096,0.07226044601864286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,10240,1.6583688524034288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,2048,0.039889779355790876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,2048,0.020435555113686454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,65536,0.2549484570821126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,10240,1.201698621114095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,4096,0.40363110436333555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,3072,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,128,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,3584,0.05794577466117012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,8192,0.013782222237851886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,16384,0.07959910896089342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,2048,0.005316444569163852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,10240,0.02709688908523983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,10240,0.006384000182151794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,7168,0.019465777609083388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,3072,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,4096,0.01166488892502255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,2048,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,7168,0.10722844468222724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,12288,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,64,0.005320888840489917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,16384,0.041830221811930336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,2560,0.005014222115278244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,10240,0.04541688826349047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,1024,0.13748089472452799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,16384,0.49571731355455184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,3072,0.030583110120561387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,1024,0.00941066692272822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,7168,0.01475288967291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,6144,0.03605511122279697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,2048,0.09184266461266412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,32,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,10240,0.48703908920288086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,256,0.03475111060672336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,256,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,7168,0.012457778056462606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,2560,0.034183998902638756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,7168,2.347432878282335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,768,0.07453244262271456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,1536,0.04289155536227756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,64,0.022708444131745234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,3072,0.12494933605194092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,4096,0.08560178014967178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,51200,0.7643022007412381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,10240,0.03680977887577481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,4096,0.013225778109497495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,65536,0.0626879996723599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,5120,0.01370666672786077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,3584,0.015519110692871941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,3072,0.008469333251317343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,1536,0.030639999442630347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,2048,0.06286578045950995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,2560,0.011693333586057028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,256,0.002542222125662698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,768,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,2560,0.1544497807820638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,256,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,7168,0.7129137780931262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,6144,0.3357404337988959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,8192,0.022952889402707417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,32,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,3072,0.023183110687467787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,65536,0.09843289189868504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,3584,0.08457333511776394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,65536,1.1784337361653645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,6144,0.01571555601225959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,16384,0.04439111219512092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,256,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,6144,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,6144,0.011759999725553723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,1024,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,3584,0.06589066982269287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,2560,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,8192,0.013924444715181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,128,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,512,0.04236888885498047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,6144,0.2583048873477512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,2048,0.01625777781009674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,16384,0.027084444959958393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,1024,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,5120,0.33559287918938535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,12288,0.010816000401973724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,1536,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,512,0.016886222693655226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,512,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,3584,0.017902221944597032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,1536,0.0038240001433425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,512,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,65536,0.157404449250963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,2560,0.39757421281602645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,10240,0.021039111746682063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,3584,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,3072,0.07943555381562975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,1024,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,1536,0.04571733209821913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,3072,0.09479555818769668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,12288,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,51200,0.12157244152492946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,3072,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,32,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,51200,0.06259999672571818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,768,0.030745777818891738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,16384,0.11515288882785374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,5120,0.0933084421687656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,16384,0.039354665411843195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,3584,0.02465244465404087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,5120,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,51200,0.125255995326572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,3072,0.007567110988828871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,65536,0.08989510933558147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,10240,0.05367733372582329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,2048,0.025553777813911438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,3072,0.025842666625976562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,32,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,12288,0.1909982230928209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,3584,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,128,0.11879200405544704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,4096,0.11882221698760986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,51200,0.09647911124759251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,768,0.004087111188305749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,4096,0.021312000022994146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,7168,0.010247111320495605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,5120,0.009541333549552495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,3072,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,12288,0.04606399933497111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,10240,0.012276444170210095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,5120,0.025464889076020982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,2560,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,6144,0.018952889574898612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,2560,0.03706399930848016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,1536,0.033478223615222506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,2560,0.048229333427217275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,2560,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,128,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,16384,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,3584,0.019886222150590684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,128,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,5120,0.18704799811045328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,3584,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,7168,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,128,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,12288,0.027105778455734253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,3072,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,3584,0.09968800014919704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,5120,0.05379377802213033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,64,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,6144,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,7168,0.08125244246588813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,2560,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,5120,0.0763617753982544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,6144,0.014168888330459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,768,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,7168,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,7168,0.11332800653245713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,16384,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,128,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,1024,0.0664737754397922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,8192,0.029434667693244085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,2560,0.01552177800072564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,7168,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,8192,0.024494222468800012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,3584,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,6144,0.009291555318567488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,12288,0.016931555337376065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,2560,0.006809777683681912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,3584,0.007770666645632849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,6144,14.990456475151909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,768,0.017676444517241586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,3584,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,8192,0.04969244533114963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,12288,0.0898088879055447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,16384,5.590911865234375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,512,0.021525333325068157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,512,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,3584,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,2560,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,32,0.018403554956118267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,256,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,768,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,1024,0.00868888861603207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,7168,0.37595555517408585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,1536,0.005028444445795483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,1024,0.0037084445357322693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,2048,0.04298044575585259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,3584,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,3072,0.01865333318710327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,256,0.015442666080262927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,8192,0.01591999994383918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,12288,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,8192,0.14017511738671198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,256,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,16384,0.016227554943826463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,1536,0.038352890147103205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,51200,0.17355733447604707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,2560,0.024488000406159297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,65536,1.024653328789605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,8192,0.0663715534740024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,65536,0.25041156344943577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,32,0.043383111556371055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,12288,0.02904355525970459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,2048,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,51200,0.05066577924622429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,51200,0.728334214952257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,5120,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,6144,0.031207111146714952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,512,0.009272000028027428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,6144,0.025633777181307476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,1024,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,32,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,64,0.03663377629386054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,10240,0.020569778150982328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,3584,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,51200,10.20377434624566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,12288,0.053435554107030235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,12288,0.2707342306772868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,128,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,7168,0.021951999929216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,65536,0.07578844494289823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,6144,0.010870222416188983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,512,0.011683555940786997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,6144,0.19000444147321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,7168,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,4096,0.023299554983774822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,6144,0.22059555848439535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,4096,0.15097244580586752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,6144,1.0050089094373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,64,0.02012711101108127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,1024,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,8192,0.039297779401143394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,12288,0.02274399995803833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,32,0.020271110865804885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,65536,1.132597287495931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,2560,0.07920444673962064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,8192,0.019155555301242404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,512,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,5120,0.023053333163261414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,12288,0.02367377777894338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,4096,0.014228444960382251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,512,0.009743111001120673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,2560,0.007058666812049017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,5120,0.060912887255350746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,6144,0.044862220684687294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,1024,0.013037333057986366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,12288,0.10815466774834527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,8192,0.014854222536087036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,128,0.009331555830107795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,5120,0.02570488883389367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,5120,0.025597333908081055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,2048,0.09842133522033691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,768,0.010496888723638324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,7168,0.01538133290078905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,5120,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,128,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,3584,0.014555555250909595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,7168,0.21356533633338082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,3072,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,6144,0.12621688842773438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,2048,0.010699555277824402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,256,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,2048,0.05095644460784065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,7168,0.06445333030488756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,51200,0.11995555294884576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,8192,0.028978665669759113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,1536,0.019191111127535503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,7168,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,2560,0.06710400183995564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,3072,0.6378764576382107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,3584,0.05146222313245138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,1536,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,2048,0.011342222491900126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,16384,0.043395555681652494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,768,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,2560,0.016153777639071148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,512,0.040051556295818753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,128,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,768,0.004591111093759537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,8192,0.012096000214417776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,4096,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,1536,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,10240,0.015590222345458137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,8192,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,5120,0.027790221903059218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,4096,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,10240,0.11297778288523357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,256,0.03035022152794732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,256,0.005401777724424998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,12288,0.08581599924299453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,65536,0.16453066137101915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,3072,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,2048,0.06755377848943074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,4096,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,1024,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,10240,0.07083466980192396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,65536,0.14244800143771702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,256,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,12288,0.46228800879584414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,4096,0.08920355637868245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,2048,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,12288,0.009140444298585257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,51200,9.346573723687067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,6144,0.01314399970902337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,768,0.023431999815834895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,51200,0.9959608713785807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,16384,0.02536444365978241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,7168,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,51200,0.07105688916312324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,4096,0.18220978313022188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,8192,0.1775724490483602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,64,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,512,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,12288,0.019646222392718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,1024,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,51200,1.929149415757921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,2048,0.13112355603112116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,512,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,51200,1.1508622699313693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,10240,0.023856888214747112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,51200,0.08053688870535956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,5120,0.0715306666162279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,32,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,10240,0.016895999511082966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,3584,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,64,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,12288,0.07552266783184476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,10240,0.028284443749321833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,512,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,8192,0.07784889141718547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,3584,0.02109777761830224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,2048,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,16384,0.13627378145853677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,65536,0.14289155271318224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,256,0.015321777926550971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,5120,0.029193778832753498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,51200,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,64,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,3072,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,1024,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,12288,0.17068533102671304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,2048,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,4096,0.11226488484276666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,16384,0.015692444311247934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,7168,0.03660533494419522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,10240,0.018515555395020377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,768,0.08255466487672594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,128,0.06815377871195476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,1536,0.05191555619239807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,7168,0.01612711118327247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,128,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,1024,0.0028782222006056043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,4096,0.06644799974229601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,8192,0.013518222504191928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,5120,0.008352888955010308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,4096,0.015976000163290236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,6144,0.01982311076588101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,768,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,65536,5.842041863335504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,2560,0.007335111498832703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,7168,0.02236355510022905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,64,0.03170222375128005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,128,0.11737955941094293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,512,0.016858667135238647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,1536,0.020075556304719713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,4096,0.01921599937809838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,5120,0.03102133340305752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,3584,0.007090667055712805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,128,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,1536,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,6144,0.1868373288048638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,10240,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,64,0.005142222262091107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,7168,0.02714755634466807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,6144,0.12111555205451117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,6144,0.04301333427429199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,1024,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,2048,0.04354755414856804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,512,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,512,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,7168,0.10904355843861897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,51200,0.06100355254279243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,10240,0.04683644572893778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,5120,0.019359111785888672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,2560,0.12278400527106391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,12288,0.024950222836600408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,10240,0.010959111154079437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,2048,0.05721777677536011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,4096,0.1590506633122762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,256,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,7168,0.09601422150929768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,12288,0.17971732881334093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,12288,0.12269155184427898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,7168,0.06344266732533772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,128,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,4096,0.01936444474591149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,65536,5.994987487792969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,3584,0.12139556143018936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,3072,0.010956444674068026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,32,0.001965333397189776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,12288,0.07179377476374309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,512,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,8192,0.034091555409961276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,3072,0.4620480007595486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,51200,0.40145243538750547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,6144,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,10240,0.03865777783923679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,7168,0.9078462388780383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,4096,0.20451111263699004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,2048,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,3072,0.029380444023344252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,16384,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,7168,0.05199022094408671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,4096,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,2048,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,6144,0.027822222974565294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,512,0.11240533987681071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,768,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,512,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,768,0.008472888833946651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,128,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,3584,0.03200444579124451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,1024,0.02683555583159129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,65536,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,3584,0.045122666491402515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,256,0.04599644409285652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,4096,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,128,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,5120,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,3072,0.0075048887067370946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,16384,0.03414488832155863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,2560,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,7168,0.029631111356947158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,51200,0.04343466626273262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,3072,0.060308443175421826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,768,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,7168,0.05595111184649997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,1536,0.023286221755875483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,7168,0.013047110703256396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,12288,0.18343733416663277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,5120,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,32,0.005375111268626318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,1536,0.006343110981914733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,3584,0.036405334870020546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,512,0.029436445898479883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,3072,0.02481866710715824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,8192,0.2391724321577284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,51200,0.21003911230299208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,256,0.012666666673289405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,10240,0.015654222832785714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,512,0.02961955467859904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,3584,0.013593778014183044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,5120,0.008385777473449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,5120,0.027715555495686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,51200,0.4136346711052789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,3584,0.021871111459202234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,2048,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,3584,0.007717333734035492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,768,0.025309332542949255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,1536,0.016523554921150208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,2048,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,1536,0.0336408880021837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,512,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,51200,1.132205327351888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,6144,0.1281422244177924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,65536,0.15459467305077448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,2048,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,65536,0.046852442953321666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,4096,0.038959999879201256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,1536,0.019398222366968792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,1024,0.037017779217825994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,7168,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,256,0.025075554847717285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,2560,0.056094222598605685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,1024,0.02604977786540985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,51200,0.8590888977050781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,8192,0.0920142200258043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,1024,0.07300533188713922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,3584,0.36832178963555234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,1536,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,16384,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,65536,0.0663271082772149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,512,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,7168,0.020448888341585796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,8192,2.78532706366645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,51200,0.18506488535139295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,16384,0.01959733333852556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,7168,0.10286844438976711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,12288,0.07096888621648152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,3584,0.02036711076895396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,768,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,16384,0.11753955152299668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,6144,0.02717511190308465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,256,0.009938666390048133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,1536,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,12288,0.21752088599734834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,1024,0.010581333604123859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,1536,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,2560,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,1024,0.005088888936572605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,12288,0.3767351044548883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,768,0.03201333350605435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,1024,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,65536,0.06277689006593493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,128,0.019596444235907663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,2048,0.009865778187910715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,2560,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,2048,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,4096,0.007751999629868402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,12288,0.013629333012633853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,10240,0.04920088913705614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,65536,0.27804088592529297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,8192,0.012796444197495779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,65536,0.07630933655632867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,10240,0.10871289173762004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,256,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,1536,0.05230844351980421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,12288,0.10689955287509495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,5120,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,2048,0.0233262214395735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,128,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,65536,0.25424443350897896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,6144,0.018764444523387484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,4096,0.0951742198732164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,1536,0.17457866668701172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,512,0.41289512316385907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,8192,0.4490453402201335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,5120,0.01947733263174693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,768,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,65536,0.01794311073091295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,512,0.020453333854675293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,4096,0.06304088566038343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,10240,0.03559911251068115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,4096,0.016122667325867545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,8192,0.10605333248774211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,1536,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,6144,0.05858666367001004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,128,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,65536,0.07975644535488553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,10240,0.03192889028125339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,256,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,4096,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,5120,0.01494311127397749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,8192,0.6443982124328613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,51200,0.2242364353603787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,10240,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,7168,0.036348445547951594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,6144,0.084806223710378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,512,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,3072,0.011200000014570026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,768,0.015491555134455362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,8192,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,2048,0.018872888551818002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,2560,0.03737599982155694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,51200,0.17932800451914468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,12288,0.0314977765083313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,4096,0.6238391134474013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,256,0.08109333117802937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,3584,0.027863999207814533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,3072,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,64,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,128,0.015097777048746744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,1024,0.020945777495702107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,32,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,1024,0.014188443620999655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,16384,0.06520089175966051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,256,0.3369128968980577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,1536,0.012796444197495779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,128,0.015523556205961438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,16384,0.026730666557947796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,3072,0.03792533278465271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,768,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,65536,0.2843662103017171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,10240,1.6810195710923936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,3584,0.012816000315878125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,32,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,3584,0.014951111541854011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,10240,0.04423377911249796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,2048,0.030085331863827173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,4096,0.0697173343764411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,768,0.05194311009513008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,1536,0.21421866946750215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,64,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,10240,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,4096,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,7168,0.09213155508041382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,3072,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,128,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,5120,0.031868444548712835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,3072,0.6978080007765027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,6144,0.012208888928095499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,5120,0.7812790870666504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,12288,0.2445004516177707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,6144,0.01812799937195248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,65536,0.8141688770718045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,1536,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,4096,0.752571529812283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,128,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,51200,1.2748906877305772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,10240,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,256,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,8192,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,12288,0.4110106627146403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,1536,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,3584,0.02700088918209076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,32,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,64,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,5120,0.02312622302108341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,256,0.011155555645624796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,12288,1.8776160346137152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,3584,0.03831555445988973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,512,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,512,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,1536,0.00872444444232517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,32,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,1536,0.008047111332416534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,2560,0.018814222680197824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,5120,0.06564178069432576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,128,0.02027466727627648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,256,0.011688888900809817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,5120,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,7168,0.004968889057636261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,2560,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,4096,0.05225422316127353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,8192,0.013693333499961428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,512,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,12288,0.735248883565267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,768,0.04807733496030172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,768,0.2920488781399197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,10240,0.03904977771970961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,128,0.06438844733768039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,128,0.04007200068897671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,16384,0.06769511434766981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,16384,0.13281155957116023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,8192,0.010769777827792697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,128,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,768,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,12288,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,2560,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,8192,0.08260355393091838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,10240,0.11160266399383545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,8192,0.014966222974989148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,64,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,2560,0.02981688910060459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,6144,0.026731555660565693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,51200,0.21252444055345324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,4096,0.09540177716149224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,1536,0.0042204442951414315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,12288,0.19505243831210664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,10240,0.023366222778956097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,1024,0.022720888257026672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,7168,0.011104889214038849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,3584,0.037188443872663714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,128,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,51200,0.2673706743452284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,4096,0.026135111848513286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,32,0.05881510840521919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,2048,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,1024,0.018961777289708454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,256,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,256,0.028858668274349634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,65536,0.41452622413635254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,5120,0.10229155752393936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,64,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,2048,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,64,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,4096,0.034712000025643244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,1536,0.003261333331465721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,8192,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,4096,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,512,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,256,0.0218551109234492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,1024,0.0058462222417195635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,64,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,3584,0.01587466730011834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,3072,0.026148445076412622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,10240,0.03860444492763943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,10240,1.5394293467203777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,10240,0.21642043855455187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,768,0.05888444185256958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,2048,0.00435644429590967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,3072,0.0630355543560452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,1024,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,8192,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,5120,0.0359280010064443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,5120,0.11215555667877197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,1024,0.004396444393528832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,65536,0.2539137734307183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,51200,0.23389599058363172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,512,0.013912000589900546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,2560,0.011707555916574268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,4096,0.020776000287797716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,5120,0.5634213553534614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,3072,0.6363866594102647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,16384,0.22360889116923013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,3584,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,7168,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,8192,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,10240,0.02551555633544922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,64,0.04391466577847799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,2560,0.008180444439252218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,7168,0.40610843234592015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,2560,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,256,0.008051555189821456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,256,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,7168,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,128,0.014503111441930136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,32,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,5120,0.8902035819159614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,65536,0.07003200054168701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,16384,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,6144,0.02792088852988349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,51200,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,2560,0.03877066572507223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,51200,0.28825778431362575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,16384,0.021293333835071985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,16384,0.1925039953655667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,5120,0.043805334303114146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,12288,0.1204035547044542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,3584,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,4096,0.008800889054934183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,8192,0.014904000692897372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,8192,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,10240,0.0540284448199802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,10240,0.01612711118327247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,16384,0.05350044369697571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,4096,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,1024,0.026494221554862127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,256,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,3584,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,16384,0.019721777902709115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,8192,0.24696087837219238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,2048,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,128,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,512,0.013527111046844058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,3072,0.016711999972661335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,6144,0.015744000673294067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,1536,0.004568888909286923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,16384,0.06431911389033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,1024,0.0185199992524253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,1536,0.017785777648289997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,1024,0.051354666550954185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,1024,0.015336889359686108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,1536,0.0058373332851462895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,8192,0.015053333507643806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,3072,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,6144,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,65536,0.06820355521308051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,3072,0.1796497768825955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,64,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,12288,0.02940800123744541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,512,0.2102186679840088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,32,0.011845333708657159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,4096,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,51200,3.634916517469618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,8192,0.015415999624464246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,256,0.004171555654870139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,4096,0.018896889355447557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,8192,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,3072,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,256,0.023154666026433308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,3584,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,5120,0.009123555488056606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,1024,0.047648890150917896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,10240,0.10874933004379272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,512,0.011200889117187925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,768,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,768,0.01774311065673828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,65536,0.1451831128862169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,3584,0.17353422111935088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,1024,0.02494755552874671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,32,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,768,0.04666666520966423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,1024,0.03016977839999729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,3072,0.01034755590889189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,1536,0.032807111740112305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,256,0.0022897778285874263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,16384,0.16249422232309976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,1024,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,2560,0.040596445401509605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,10240,0.021595555875036452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,16384,0.03452800048722161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,6144,0.030560887522167627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,512,0.0034666665726237824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,1536,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,32,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,512,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,16384,0.08620978064007229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,16384,0.07543022102779813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,512,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,51200,0.05320622192488777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,2048,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,3072,0.017522666189405654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,3584,0.031181332137849595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,64,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,2048,0.04432177874777052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,51200,0.045406222343444824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,768,0.003900444342030419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,6144,0.009409777820110321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,65536,0.10668800274531047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,12288,1.4460773468017578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,65536,0.08868710862265693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,768,0.11871911419762506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,10240,0.01795377830664317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,12288,0.4531715710957845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,1536,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,2048,0.0041759999261962045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,7168,0.06645155615276761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,6144,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,2048,0.007349333001507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,10240,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,2560,0.010245333115259806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,1024,0.003119999956753519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,2560,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,512,0.0032275555034478507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,2048,0.018070222602950204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,768,0.014775110615624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,128,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,16384,0.07720533344480726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,32,0.017229333519935608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,256,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,1536,0.03259377678235372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,16384,0.4994364314609104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,51200,2.6695919036865234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,65536,0.14533778031667074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,5120,0.15516889095306396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,2048,0.0990791122118632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,1536,0.007121777368916406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,10240,0.2524293263753255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,512,0.0028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,10240,0.3237537807888455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,1024,0.034781333472993635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,1536,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,2560,0.009641778137948778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,6144,0.050028443336486816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,64,0.006350222147173352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,12288,20.876719156901043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,2048,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,65536,0.06993066602283053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,4096,0.01400888959566752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,1024,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,7168,0.13298222753736708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,5120,0.22417243321736655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,2048,0.02069155540731218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,5120,0.04827910992834303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,512,0.02474311159716712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,1536,0.030589332183202107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,5120,0.01475911173555586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,32,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,51200,0.06050133042865329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,32,0.05272000034650167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,7168,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,51200,0.06907466385099623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,3072,0.06736710998747084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,768,0.009876444107956356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,768,0.04071999920739068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,1024,0.009378666679064432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,3072,0.02892177634769016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,4096,0.28394489818149143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,3072,0.1363511085510254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,7168,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,7168,2.342450671725803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,1536,0.02363555630048116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,256,0.026778666509522334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,7168,0.05227288934919569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,65536,0.029599110285441082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,128,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,2560,0.008043555749787224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,8192,0.02241155505180359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,5120,0.03133777777353922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,8192,0.09461510843700832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,10240,0.030636443032158747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,7168,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,65536,0.0598506662580702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,128,0.01533777846230401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,7168,0.02090577781200409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,2560,0.9749849107530383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,128,0.32927465438842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,10240,0.15464888678656682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,128,0.005623111294375525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,256,0.3547031084696452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,64,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,6144,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,3072,0.023748444186316595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,7168,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,5120,0.05753778086768257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,7168,0.012273777690198688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,768,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,8192,0.036554667684766985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,512,0.015808888607554965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,6144,0.012488889197508493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,768,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,64,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,6144,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,1536,0.016889777448442247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,1536,0.060441778765784375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,10240,0.024869332710901897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,256,0.018182223041852314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,1536,0.07858044571346708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,768,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,512,0.028565333949195013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,64,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,10240,0.03147555390993754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,256,0.009053332938088311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,7168,0.017483555608325534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,65536,0.5337128639221191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,3584,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,128,0.010738666686746808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,64,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,8192,0.1383591095606486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,3584,0.021253334151373968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,3072,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,2560,0.17654666635725233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,3072,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,12288,0.03247022297647264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,32,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,7168,0.04104355639881558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,768,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,16384,0.03212622139188979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,128,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,768,0.029093331760830347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,7168,1.5107475916544597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,1024,0.03322488731808133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,1024,0.01538133290078905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,4096,0.019274666905403137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,2560,0.03839110996988084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,1024,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,3072,0.009573332965373993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,128,0.08698044220606486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,12288,0.037124445041020714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,5120,0.0322017769018809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,3584,0.11674133936564128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,768,0.002977777686383989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,2048,0.022200889057583276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,65536,0.1604942215813531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,768,0.0034053333931499054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,51200,0.3124897744920519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,10240,0.01997777819633484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,16384,0.07856533262464735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,512,0.007726222276687622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,3072,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,32,0.001791999985774358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,16384,0.24648356437683105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,6144,0.030253334177864924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,7168,0.10374222199122111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,128,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,512,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,10240,0.1599386665556166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,2048,0.0041911109454101985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,7168,0.019260444574885897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,768,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,6144,0.1309973398844401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,768,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,1536,0.039320889446470476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,16384,0.32302133242289227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,2560,0.2692106564839681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,8192,0.02870755394299825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,1536,0.0199297782447603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,5120,0.11663466029696995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,32,0.002279111080699497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,12288,0.04834666517045763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,51200,0.05958133273654514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,16384,0.018333334061834548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,4096,0.0087013335691558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,128,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,128,0.003600888782077365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,512,0.015508444772826301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,1024,0.01184622198343277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,512,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,256,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,65536,0.09095111158159043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,4096,0.04009688893953959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,3072,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,6144,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,5120,0.010651555326249866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,1024,0.011976000335481433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,6144,0.022385777698622808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,256,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,5120,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,2560,0.05453422334459093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,8192,0.02516977820131514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,128,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,16384,0.030431999100579157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,32,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,10240,0.039495110511779785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,65536,0.10608088970184326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,256,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,10240,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,256,0.203658660252889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,4096,0.09414222505357529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,65536,0.37149421374003094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,1024,0.006873777757088344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,5120,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,64,0.006676444576846228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,3584,0.014944000376595391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,6144,0.11632443798912896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,5120,0.1509715583589342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,3072,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,5120,0.027223111854659185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,32,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,51200,1.0393697950575087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,6144,0.015233778291278414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,12288,0.04562222295337253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,32,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,128,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,768,0.021631111701329548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,8192,0.011418666276666852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,6144,0.05470844440990024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,512,0.009713778065310584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,51200,0.22534222073025176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,12288,1.277374267578125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,7168,0.022218666142887537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,16384,0.022262222237057153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,51200,0.09382933378219604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,128,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,2048,0.026260443859630164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,7168,0.06320355335871379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,512,0.013153778182135688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,65536,1.4860435061984594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,8192,0.05067377620273166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,2048,0.10875999927520752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,10240,0.3654799991183811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,51200,0.41504266526963973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,2560,0.02812533246146308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,6144,0.05743110842174954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,8192,0.03843999902407328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,4096,0.13785333103603786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,2048,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,3584,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,6144,0.03299466768900553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,768,0.02075466679202186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,10240,0.03084355592727661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,3072,0.02311466634273529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,8192,0.027850667635599773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,6144,0.025239111648665533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,16384,0.718031088511149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,1536,0.013826666606797112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,512,0.0027111110587914786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,32,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,512,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,512,0.013226666384273104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,3072,0.038188444243537054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,65536,0.027446221974160936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,2048,0.010420444111029306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,51200,0.9100284576416016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,6144,0.03311644328965081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,51200,0.11445155408647324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,8192,0.03320533368322585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,4096,0.23673155572679308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,4096,0.14965866671668157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,51200,0.9694275326199002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,10240,0.05811466773351034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,768,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,2560,0.00945955514907837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,5120,0.0086986662613021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,128,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,64,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,8192,0.11783377329508464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,7168,0.0664186676343282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,5120,0.17432444625430635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,16384,0.06327022446526422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,2560,0.020921778347757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,7168,0.12261777453952366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,8192,0.007928888830873702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,64,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,6144,0.030288888348473444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,4096,0.05209066801600986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,65536,0.18068710962931314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,1536,0.01646311084429423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,1024,0.02570399973127577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,12288,0.09934755828645493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,256,0.0288080010149214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,2048,0.022163555026054382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,3072,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,768,0.02454400062561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,256,0.01572533282968733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,51200,0.05134400063090854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,7168,0.03229066729545593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,256,0.0070053330726093715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,2560,0.039106667041778564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,2560,0.03128088845147027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,1536,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,7168,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,768,0.01107911103301578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,128,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,7168,0.10222133662965562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,7168,0.14461778269873724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,128,0.02238311039076911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,128,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,256,0.015629332926538255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,768,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,256,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,256,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,65536,1.5162791146172419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,65536,0.0657235582669576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,7168,0.2638808886210124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,3072,0.00872622181971868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,64,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,4096,0.017264889346228707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,256,0.02108000053299798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,3584,0.06842311223347981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,5120,0.017619555195172627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,1024,0.00361333332127995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,512,0.014397333065668741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,3584,0.013224889006879596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,1024,0.005483555710977978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,8192,0.5831182267930772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,128,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,5120,0.019183110859658983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,512,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,512,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,65536,0.5574675665961372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,2560,0.033026665449142456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,10240,0.015440000428093804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,512,0.006804444309737947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,256,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,2048,0.006846222198671765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,16384,0.21959022680918375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,8192,0.012668444050682915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,768,0.07903289132648043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,5120,0.01089866707722346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,8192,0.0381040010187361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,512,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,768,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,10240,0.02683466672897339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,1536,1.3845858044094508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,51200,0.14218844307793513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,1024,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,3584,0.008362666600280339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,128,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,512,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,7168,0.01720266706413693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,5120,0.21867199738820395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,6144,0.2201626698176066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,1024,0.06872800323698255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,16384,0.0409822232193417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,5120,0.016216889023780823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,16384,3.161466598510742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,512,0.12975377506679958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,128,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,128,0.003107555624511507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,128,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,7168,0.03088266650835673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,12288,0.01704088846842448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,5120,0.03854488995340135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,4096,0.301055113474528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,12288,0.017616000440385606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,10240,0.032905777295430504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,51200,6.507464938693577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,3072,0.02355555527740055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,1024,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,2560,0.02569244470861223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,512,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,768,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,128,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,768,0.14064710670047337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,1536,0.019694222344292533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,8192,0.016892444756295946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,12288,0.019889778561062284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,3584,0.0288080010149214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,2048,0.019292443990707397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,128,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,12288,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,7168,0.04639288783073425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,128,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,6144,0.02924355533387926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,768,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,8192,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,7168,0.0366231103738149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,1024,0.01831377711560991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,768,0.01778133379088508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,5120,0.1111875507566664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,1536,0.027871112028757732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,4096,0.01714933415253957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,512,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,5120,0.0177502218219969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,768,0.01680799987581041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,51200,1.2556631300184462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,32,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,65536,0.3741893238491482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,4096,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,2560,0.03623200125164456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,2048,0.03521155648761325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,12288,0.062112887700398765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,256,0.024305777417288885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,6144,0.07461688915888469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,51200,0.324360900455051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,6144,0.04405155446794298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,12288,0.03842844565709432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,8192,0.018138666947682697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,3584,0.04306755463282267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,16384,0.16020177470313177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,2560,0.021935110290845234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,7168,0.22048710452185738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,256,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,12288,0.02219377789232466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,51200,0.3999839888678656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,2560,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,6144,0.03377155462900797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,51200,0.06552888949712117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,256,0.10366933213339911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,768,0.013414222333166333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,3072,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,5120,0.018443556295500863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,3072,0.008446222378147973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,3584,1.4175057941012914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,16384,0.030807998445298936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,64,0.0025600000388092464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,7168,0.026777777406904433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,2560,0.047621332936816745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,1536,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,1536,0.0736808909310235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,8192,0.018975110517607797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,3584,0.03125422199567159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,1024,0.006319111006127463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,16384,0.0381626652346717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,16384,0.01404266721672482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,512,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,2048,0.031276444594065346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,65536,0.23760000864664713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,3072,0.0075093333919843035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,1024,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,256,0.016879111528396606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,3584,0.012831110921170978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,1024,0.04120622078577677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,64,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,16384,0.26047910584343803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,3584,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,16384,0.0880533324347602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,4096,0.01961866683430142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,8192,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,8192,0.03396711084577773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,768,0.029121776421864826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,12288,0.06498310963312785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,10240,0.012495111260149213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,2560,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,7168,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,7168,0.2745644516415066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,8192,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,3072,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,128,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,12288,0.026723555392689172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,10240,0.01744888888465034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,1024,0.09223377704620361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,256,0.023034666975339253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,3072,0.023448000351587932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,12288,0.0748879975742764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,8192,0.2528231143951416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,128,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,12288,0.1086568898624844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,256,0.010636444720957013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,6144,0.0203457772731781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,10240,0.07980266544553968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,65536,2.4480311075846353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,8192,0.2868844403160943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,1536,0.05786755349900988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,7168,0.027407111393080816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,1536,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,65536,0.2820142110188802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,1536,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,4096,0.008157333566082848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,768,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,8192,0.05914755662282308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,51200,0.05262488789028592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,2048,0.028522666957643297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,768,0.00905600024594201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,8192,0.01571466690964169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,768,0.07608799801932441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,512,0.0025519999778932994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,4096,0.4012506537967258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,65536,0.07096533642874824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,6144,0.016315556234783597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,65536,3.5577057732476125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,65536,0.06787288851208158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,2048,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,1536,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,10240,0.02649066680007511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,1024,0.0738275580936008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,2560,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,7168,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,256,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,4096,0.06195555792914497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,51200,0.05089599887530009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,16384,0.04040177663167318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,2560,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,3072,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,3072,0.039559112654791936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,8192,0.02661777701642778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,6144,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,3584,0.0753119985262553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,6144,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,5120,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,2560,0.014395554860432943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,768,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,64,0.0019404445257451797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,5120,0.060805333985222705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,5120,0.00942311104800966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,768,0.019296000401178997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,3072,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,6144,0.02658577760060628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,10240,0.16221243805355495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,16384,0.2767724461025662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,768,0.024724443753560383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,8192,0.0244862222009235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,16384,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,5120,0.033168001307381526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,3584,0.026299556096394856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,1536,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,7168,0.010068444742096795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,3584,0.03924533393647935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,512,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,768,0.024086222052574158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,8192,0.19195111592610678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,16384,0.3092435465918647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,2560,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,128,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,3072,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,768,0.004743111216359668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,7168,0.09717422061496311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,7168,0.01276711126168569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,65536,0.1904933320151435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,16384,0.11726133028666179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,2560,0.005087999833954705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,8192,0.14697689480251735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,1024,0.020643555455737643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,3584,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,6144,0.07684622208277385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,12288,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,51200,0.21844532754686144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,5120,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,3584,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,768,0.08075111442142062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,4096,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,5120,0.11528266800774468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,3072,0.03795466820398966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,12288,0.15208621819814047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,12288,0.040707555082109236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,2048,0.013411555853154926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,7168,0.141975998878479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,12288,0.04059200154410468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,3072,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,12288,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,10240,0.012176888684431711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,256,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,51200,0.05788533555136787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,51200,0.46723376380072695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,5120,0.051727109485202365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,2048,0.024096000525686476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,16384,0.1117946704228719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,65536,0.06829333305358887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,7168,0.017461333009931777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,4096,0.7386879920959473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,51200,0.04995377858479818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,4096,0.06831822130415174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,8192,0.023764444722069636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,256,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,768,0.022288888692855835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,4096,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,3072,0.025219556358125474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,51200,0.7686346901787652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,8192,0.0912640028529697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,10240,0.024719999896155462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,3072,0.007818666597207388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,4096,0.023759111762046814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,51200,0.13603822390238443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,128,0.07699200179841784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,16384,0.1785635550816854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,1536,0.3906124432881673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,51200,1.6033271153767903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,512,0.030982222821977403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,1536,0.010112889111042023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,51200,0.07766933573616876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,512,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,32,0.009568000005351173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,768,0.0026844443960322272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,3584,0.008550222549173567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,1536,0.056259552637736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,1536,0.04352088769276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,51200,0.05024622215165032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,4096,0.29730044470893013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,768,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,5120,0.010433778166770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,4096,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,4096,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,7168,0.22559555371602377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,2048,0.005710222240951326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,3584,0.04416266745991177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,12288,0.2162862221399943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,768,0.004671999977694618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,6144,0.0748702221446567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,7168,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,256,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,5120,0.02379111117786831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,768,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,5120,0.01722755531469981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,64,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,2048,0.017210667332013447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,5120,0.02913155489497715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,10240,0.03975644376542833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,3584,0.06609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,2048,0.14634754922654894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,51200,0.06334133280648126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,4096,0.02509422269132402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,4096,0.09587466716766357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,6144,0.41393423080444336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,5120,0.011151110960377587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,10240,0.09852088822258843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,65536,0.07818044556511773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,3584,0.10635733604431152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,51200,0.028663999504513208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,7168,0.0455715556939443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,64,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,128,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,8192,0.01535022258758545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,8192,0.05117511086993747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,2048,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,768,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,128,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,3072,0.02683199942111969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,1024,0.01684711045689053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,5120,0.018958222534921434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,16384,0.020263110597928364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,1024,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,4096,0.040808889600965716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,12288,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,3584,0.03535022338231405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,256,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,10240,0.014126222994592456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,3584,0.06458844741185506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,768,0.0038782221575578055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,4096,0.01828533411026001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,8192,0.033486220571729876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,6144,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,3072,0.037610666619406805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,2560,0.01622222198380364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,2048,0.007104888558387756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,768,0.0058648888435628675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,3072,0.03910933269394769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,1536,0.027630221512582567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,1024,0.02681066592534383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,10240,0.06451911396450467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,32,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,51200,2.5398854149712458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,51200,0.04864355590608385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,8192,0.15599821673499212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,32,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,1024,0.04012088974316915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,8192,0.029858665333853826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,4096,0.033674667278925575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,51200,0.09855555825763279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,12288,0.10797778103086685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,5120,0.0254711111386617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,4096,0.008959111240175035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,512,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,768,0.0030871110243929755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,16384,0.11236711343129475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,16384,0.048229333427217275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,3072,0.012649777862760754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,10240,0.014896889527638754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,10240,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,3584,0.014753777119848462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,16384,0.5059350861443414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,128,0.012404444317022959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,8192,0.019991111424234178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,7168,0.1299617820315891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,51200,0.47828091515435117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,4096,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,3072,0.0377786656220754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,32,0.007715555528799693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,3584,0.07476800017886691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,3584,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,16384,0.14535555574629042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,3584,0.29031644927130806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,3072,0.060694224304623075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,512,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,768,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,51200,7.9823252360026045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,16384,0.04654666781425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,2048,0.02943555514017741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,7168,0.06039377715852526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,16384,0.1817706690894233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,65536,0.2959057754940457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,64,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,3072,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,51200,0.052108443445629544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,5120,0.021853332718213398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,2560,1.9201884799533422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,768,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,2560,0.01834222177664439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,2560,0.042415112257003784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,512,0.025535111625989277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,16384,0.04647377795643277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,3072,0.09811644421683417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,2048,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,7168,0.014214222629865011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,2560,0.02763644523090786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,256,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,10240,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,10240,0.14354933632744685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,32,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,51200,0.9674239688449435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,3584,0.22070133686065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,2560,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,6144,1.0786586337619357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,1024,0.2872142261928982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,3072,0.005856888989607493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,128,0.010899555351999072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,8192,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,7168,0.4876577589246962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,51200,0.11123643981085883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,2048,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,4096,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,1536,0.022916444473796423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,32,0.006868444383144379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,2048,0.013392888837390475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,128,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,2048,0.03348088926739163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,3584,0.0743502246008979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,8192,0.14045243793063694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,12288,0.0075804442167282104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,4096,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,7168,1.0322471194797092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,512,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,128,0.02976622184117635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,6144,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,1024,0.00980355590581894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,3584,0.018802666001849704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,3072,0.021669333179791767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,2560,0.11037866936789618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,10240,0.04622488882806566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,51200,0.19981866412692598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,1024,0.03423022230466207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,65536,0.1333742274178399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,768,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,2048,0.0047662220895290375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,51200,0.04065955678621928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,7168,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,5120,0.009363555245929295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,256,0.030125333203209773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,12288,0.02720355490843455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,128,0.024775111012988623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,4096,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,16384,0.04468355576197306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,512,0.020929776959949072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,5120,0.07200000021192762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,768,0.21791288587782118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,3072,0.16980977853139242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,128,0.010959111154079437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,16384,0.12486310799916585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,8192,0.009462222456932068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,16384,0.021359999974568684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,5120,0.009386666946940953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,5120,0.11822844876183404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,1024,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,6144,0.5768346786499023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,256,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,16384,0.08491110801696777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,1536,0.0038862224254343244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,3584,0.03847288754251268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,3072,0.031344887283113264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,16384,0.05835821893480089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,256,0.015105777316623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,12288,0.13349244329664442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,51200,0.13259022765689427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,51200,0.011023999916182624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,768,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,16384,0.031919111808141075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,1536,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,512,0.04673688941531711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,51200,0.05099733339415657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,768,0.03017866611480713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,512,0.05607199668884277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,512,0.013820444544156393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,768,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,10240,0.011795555551846823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,16384,0.024998222788174946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,256,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,12288,0.3095528814527723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,512,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,2048,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,3584,0.018515555395020377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,32,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,8192,0.035975999302334256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,64,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,256,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,3072,0.012447111308574677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,12288,0.03090488910675049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,2560,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,1536,0.1534284485710992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,3072,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,1024,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,3072,0.012704888979593912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,128,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,512,0.013812444276279874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,3584,0.10018399688932632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,5120,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,256,0.0057591112951437635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,65536,0.30051199595133465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,2560,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,51200,0.1459831131829156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,1536,0.34297866291469997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,6144,0.33040356636047363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,5120,0.3189946545494927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,1536,0.42058398988511825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,65536,0.0510755545563168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,256,0.008375110725561777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,6144,0.044996443721983165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,256,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,256,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,2560,0.03043466806411743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,16384,0.04447466797298855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,3584,0.2517777813805474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,16384,0.06071644359164768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,256,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,12288,0.12076000372568767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,8192,0.1466933356391059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,32,0.01836266616980235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,5120,0.020184889435768127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,4096,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,10240,0.04141955574353536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,10240,0.10417066680060492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,8192,0.042176889048682324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,1536,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,6144,0.08153066370222303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,3072,0.015885333220163982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,5120,0.008216888540320927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,65536,0.08163022332721286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,3072,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,64,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,51200,0.05627555317348904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,12288,0.11982844935523139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,128,0.024999111890792847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,768,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,768,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,16384,0.051968889103995435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,65536,0.47711557812160915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,768,0.018183999591403537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,8192,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,1536,0.138246218363444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,10240,0.0281333327293396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,51200,0.4277013407813178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,3584,0.01590399940808614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,128,0.03640888796912299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,3584,0.02165777815712823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,3072,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,5120,0.031230221192042034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,12288,0.037795556916130915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,7168,0.14077422353956434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,65536,0.13844533761342367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,4096,0.011454222102959951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,1536,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,64,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,2048,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,5120,0.06340889135996501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,51200,0.06555289030075073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,5120,0.12103911240895589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,64,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,128,0.0019235555082559586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,32,0.00870577742656072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,16384,0.11548356215159099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,768,0.003584889074166616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,512,0.15299644735124376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,1024,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,51200,0.12499822510613336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,51200,0.07207289006974962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,1024,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,51200,0.9635253482394748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,4096,0.01519377695189582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,32,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,16384,0.05000799894332886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,7168,0.025447999437650044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,51200,1.1915884017944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,8192,0.5267430941263834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,2048,0.008928889201747047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,16384,0.02166044380929735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,1024,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,3072,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,12288,0.23840000894334581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,128,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,3584,0.08212711413701375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,8192,0.029162665208180744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,8192,0.03378488951259189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,64,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,65536,0.049475553962919444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,1536,0.026032888227038916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,51200,0.1826453341378106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,10240,0.13431821929083929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,6144,0.08333866463767158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,1024,0.0073724447025193115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,4096,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,1536,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,3584,0.3758080005645752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,16384,0.07037599881490071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,64,0.042958223157458834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,51200,0.05035822259055244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,1536,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,10240,0.05040711164474487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,51200,0.29007289144727916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,7168,0.6983048650953504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,1536,0.01612444387541877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,32,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,2560,0.020900444851981267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,512,0.01681866745154063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,2048,0.028154667880800035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,1536,0.29353777567545575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,1536,0.025120000044504803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,1536,0.016591111818949383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,3072,1.555029339260525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,7168,0.04000000158945719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,64,0.04132444328731961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,3584,0.033561779393090144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,65536,0.03793155484729343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,12288,0.01637066735161675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,1024,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,1024,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,2048,0.01958755486541324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,5120,0.05982933441797892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,256,0.08486488792631362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,7168,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,512,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,512,0.01849600010448032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,16384,1.5295528835720484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,8192,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,8192,0.016590222716331482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,1024,0.012152888708644442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,2560,0.09551111194822524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,3584,0.06773244672351413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,16384,0.03558133376969232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,16384,0.02796444296836853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,10240,0.6429697672526041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,12288,0.014103111293580798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,256,0.0028719999310043124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,3072,0.04833777745564779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,3072,0.08077155881457858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,4096,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,2048,0.05968711111280653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,128,0.016733333468437195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,65536,1.0461609098646376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,12288,0.08520711130566067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,10240,0.0255724440018336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,3072,0.04680977927313911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,1024,0.032122668292787336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,6144,0.0105591110057301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,2560,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,512,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,768,0.0548257761531406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,4096,0.023984000086784363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,1536,0.022107554806603327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,3584,0.046784000264273755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,12288,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,4096,0.061884442965189614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,16384,0.016321778297424316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,2560,0.030117332935333252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,3072,0.009639110830095079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,2560,0.015221332510312399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,6144,0.8030568758646647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,2048,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,51200,1.4937796062893336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,2048,0.010100444157918295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,5120,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,10240,0.020608888732062448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,3584,0.46284442477756077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,256,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,8192,0.03897777862019009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,32,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,6144,0.1534764501783583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,512,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,768,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,51200,0.043102221356497876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,6144,0.07037333647410075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,8192,0.015561777684423657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,1536,0.04052799940109253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,768,0.010048000348938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,65536,0.7997466723124186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,7168,0.06476266516579522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,12288,0.1205857727262709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,6144,0.04117688867780898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,7168,0.03491288754675124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,1024,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,1536,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,128,0.014714666538768344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,3584,0.007728888756699032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,2048,0.03695111142264472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,12288,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,1536,0.018766222728623282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,64,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,64,0.09466666645473903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,5120,0.00904266701804267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,1024,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,1024,0.04548977812131246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,1536,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,1024,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,32,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,64,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,4096,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,51200,0.9242382049560547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,65536,0.05199822121196323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,6144,0.11225777202182347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,768,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,2048,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,1024,0.0163128889269299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,3072,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,32,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,4096,0.04020444552103678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,10240,0.32286045286390513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,2560,0.03029600116941664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,128,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,65536,0.06487199995252821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,2560,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,768,0.013548444542619916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,768,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,4096,0.01110400011142095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,4096,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,51200,0.1555751164754232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,10240,0.6941644350687662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,512,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,4096,0.16619200176662868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,1536,0.006122666514582104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,4096,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,4096,0.1813031170103285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,7168,0.022038223014937505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,768,0.02478222217824724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,3584,0.03772177630000644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,65536,0.38477155897352433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,8192,1.4165973663330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,1536,0.05108977688683403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,4096,0.020554666717847187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,32,0.005977777971161737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,512,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,1024,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,3072,0.03215377860599094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,10240,0.06651910808351305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,3072,0.24952265951368544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,1024,0.03135200010405646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,4096,0.13870222038692898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,65536,0.12471911642286514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,7168,0.012144000165992312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,8192,0.014706666270891825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,1536,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,1536,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,65536,0.66829956902398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,65536,0.26643289460076225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,10240,0.028367112080256145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,8192,0.06731200218200684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,7168,0.16007199552324083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,65536,3.547021230061849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,6144,0.019088000059127808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,16384,0.016203555795881484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,10240,0.0972248911857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,768,0.3617848820156521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,3584,0.00942133367061615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,128,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,10240,0.03340977761480544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,5120,0.034895999564064875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,10240,0.016456888781653512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,256,0.013285333083735572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,64,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,3584,0.026075555218590632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,3072,0.019527999891175162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,1536,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,65536,2.7853501637776694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,5120,0.06925421953201294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,2560,0.02823822365866767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,3584,0.03314400050375197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,512,0.009913778139485253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,10240,0.025507556067572698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,128,0.014091556270917257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,64,0.09088533454471165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,768,0.010741333166758219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,4096,0.033478223615222506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,51200,1.5802169375949437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,6144,0.459828429751926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,4096,0.01824355622132619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,512,0.01168444421556261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,8192,0.03192977772818671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,3072,0.06400355365541247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,256,0.0042453333735466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,7168,0.026752000053723652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,32,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,3072,0.0075706665714581805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,512,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,65536,0.33732443385654026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,6144,0.09875910811954075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,7168,0.14001154899597168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,256,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,1024,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,5120,3.0165280236138234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,5120,0.11836800310346816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,12288,0.04911288950178358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,51200,7.923261854383681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,768,0.03866577810711331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,5120,0.0404079986943139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,512,0.08398844136132134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,16384,0.1127928892771403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,768,0.025388444463411968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,2560,0.012178666889667511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,4096,0.03155466583040025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,2048,0.03846933444341024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,3072,0.021370665894614324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,51200,42.753028021918404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,5120,0.03851822349760268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,512,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,768,0.022308443983395893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,65536,0.2835306591457791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,7168,0.015171556009186639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,8192,0.014691554837756686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,16384,0.20542221599155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,1024,0.0573128859202067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,51200,0.12981777720981175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,1536,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,8192,0.015200000670221118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,768,0.02462933295302921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,64,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,32,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,10240,0.1850177711910672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,256,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,1024,0.017954667409261067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,10240,0.02676533328162299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,16384,0.16678310765160453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,12288,0.02681155502796173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,256,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,10240,0.23873066902160645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,32,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,2048,0.0616168909602695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,768,0.0439573327700297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,12288,0.07664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,256,0.0199644449684355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,8192,0.013361777696344586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,4096,0.011916444533401065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,2048,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,51200,0.5754053327772353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,7168,0.030776888132095337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,8192,0.01761511133776771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,12288,0.0348293317688836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,2560,3.9779476589626737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,3584,0.012819555898507437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,64,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,7168,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,65536,0.9740720325046115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,1536,0.009718221922715506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,4096,0.040433777703179255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,65536,0.07798310783174303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,128,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,5120,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,3584,0.010998222563001843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,65536,0.22870400216844347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,8192,0.022309333086013794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,4096,0.03087466624048021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,7168,0.1331982215245565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,512,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,256,0.01904444396495819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,51200,0.019902222686343722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,1024,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,2560,0.0125608891248703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,64,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,3072,0.05659111340840658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,51200,0.05051022105746799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,2048,0.0233262214395735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,1536,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,3072,0.008270222279760573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,65536,0.06188178062438965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,128,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,128,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,5120,0.02232622272438473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,2048,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,6144,0.1415991120868259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,8192,0.02641422218746609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,12288,0.027172444595230952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,7168,0.026748443643252056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,2048,0.014507555299335055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,4096,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,256,0.15212533209058973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,7168,2.3423298729790583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,1024,0.07791199949052599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,4096,0.024677332904603746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,32,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,768,0.21171022786034477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,7168,0.010218666659461128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,2048,0.004549333204825719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,16384,0.029002666473388672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,7168,0.9446026484171549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,512,0.030976888206270006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,3072,0.05141422152519226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,3584,0.039621333281199135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,12288,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,1536,0.034993777672449745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,8192,0.026093333959579468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,2048,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,256,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,51200,0.12787111600240073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,6144,0.05539110965198941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,256,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,2560,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,7168,0.020728000336223178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,2560,0.01644711030854119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,4096,0.03527644607755873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,3584,0.25320090187920463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,768,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,6144,0.02580533259444767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,65536,0.09511822462081909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,1536,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,10240,0.23489689826965332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,128,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,3584,0.0425217780801985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,1536,0.026141333911154006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,12288,0.02823822365866767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,128,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,65536,0.2880764537387424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,10240,0.04240533378389147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,1536,0.06982133123609754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,5120,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,128,0.012292444705963135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,768,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,4096,0.13316888279385036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,12288,3.3344169192843967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,5120,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,256,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,8192,0.07470844189325969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,512,0.010404444403118556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,3584,0.0105608892109659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,4096,0.034204443295796715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,3584,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,2048,0.013394667042626275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,7168,0.5302729076809353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,2048,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,512,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,512,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,1024,0.022524444593323603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,16384,0.267722659640842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,51200,4.703769683837891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,1024,0.027722666660944622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,2048,0.09236800008349949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,2560,0.10707822110917832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,768,0.034853332572513156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,512,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,1536,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,32,0.0525493323802948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,12288,0.02032711108525594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,5120,0.04063733418782552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,6144,0.0928737786081102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,10240,3.590246412489149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,3072,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,32,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,32,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,7168,0.013145777914259167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,7168,0.08422933022181193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,12288,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,3584,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,8192,0.13260089026557073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,3072,0.041987554894553296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,2048,0.017210667332013447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,1536,0.017616889543003507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,3072,0.0910373330116272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,1024,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,3072,0.09078755643632676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,51200,0.05634311172697279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,7168,0.012465777496496836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,1536,0.008941333327028487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,6144,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,1024,0.024373332659403484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,32,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,32,0.0018684443914228016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,2048,0.0199626667631997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,4096,0.1936204433441162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,6144,0.011588444312413534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,128,0.006213333457708359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,51200,0.3281804455651177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,2048,0.006107555495368109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,64,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,6144,0.12081333001454671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,65536,0.04678755667474535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,1024,0.006200888918505774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,2560,0.03640977872742547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,768,0.038252443075180054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,512,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,65536,0.1799688869052463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,256,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,2560,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,8192,0.06055733230378893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,3072,0.07315555546018812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,4096,0.011765333513418833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,12288,0.0517155561182234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,1536,0.006674666785531574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,128,0.023344889283180237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,65536,0.08677422338061863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,2048,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,51200,0.986387570699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,8192,0.04099999864896139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,768,0.022288888692855835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,3072,0.47114308675130206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,12288,0.10386044449276394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,1024,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,3072,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,16384,0.01348088930050532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,128,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,10240,0.19013688299391004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,16384,0.10693333546320598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,8192,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,65536,0.16371466053856745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,128,0.00941244430012173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,16384,0.23496799998813203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,3072,0.009477333062224919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,3584,0.014528888795110913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,4096,0.02381333377626207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,65536,0.046903110212749906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,3584,0.027104889353116352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,4096,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,128,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,256,0.008738666772842407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,5120,0.11092711157268947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,10240,0.01773955590195126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,7168,0.0229093333085378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,2048,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,1024,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,12288,0.02886399957868788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,8192,0.22664443651835123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,7168,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,256,0.02774133284886678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,128,0.01237066669596566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,7168,0.02153066628509098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,64,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,65536,0.9726311365763346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,12288,0.10322488678826226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,512,0.054343998432159424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,256,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,1024,0.003548444559176763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,2560,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,51200,1.0442320505777996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,2560,0.057132442792256675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,8192,0.2596248785654704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,3584,0.018449778358141582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,5120,0.03050044510099623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,128,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,768,0.03286400106218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,5120,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,3072,0.029556443293889362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,512,0.027519999278916255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,8192,0.019917334119478863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,65536,0.48256444931030273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,5120,0.07563289006551106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,65536,0.16465155283610025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,65536,0.27271466785007054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,1024,0.020211555891566806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,1024,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,8192,0.02884355518552992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,6144,0.010623999767833285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,7168,0.04965333475006951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,3584,0.029080000188615587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,3072,0.48120800654093426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,512,0.052517334620157875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,3072,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,6144,0.09449066718419392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,65536,0.06760533650716145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,2560,0.0135057775510682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,16384,0.13602488570743138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,51200,3.3268053266737194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,2048,0.22866490152147081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,32,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,51200,0.06298400296105279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,6144,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,51200,0.3849537902408176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,256,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,8192,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,8192,0.0910266637802124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,1536,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,12288,0.015917332635985482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,1536,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,64,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,6144,0.024658666716681585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,3584,0.01650311052799225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,1024,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,65536,0.6136870914035374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,8192,0.02160266704029507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,1536,0.030729777283138696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,2048,0.02996888756752014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,512,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,5120,0.08878933058844672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,512,0.005440000030729506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,12288,0.018168000711335074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,32,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,8192,0.05890933010313246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,3072,0.014936889211336771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,4096,0.03451111250453525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,1536,0.02036355601416694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,10240,0.05242844422658285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,128,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,8192,1.0125324461195204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,51200,0.06281866629918416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,7168,0.03387466735310025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,65536,1.8462329440646703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,7168,0.08932355377409194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,65536,0.09089422225952148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,5120,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,128,0.17052088843451607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,3584,0.011305777562989129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,512,0.1758355564541287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,12288,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,1536,0.09550933043162028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,768,0.06229333082834879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,5120,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,32,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,4096,0.03874933388498094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,64,0.02591200007332696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,12288,0.090865777598487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,256,0.015936000479592215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,16384,0.09056444300545587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,2048,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,3584,0.023082666926913794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,6144,0.018858666221300762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,3072,0.04782844583193461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,8192,0.013379555609491138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,5120,0.6656204329596626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,2560,0.019575110740131803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,1536,0.16871378156873915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,5120,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,12288,0.011338666909270816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,1536,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,5120,0.06059733364317152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,51200,0.27206312285529244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,128,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,6144,0.10316355360878839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,4096,0.07138399945365058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,10240,0.0912160012457106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,32,0.008910222185982598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,5120,0.16095644897884792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,16384,0.3728791077931722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,1536,0.03202755583657159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,7168,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,51200,3.4974596235487194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,3072,0.03028977910677592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,51200,0.07021511263317533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,2048,0.06098133325576782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,256,0.01148622234662374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,65536,1.6055386861165364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,8192,0.6729057629903158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,8192,0.03696977761056688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,2048,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,3584,0.03221244282192654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,32,0.018580444984965853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,1536,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,7168,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,768,0.00665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,6144,0.03491910960939195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,6144,0.03615555498335097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,256,0.005499555418888728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,768,0.0797377758555942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,512,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,512,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,10240,0.04720177915361193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,64,0.010635555618339114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,5120,0.01716800034046173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,10240,0.018654222289721172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,12288,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,5120,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,12288,0.026777777406904433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,3072,0.009880888793203566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,512,0.19404178195529512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,3072,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,128,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,3072,0.15922399361928305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,512,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,7168,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,2560,0.02126933303144243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,128,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,12288,0.22381154696146646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,128,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,51200,0.3640631039937337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,5120,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,2560,0.03559733430544535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,65536,0.4576791127522786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,16384,0.2582648860083686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,16384,0.017645332548353408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,1024,0.006364444477690591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,2560,0.012095111111799875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,64,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,3072,0.01202400028705597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,4096,0.02808533443344964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,10240,0.0269804447889328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,12288,0.07069600290722318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,3072,0.009465778039561378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,768,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,7168,0.04053599966896905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,8192,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,65536,0.06759555472267999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,8192,0.041139556301964655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,768,0.019725332657496136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,51200,0.10432800319459702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,5120,0.01668177710639106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,10240,0.025731555289692346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,65536,0.4844960106743707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,768,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,6144,0.01882577770286136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,256,0.0024293334119849736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,7168,0.00996888925631841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,10240,1.4522204928927953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,3584,0.036588443650139704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,2560,0.006017777654859755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,2560,0.02478755513827006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,65536,0.0694426695505778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,256,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,512,0.01349511080318027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,64,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,128,0.011733333269755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,5120,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,7168,0.12251555919647217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,16384,0.027271999253167048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,10240,0.01612977849112617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,1024,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,51200,4.1014993455674915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,5120,0.034801777866151594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,3072,0.013387555877367655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,3072,0.005782222168313132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,2560,0.13128356138865152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,51200,0.06468622552023993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,3584,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,51200,0.05887999799516466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,16384,0.014651555154058667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,65536,0.46626843346489805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,3072,0.00885333369175593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,65536,0.724914656745063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,128,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,4096,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,256,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,2048,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,512,0.009303111169073317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,2048,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,16384,0.11369600560930039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,128,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,6144,0.010274666878912184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,12288,0.2863493230607775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,10240,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,128,0.022509333160188463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,128,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,5120,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,51200,1.6013511021931965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,128,0.0284826656182607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,3072,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,1024,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,2560,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,51200,0.050624890459908374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,64,0.01889600025282966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,4096,0.011869333684444427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,2048,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,4096,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,10240,0.05129955543412102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,6144,1.4171910815768773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,16384,0.08155288961198595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,512,0.022208889325459797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,1536,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,4096,0.01015911085738076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,256,0.0041866666740841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,512,0.030517333083682593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,2048,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,256,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,3584,0.009093333449628618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,512,0.021324444148275588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,1536,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,1024,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,128,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,2560,0.012160888976520963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,6144,0.2790284421708849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,3072,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,32,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,6144,0.014069333672523499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,256,0.01442311041884952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,4096,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,2560,0.027452444036801655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,1024,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,12288,0.14021244313981798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,4096,4.18427997165256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,3072,0.19200355476803252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,7168,0.03476622369554307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,256,0.017128000656763714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,16384,1.8288959927029078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,3072,0.018000889155599806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,3584,0.01995822290579478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,2560,0.01442399952146742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,51200,1.9498382144504125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,5120,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,32,0.003162666741344664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,10240,0.06270488765504625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,1024,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,768,0.011872000164455838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,8192,0.11844533019595677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,2048,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,1024,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,32,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,6144,0.03163911236657037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,8192,0.033871110942628645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,4096,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,6144,0.4443022145165338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,128,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,3584,0.07439377572801378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,256,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,3072,0.012887111140622033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,10240,0.036906666225857206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,4096,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,3072,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,1536,0.020875554945733812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,7168,0.015053333507643806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,10240,0.07122755712933011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,12288,2.1562612321641708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,16384,0.5519626405504015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,8192,0.1848684416876899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,4096,0.09583022197087605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,768,0.004311999926964442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,5120,0.12686222129397923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,7168,0.011391110718250275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,64,0.0022026666750510535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,10240,0.01645866698688931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,3072,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,65536,0.2929244571261936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,768,0.0162453336848153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,2560,0.005432888865470886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,2048,0.014176888598336114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,128,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,7168,0.05287555522388882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,5120,0.03211022085613675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,12288,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,64,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,65536,0.13679822285970053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,16384,0.25152089860704213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,7168,0.03522933191723294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,6144,1.4021199544270833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,7168,0.1251200040181478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,6144,0.0192266669538286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,768,0.055480887492497764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,128,0.002255111104912228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,6144,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,7168,0.06618222263124254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,128,0.012442666623327466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,128,0.030064887470669214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,65536,0.08030933141708374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,5120,0.13028532928890652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,768,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,65536,0.3057333363427056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,768,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,5120,0.0983475579155816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,5120,0.0351839992735121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,10240,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,51200,1.1757048500908744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,256,0.10345066918267144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,65536,0.411589331097073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,2560,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,256,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,256,0.04150844282574124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,10240,0.017674666312005784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,12288,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,10240,0.02107644412252638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,256,0.017265778448846605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,8192,0.2992222309112549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,7168,0.070470220512814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,512,0.011509333219793109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,3584,0.03214933474858602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,2048,0.10685244533750747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,128,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,10240,0.042711112234327525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,1024,0.06027555465698242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,16384,0.037669334146711565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,512,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,6144,0.04312444395489163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,3584,0.009547555612193214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,1536,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,512,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,51200,3.8424729241265188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,6144,0.013435555828942193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,4096,0.009497777455382878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,6144,0.03897866606712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,256,0.03509333398607042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,32,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,12288,0.09286755323410034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,65536,0.21696710586547852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,65536,0.6663617557949489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,65536,0.5855209032694498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,1024,0.17451288965013292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,128,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,512,0.020121778051058452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,512,0.009915555516878763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,5120,0.8593733575608996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,768,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,8192,0.034656001461876765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,10240,0.048587554030948214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,10240,0.09110133515463935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,3584,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,1536,0.03302133414480422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,10240,0.037194665935304426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,768,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,5120,0.030123554997973975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,7168,0.008263111114501953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,4096,0.06240355306201511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,2048,0.10484177536434597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,6144,0.1149653328789605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,3072,0.014110222458839417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,51200,0.04752622379197014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,10240,0.008053333395057255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,64,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,64,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,8192,0.1372284491856893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,2560,0.033061332172817655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,512,0.01923644377125634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,6144,0.10286311308542888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,3584,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,12288,0.18864088588290742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,3584,0.011485333244005838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,7168,0.1311439938015408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,4096,0.004567111117972268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,1024,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,2048,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,1024,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,7168,0.01754666699303521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,65536,0.046859555774264865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,7168,0.04564088914129469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,5120,0.03138844503296746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,32,0.022454222043355305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,51200,0.07517511314815946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,8192,0.06487288739946154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,2048,0.018526222970750596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,3072,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,1024,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,1536,0.07550311088562012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,7168,0.06383911106321546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,2560,0.05901777744293213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,1024,0.0515342222319709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,10240,0.06771022081375122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,768,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,7168,0.17281244860755074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,256,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,5120,0.02683199942111969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,32,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,4096,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,16384,0.060995552274915904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,7168,0.071670220957862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,768,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,8192,0.07752177450391981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,768,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,1024,0.003175111073586676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,2048,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,6144,0.03375644485155741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,8192,0.024700444605615404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,65536,0.07492711146672566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,4096,0.03230844603644477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,4096,0.01461955573823717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,3072,0.006682666639486949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,5120,0.012675555215941535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,51200,0.03867644402715895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,64,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,65536,0.24796178605821398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,10240,0.04203733470704821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,5120,0.017874666386180453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,10240,0.017249777913093567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,12288,0.05032799972428215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,64,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,12288,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,1536,0.7822977701822916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,3072,0.035727110173967146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,65536,0.1848844422234429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,256,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,1024,0.012728888955381183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,7168,0.09107733435100979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,3584,0.015797333584891427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,16384,0.046160889996422656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,768,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,6144,0.032925334241655134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,5120,0.13622577985127768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,7168,0.027434666951497395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,2048,0.05754577451282077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,6144,0.15152888827853733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,12288,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,4096,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,2560,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,768,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,3072,0.007377777662542131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,65536,0.06805600060356988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,1024,0.005487999982304043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,256,0.02325155503220028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,51200,0.08599999878141616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,64,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,3072,0.009588444398509132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,8192,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,512,0.0070222218831380205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,128,0.03352355625894334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,128,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,768,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,512,0.18803111712137857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,6144,0.03776799970202976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,10240,0.020280889338917203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,2048,0.02330133318901062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,16384,0.11348977353837754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,65536,0.09883822335137261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,6144,0.029600001043743555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,16384,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,1024,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,3072,0.012650666965378655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,8192,0.5402542220221626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,5120,0.02515911062558492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,1024,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,6144,0.010800888968838586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,65536,0.6608737839592828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,8192,0.01921777758333418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,3584,0.01903199983967675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,128,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,2048,0.014739554789331226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,2560,0.014873777826627096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,7168,0.024933333198229473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,2048,0.03898222247759501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,2560,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,12288,0.017346666918860543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,10240,0.015764445066452026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,7168,0.058442665470970996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,128,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,768,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,1024,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,5120,0.012348444097571902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,2048,0.16668621699015299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,2048,0.016558221644825406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,2048,0.4251893361409505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,3584,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,12288,0.020425778296258714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,128,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,64,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,128,0.015989333391189575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,51200,0.6024799876742893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,51200,0.16398578219943577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,16384,0.025377778543366328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,51200,0.07630844248665704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,1024,0.02055999967787001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,4096,0.008418666819731394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,10240,0.02787644333309597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,3584,0.02438844409253862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,1024,0.0033884443756606844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,1536,0.003531555748648114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,128,0.012705778082211813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,3072,0.019938665959570143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,3072,0.028088887532552082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,3584,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,1536,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,12288,0.05490933193100823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,128,0.016887111796273124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,6144,0.010247999595271217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,12288,0.05701155794991387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,16384,3.430996365017361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,51200,0.19332622157202828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,2560,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,2560,0.02453244394726223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,65536,4.421566433376737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,512,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,3072,0.17872799767388237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,65536,0.12264711327022976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,65536,12.39329867892795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,10240,0.6070586840311686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,5120,0.05713599920272827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,12288,0.0392577780617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,7168,0.05282400051752726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,5120,0.12491733498043484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,4096,0.014859555496109856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,3072,0.06569511360592313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,5120,0.027228444814682007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,256,0.009789333575301701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,6144,0.033215999603271484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,3584,0.4111226664649115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,512,0.023578666978412207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,768,0.007870222131411234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,10240,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,16384,0.05372533202171326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,1536,0.018217777212460835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,2560,0.1046560009320577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,4096,0.019018666611777413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,512,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,12288,0.05357244279649523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,6144,0.010842666857772403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,128,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,51200,0.11288977993859185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,7168,0.013051555388503604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,51200,0.1838222212261624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,6144,0.01221333361334271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,3072,0.1462915605968899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,10240,0.09284177753660415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,65536,0.43815912140740293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,6144,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,6144,0.018686221705542672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,6144,0.009464888936943477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,6144,0.02166755497455597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,5120,0.017127111554145813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,51200,0.12026845084296332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,5120,0.023598222268952265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,512,0.00923111124171151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,65536,0.23340890142652723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,10240,0.05169511172506544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,768,0.006520888871616787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,5120,0.06560444169574313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,7168,0.016958221793174744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,128,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,3584,0.012869333227475485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,2560,0.033837334977255926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,256,0.052335109975602895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,128,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,5120,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,10240,0.02199111051029629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,2560,0.043712000052134194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,16384,0.020790222618314955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,7168,0.07323733303281996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,768,0.005774222314357758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,10240,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,16384,0.04704711172315809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,2048,0.03659822212325202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,32,0.04788177874353197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,10240,0.019852444529533386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,1024,0.003175111073586676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,65536,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,65536,0.39971110555860734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,2560,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,7168,0.01789955629242791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,8192,0.04746311240726047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,3584,0.38647998703850645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,512,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,1024,0.029008888536029395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,2048,0.011055111057228513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,1024,0.018942221999168396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,1024,0.016164445214801364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,65536,0.26507200135125053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,256,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,3072,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,6144,0.06286666790644328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,768,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,1024,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,7168,0.10610044664806789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,65536,0.17350310749477812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,128,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,3584,0.02927288744184706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,768,0.06842488712734647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,3072,0.007368889119890001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,3584,0.010330666270520952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,65536,0.1632257832421197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,768,0.012019555601808759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,64,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,2560,0.01886488828394148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,1536,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,2048,0.00481599983241823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,2560,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,128,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,65536,0.1454977724287245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,7168,0.03773422373665704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,2048,0.009851555857393477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,16384,0.1373573276731703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,64,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,8192,0.04103911254141066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,64,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,4096,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,7168,1.6729520161946614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,1024,0.01608711149957445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,12288,0.038402666648228966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,3584,0.05561511384116279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,2048,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,2048,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,2048,0.017830222845077515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,1536,0.015765332513385348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,3072,0.014662222729788886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,12288,0.0639537771542867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,51200,0.05133777856826782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,2560,0.0669599970181783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,1536,0.030921777089436848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,768,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,12288,0.03686577743954129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,10240,0.03569777806599935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,1536,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,10240,0.0420702232254876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,2048,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,3072,0.031210667557186548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,5120,0.00868888861603207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,4096,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,16384,0.21246311399671766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,3072,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,12288,1.1808044645521376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,6144,0.02021511064635383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,3584,0.5684995651245117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,51200,0.25214844279819065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,51200,0.13315378295050725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,2560,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,7168,0.011578666667143503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,2048,0.020986666282018025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,32,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,1536,0.008901333643330468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,2048,0.04786577820777893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,7168,0.03552177879545424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,5120,0.0047688889834615915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,65536,0.14400800069173178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,2560,0.02216533323129018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,64,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,768,0.037779556380377874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,1536,0.009789333575301701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,16384,0.25667643547058105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,10240,0.041434665520985924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,3072,0.06371200084686279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,32,0.003903111235962974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,1024,0.01716888944307963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,10240,0.2021324502097236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,7168,0.02426044477356805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,12288,0.4704586664835612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,3584,0.0635075569152832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,512,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,64,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,2048,0.021016889148288306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,65536,0.1297395494249132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,51200,0.0606328911251492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,65536,0.12091999583774143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,12288,0.06398489077885945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,32,0.0063591111037466265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,12288,0.015228443675571017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,6144,0.05673778057098389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,1024,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,4096,0.0295688874191708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,128,0.009718221922715506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,65536,0.03534844517707825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,32,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,16384,0.01648888819747501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,16384,0.04116888840993246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,1024,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,51200,0.266628450817532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,128,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,1536,0.006892444358931647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,12288,0.09224177731408013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,16384,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,128,0.005161777966552311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,6144,0.01680088871055179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,256,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,4096,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,8192,0.02776799930466546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,10240,0.01498311095767551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,12288,0.01701866587003072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,1536,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,3584,0.007870222131411234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,12288,0.8102684550815158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,10240,0.04239111145337423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,1024,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,4096,0.02605777813328637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,2560,0.03325155708524916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,6144,0.016419554750124615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,10240,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,1536,0.05584355857637194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,512,0.2217226690716214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,5120,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,3584,0.05267822080188327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,10240,0.028135998381508723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,2560,0.06984000073538886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,256,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,32,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,768,0.018968888454967074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,32,0.036991112762027316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,7168,0.010871999793582492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,7168,0.020590222544140287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,10240,0.04186488853560554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,6144,0.030673778719372217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,1024,0.022146667043368023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,1024,0.04317333300908407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,1536,0.015463110473420886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,3584,0.03288977675967746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,256,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,5120,0.021129777034123737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,3584,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,16384,0.04492977923817105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,4096,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,512,0.015083556373914083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,16384,2.46778318617079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,10240,0.011767999993430244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,16384,0.07366222143173218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,8192,0.14529866642422146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,2048,0.019066666563351948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,8192,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,1536,0.06663111183378431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,256,0.019633778267436557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,3072,0.019527110788557265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,256,0.01220533334546619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,65536,2.4295546213785806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,65536,0.25984621047973633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,1024,0.00565155554148886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,512,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,12288,0.011924444801277585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,51200,0.6694080034891764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,65536,0.23006399472554526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,5120,0.30976888868543834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,512,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,512,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,16384,0.1367315583758884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,12288,0.037003556887308754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,51200,0.12191822793748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,2560,0.016207110550668504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,512,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,51200,0.2046346664428711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,3584,0.011491555306646558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,65536,0.3410968780517578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,4096,0.013219555219014486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,256,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,12288,0.01920977731545766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,7168,0.10347466336356269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,3584,0.05760622024536133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,1536,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,768,0.18499199549357095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,10240,0.08800444338056777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,12288,0.018573333819707234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,2048,0.037217779292000666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,7168,0.02546311087078518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,6144,0.019471999671724107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,16384,0.02075466679202186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,768,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,1024,0.05269777774810791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,65536,7.000208112928603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,2048,0.02272355556488037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,51200,0.059623108969794385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,65536,0.09983022345436944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,12288,0.04248266749911838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,32,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,1536,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,256,0.021881777379247878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,4096,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,3072,0.031071109904183283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,3584,0.01887822151184082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,12288,0.020891555481486853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,8192,0.013882666826248169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,3584,0.010985777609878115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,32,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,4096,0.046451555358039014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,12288,0.042423999971813627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,32,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,51200,10.041247049967447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,3584,0.047729776965247266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,5120,0.01754133403301239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,2048,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,1024,0.005977777971161737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,65536,0.07086844576729669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,4096,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,7168,0.028174221515655518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,32,0.0019537777536445195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,768,0.004927111168702443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,3072,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,2048,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,3584,0.02771200074089898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,1536,0.028359999259312946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,12288,0.031344887283113264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,1536,0.013440888788965015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,16384,0.09281422032250299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,768,0.019443555010689628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,65536,0.07090400324927436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,3072,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,768,0.01791288952032725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,4096,0.011133333047231039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,7168,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,12288,0.030227555168999568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,256,0.012715555727481842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,3072,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,5120,0.3318720128801134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,5120,0.01889333294497596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,8192,0.016927111479971144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,4096,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,3072,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,8192,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,7168,0.02202488813135359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,16384,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,10240,0.2077102263768514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,65536,0.034625778595606484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,128,0.010203555226325989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,12288,0.017096888687875535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,65536,1.0905698140462239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,2048,0.05854755640029907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,12288,0.08851289086871678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,3072,0.06574489010704888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,5120,0.08066755533218384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,768,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,5120,0.45330222447713214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,1536,0.004925333377387789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,65536,0.0745039979616801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,2048,0.023689778314696416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,65536,0.06628888845443726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,768,0.018283555905024212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,768,0.014597333139843412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,1024,0.015455111861228943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,1024,0.1702880064646403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,1536,0.017166222135225933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,16384,0.026047999660174053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,6144,0.03560622202025519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,51200,0.04907288816240099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,10240,0.03269955515861511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,2560,0.015034667319721647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,1536,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,51200,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,2560,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,3072,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,8192,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,1536,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,8192,0.01408266690042284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,12288,0.09814933273527358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,10240,0.02887466549873352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,1024,0.0427706672085656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,512,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,16384,0.057793776194254555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,512,0.010216889282067617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,2560,0.0075528886583116316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,8192,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,128,0.02113955550723606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,51200,0.03992444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,7168,0.2663608921898736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,5120,2.6726356082492404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,65536,0.4215804470909967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,5120,0.009285333255926767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,65536,0.05652888615926107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,8192,0.23550311724344888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,2048,0.08878933058844672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,2048,0.014953777194023132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,3072,0.02979288829697503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,512,0.012858666479587555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,51200,0.4701111051771376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,2048,0.018124444617165458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,2560,0.03278222348954942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,10240,0.015273777974976433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,2560,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,3072,0.0834346678521898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,1024,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,64,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,128,0.0019244444039132858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,4096,0.017498667041460674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,3072,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,3584,0.008578666382365758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,256,0.009303111169073317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,51200,0.5184177822536892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,10240,0.13812355200449625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,10240,0.164337780740526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,1536,0.06394400199254353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,6144,2.26236809624566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,51200,0.018515555395020377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,768,0.023992000354660883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,10240,0.014885332849290637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,1024,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,3072,0.007690666450394525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,256,0.015288889408111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,2048,0.018613333503405254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,65536,0.07064000103208753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,5120,0.017161778277821012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,1024,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,6144,0.02778488728735182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,3072,0.015052444405025907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,1024,0.10388000143898858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,3072,0.020124443703227572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,6144,0.06157244576348198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,3072,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,64,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,5120,0.05916088819503784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,3584,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,4096,0.01891644464598762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,51200,0.0708835588561164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,8192,0.03360711203681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,3584,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,5120,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,768,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,65536,0.05913066864013672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,4096,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,12288,0.10277866654925877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,12288,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,8192,0.1047075523270501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,51200,1.5727173487345378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,768,0.007672888537247975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,7168,0.14395288626352945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,8192,0.014328888720936246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,4096,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,10240,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,768,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,12288,0.027311111489931744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,4096,0.008616888688670265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,7168,0.023990222149425085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,8192,0.20175466272566053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,3072,0.08028622468312581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,5120,0.025809778107537165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,4096,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,12288,0.0886311133702596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,512,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,6144,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,5120,0.012862222062216865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,64,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,1536,0.008919110728634728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,10240,0.01329955541425281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,2048,0.009397333694828881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,32,0.00609422226746877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,1536,0.02010222276051839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,512,0.012270222107569376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,128,0.010514666636784872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,51200,0.048911998669306435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,768,0.0269128895468182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,256,0.05228711167971293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,3584,0.02168088820245531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,256,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,768,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,128,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,16384,0.14628622266981336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,768,0.020248000820477802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,65536,1.1644826465182834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,768,0.0025600000388092464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,10240,0.015466666883892484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,6144,0.02420799930890401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,768,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,4096,0.02128977742460039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,1536,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,8192,0.10337511036131118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,1536,0.16098933749728733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,512,0.015410666664441427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,10240,0.044997334480285645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,256,0.027274666561020747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,768,0.017907554904619854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,32,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,2560,0.00755555596616533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,128,0.027715555495686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,5120,0.017822222577200997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,3072,0.03223022156291538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,3072,0.03781422310405307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,7168,0.021957332889238994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,7168,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,32,0.006180444525347815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,2560,0.3946231206258138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,65536,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,256,0.09089155329598321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,8192,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,65536,0.1359377834531996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,2048,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,256,0.9043501747979058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,2048,0.011647111011876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,1024,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,10240,0.07699555820888944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,5120,0.08795555432637532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,8192,0.33249955707126194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,3072,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,4096,0.24295555220709908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,5120,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,10240,0.09433688720067342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,2560,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,2560,0.03490399983194139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,10240,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,64,0.002032888846264945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,128,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,512,0.011926222178671094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,12288,0.5320933130052354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,5120,0.16554578145345053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,256,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,8192,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,128,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,6144,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,1024,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,256,0.019145778483814664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,65536,0.49430402119954425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,4096,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,768,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,7168,0.023052444060643513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,51200,0.10578666792975532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,6144,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,2048,0.13110844294230142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,7168,0.013737777868906656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,3584,0.023688889212078516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,128,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,16384,0.6085253291659886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,10240,0.6135777897304958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,6144,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,3072,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,12288,0.054413331879509814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,51200,0.07414666811625163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,1024,0.018057778477668762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,512,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,128,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,8192,0.23452888594733345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,2048,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,7168,0.022037333912319605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,2048,0.05977511405944824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,12288,0.01720177796151903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,2048,0.012741333080662621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,5120,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,512,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,51200,0.037979556454552546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,12288,0.017709333035680983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,256,0.013214222258991666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,128,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,65536,0.5406497849358453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,1024,0.020671111014154222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,128,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,5120,0.023183110687467787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,16384,0.015844444433848064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,4096,0.038336889611350164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,768,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,128,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,128,0.015841777125994366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,6144,0.008479110896587372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,51200,0.3068471219804552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,768,0.01828977796766493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,16384,0.10926044649548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,65536,0.15602844291263157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,4096,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,5120,0.009759110709031424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,65536,0.06615289052327473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,10240,0.23975910080803764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,768,0.005700444595681296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,12288,0.03459466828240289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,16384,0.09904622369342381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,12288,0.0737724436653985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,12288,0.029616001579496596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,10240,0.03707022137112088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,6144,0.010316444767846001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,1536,0.019866666860050626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,3584,0.04460710949367947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,51200,0.8287750879923502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,12288,0.08219822247823079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,512,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,3584,0.016220443778567843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,2048,0.010445333189434476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,4096,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,51200,0.11537155840131973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,512,0.007154666715198093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,512,0.01591822173860338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,7168,0.014581332604090372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,10240,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,768,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,8192,0.07778311438030668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,512,0.025179555018742878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,32,0.10461422469880845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,2048,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,5120,0.022114665971861944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,3584,0.007744889292452071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,3072,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,128,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,2048,0.028547555208206177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,1024,0.00904977818330129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,12288,0.16212799814012316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,1536,0.013831111292044321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,65536,0.1614746650060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,4096,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,512,0.002209777840309673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,128,0.012479999827014076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,2560,0.03614844547377692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,5120,0.01220266686545478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,5120,0.7677849133809408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,1024,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,32,0.013817778064144982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,3584,0.04432622260517544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,2560,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,6144,0.008823999928103553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,4096,0.07384177711274889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,8192,0.13879556126064724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,64,0.006052444378534953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,2048,0.006699555450015598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,51200,0.36223822169833714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,1536,0.012117333710193634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,8192,0.017704889178276062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,16384,0.05754577451282077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,2560,0.0624542236328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,12288,0.027635556128289964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,1024,0.030861334668265447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,32,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,10240,0.1625235610538059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,5120,0.012588444683286877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,16384,0.04443733228577507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,256,0.04655111167165968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,1536,0.005666666560702854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,10240,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,1536,0.0047564444442590075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,6144,0.04595999916394552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,1024,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,1536,0.017263111140992906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,3072,0.05656177467770047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,8192,0.1363111072116428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,128,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,12288,0.042043556769688926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,6144,0.012283555335468717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,128,0.08254044585757785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,1024,0.05818844503826565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,768,0.0053484443989064955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,1536,0.02742133372359806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,6144,0.00962044464217292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,64,0.009841777384281158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,2560,0.4242497815026178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,5120,0.05249155561129252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,8192,0.04532266656557719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,512,0.004250666747490565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,2560,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,5120,0.01648177703221639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,4096,0.010849778023031024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,1024,0.07095999850167169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,512,0.012793777717484368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,65536,0.1290257771809896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,512,0.00720711135201984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,7168,0.2455271085103353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,6144,0.1367031070921156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,768,0.0039057777159743835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,256,0.004651555584536659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,2560,0.013198222551080914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,256,0.022268444299697876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,2048,0.019161777363883126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,10240,0.10893866750929092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,2560,0.06177510817845663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,5120,0.01700177788734436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,128,0.08228355646133423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,768,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,128,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,4096,0.048954665660858154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,16384,0.18502044677734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,1536,0.00618488879667388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,512,0.02648266653219859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,12288,0.02898400028546651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,8192,0.03771022293302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,10240,0.8419929080539279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,1024,0.024882665938801233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,65536,0.9143786960177951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,2048,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,4096,0.2249831093682183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,10240,0.03158400124973721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,768,0.01995377739270528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,6144,0.13623555501302084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,51200,0.1653697755601671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,768,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,12288,0.14852000607384577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,256,0.007730666961934831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,2048,0.017112889223628573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,1024,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,65536,0.8333466847737631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,32,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,768,0.00959288908375634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,51200,0.353765328725179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,768,0.00902844468752543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,10240,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,768,0.017916444275114272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,4096,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,7168,0.20284799734751383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,16384,1.1469892925686307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,5120,0.020672000116772123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,51200,0.8954337437947592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,5120,0.06404622395833333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,6144,0.04976888828807407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,1536,0.023418666587935552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,2048,0.045392000012927584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,6144,0.014183110660976835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,2560,0.020763556162516277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,5120,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,2048,0.08589510785208808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,256,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,6144,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,16384,0.021635555558734473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,51200,8.15573459201389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,768,0.017874666386180453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,4096,0.03178400132391188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,2048,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,51200,0.9072106679280599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,128,0.019907555646366544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,7168,0.009837333526876237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,2048,0.03896533449490865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,7168,0.025309332542949255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,3072,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,64,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,1536,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,1536,0.0039013334446483185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,64,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,768,0.028098666005664404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,3584,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,7168,0.042656888564427696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,64,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,256,0.01678311162524753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,65536,0.06379022200902303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,2560,0.25828801261054146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,6144,0.11161243915557861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,7168,0.029286222325430974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,2560,0.011724444727102915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,5120,0.022042666872342426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,5120,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,1536,0.33190666304694283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,2048,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,768,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,6144,0.04173688756095039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,512,0.014730667074521383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,3072,0.0956453349855211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,10240,0.02908888790342543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,128,0.020447999238967896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,3072,0.03000266684426202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,8192,0.4715333514743381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,16384,0.021819555097156104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,256,0.003549333247873518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,10240,0.01714933415253957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,3584,0.010435555544164447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,7168,0.16108177767859563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,65536,0.11921421686808269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,16384,53.15777587890625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,12288,0.03174577818976508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,8192,0.1387315591176351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,10240,0.15317422813839382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,6144,0.06193777587678698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,7168,1.1613332960340712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,1536,0.016599999533759225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,1024,0.014315555493036905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,7168,0.10100266668531631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,768,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,2560,0.16453244951036242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,65536,1.035721778869629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,6144,0.011113777756690979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,256,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,1024,0.09711822536256577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,4096,0.17326222525702584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,4096,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,2048,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,1536,0.003909333298603694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,10240,0.029935999049080744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,5120,0.024591111474566992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,2560,0.1328631109661526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,3072,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,8192,0.0358515547381507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,2048,0.019691555036438834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,512,0.00963199966483646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,12288,0.16781777805752227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,10240,0.019718221492237516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,4096,0.0760542220539517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,8192,0.27340179019504124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,3584,0.013305777476893531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,256,0.017878222796652053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,10240,0.6160711182488335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,7168,0.03526666760444641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,10240,0.1606880029042562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,51200,0.05763733386993408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,2048,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,3072,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,4096,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,3584,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,64,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,128,0.029779556724760268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,5120,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,5120,0.024315555890401203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,16384,0.043493333790037364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,8192,0.014853333433469137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,8192,0.09556177589628433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,6144,0.052468445565965444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,8192,0.011428444749779172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,256,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,3584,0.046055999067094594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,1536,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,1024,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,3584,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,2048,0.006471111128727595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,7168,0.020055111911561754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,8192,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,128,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,768,0.02459822263982561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,2048,0.06378044684727986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,64,0.006680888848172293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,6144,0.040222220950656466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,65536,1.2491653230455186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,10240,0.24824089474148223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,3584,0.6934186617533366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,1024,0.012191111014948951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,3584,0.12708711624145508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,256,0.0229795558585061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,7168,0.04879200127389696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,7168,0.034620443979899086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,768,0.00360000009338061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,65536,0.2491146723429362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,8192,0.01643288963370853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,3584,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,8192,0.02996088729964362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,3072,0.2170462210973104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,12288,0.058279110325707324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,7168,0.014315555493036905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,6144,0.015881778465376962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,65536,0.07932355668809679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,6144,0.25552622477213544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,3072,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,4096,0.03514133228196038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,65536,6.505643632676866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,3584,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,8192,0.056278222137027316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,768,0.04437866806983948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,4096,0.012308444413873883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,768,0.010126222338941362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,6144,0.008719111482302347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,12288,0.043960889180501304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,4096,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,12288,0.0705359975496928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,768,0.01510933372709486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,2048,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,128,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,4096,0.6120622422960069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,5120,0.012973333398501078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,2048,0.17492089006635878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,3584,0.06536355283525255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,3072,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,2560,0.08159199688169691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,768,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,1024,0.048735999398761325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,5120,0.11007822222179836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,65536,0.0850284430715773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,32,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,5120,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,5120,0.038338667816585965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,65536,0.08146666818194918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,2560,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,10240,0.012602667013804117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,3584,0.7277733484903971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,4096,0.07581066422992282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,256,0.0032559999575217566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,768,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,256,0.009114666945404476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,3072,0.021856889128684998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,128,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,3072,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,2048,0.028780443800820246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,12288,0.049583110544416636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,2560,0.059360000822279185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,6144,0.0440631111462911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,256,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,6144,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,12288,0.012913777596420713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,4096,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,2048,0.0626879996723599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,2048,0.030942221482594807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,51200,0.06500799788369073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,5120,0.07669511106279162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,256,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,2560,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,51200,0.2589920096927219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,8192,0.030003554291195337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,128,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,128,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,768,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,16384,0.3246133327484131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,768,0.007111111448870764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,5120,0.00869599978129069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,7168,0.05228799912664625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,3072,0.02232711182700263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,7168,5.6836658053927955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,8192,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,5120,0.011517333487669626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,3584,0.07959733406702678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,768,0.0031946665710873077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,12288,0.017456889152526855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,6144,0.03349511159790887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,512,0.01519377695189582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,512,0.02535022298494975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,3584,0.13507378101348877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,2048,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,128,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,32,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,16384,0.1017120016945733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,4096,0.019392000304328073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,2560,0.2094515562057495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,65536,0.21792622407277426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,8192,0.008918222453859117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,12288,0.02055111030737559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,256,0.03363555669784546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,1024,0.002663111107216941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,51200,0.5463351143731011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,7168,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,5120,0.016247999336984422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,65536,0.055997331937154136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,65536,0.023631999890009563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,128,0.014495111174053617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,6144,0.017712000343534682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,10240,0.00851822230550978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,3584,0.047714667187796704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,5120,0.0199617776605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,1536,0.0042257776690853965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,65536,3.432816823323568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,3072,0.032227555910746254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,51200,0.06558399730258517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,1536,0.08605155679914687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,5120,0.08980977535247803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,2560,0.07981955342822604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,128,0.004211555752489301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,10240,0.07135111093521118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,4096,0.05622400177849663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,7168,0.010750222537252637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,512,0.05324800146950615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,51200,0.6609608862135146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,16384,0.11394222577412923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,16384,0.08182222313351102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,2560,0.02091022166940901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,1024,0.034550223085615374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,51200,0.13878666030036077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,3584,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,3584,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,12288,2.4437323676215277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,256,0.01016266644001007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,2560,0.00702133360836241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,65536,0.08948355250888401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,128,0.019909333851602342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,16384,0.027142221728960674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,1024,0.019468444916937087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,5120,0.034303999609417386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,3584,0.04201066825124952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,32,0.09546489185757107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,128,0.010494222243626913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,51200,0.10650933451122707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,8192,0.05720444520314535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,3584,0.030232889784706965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,65536,0.08966933356391059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,6144,0.018231999542978074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,7168,0.41221777598063153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,4096,0.00792266676823298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,3072,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,4096,0.039525333378050066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,1536,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,256,0.013529777526855469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,6144,0.027148443791601393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,3072,0.05299644337760078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,16384,0.0742959976196289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,64,0.004001777619123459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,2048,0.02437066700723436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,3072,0.032454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,2048,0.010775110787815519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,6144,0.4927413198682997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,5120,0.01201777739657296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,1536,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,3072,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,8192,0.025612443685531616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,3584,0.012599111431174807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,16384,0.08682222498787774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,3584,0.022966222630606756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,64,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,3072,0.0427706672085656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,1536,0.07573955588870578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,65536,0.21985510985056558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,2560,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,768,0.0058355554938316345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,8192,0.24692710240681967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,4096,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,4096,0.01997955474588606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,16384,0.06084089146720039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,2560,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,16384,0.0413333343134986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,10240,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,5120,0.07246400250328912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,7168,0.02509333358870612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,2048,0.016901334126790363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,65536,0.46211735407511395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,5120,0.014075555735164218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,256,0.012305777933862476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,3584,0.02327377763059404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,16384,0.00565422202150027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,10240,0.18822666009267172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,256,0.02611555490228865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,3584,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,51200,0.14042399989234075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,1536,0.020672889219390023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,16384,0.02753511071205139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,10240,0.021394666698243883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,65536,0.35430312156677246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,6144,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,6144,0.035000890493392944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,3072,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,12288,0.05189777745140923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,2560,0.01219288922018475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,2048,0.022170666191313002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,65536,0.19828444057040742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,2048,0.01108088923825158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,8192,0.07869333028793335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,3072,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,1024,0.033273776372273765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,8192,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,1536,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,7168,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,64,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,16384,0.061456885602739125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,2560,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,5120,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,3584,0.01963555481698778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,2048,0.004253333227501975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,1024,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,1024,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,5120,0.010479999913109673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,65536,0.0627617769771152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,10240,1.8687474992540147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,3584,0.08880711264080471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,65536,4.380407969156901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,5120,0.02774577836195628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,256,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,7168,0.009537777966923183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,512,0.02544533378548092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,3584,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,8192,0.48883644739786786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,7168,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,16384,0.38766932487487793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,128,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,51200,0.04643288917011685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,12288,1.7159137725830078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,3584,0.018199111024538677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,4096,0.1694880061679416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,1536,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,1024,0.011866666376590729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,64,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,128,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,65536,1.6814311345418294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,1536,0.026375111606385972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,5120,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,512,0.008193777667151557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,128,0.006708444406588872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,6144,0.02155911094612545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,12288,0.056403557459513344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,768,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,7168,0.020012444920010038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,2560,0.014112888111008538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,512,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,8192,0.02749155627356635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,12288,0.017864000466134813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,6144,0.009848888549539778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,32,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,6144,0.13861778047349718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,1536,0.055544889635509916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,16384,0.06586133109198676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,256,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,768,0.002557333351837264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,10240,0.037103109889560275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,51200,0.7542568842569987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,3072,0.007867555651399825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,3584,0.12400000625186497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,128,0.014849777022997538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,65536,0.17263910505506727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,65536,0.06299999687406752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,51200,0.28680266274346244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,1024,0.012450666891203986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,1536,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,65536,0.09319733248816596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,65536,0.02516533268822564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,5120,0.06356177727381389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,768,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,3584,0.40511110093858504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,4096,0.009684444301658207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,1536,0.12379288673400879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,10240,0.03598577777544657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,16384,0.07847377989027235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,65536,0.411514679590861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,65536,0.07679555813471477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,4096,0.10527466403113471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,128,0.021205332544114854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,512,0.015578667322794596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,2560,0.01734133395883772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,256,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,256,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,2560,0.020817778176731534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,6144,0.18667644924587676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,4096,0.020836444364653692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,10240,0.042724443806542285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,256,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,256,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,51200,0.05237333310974968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,7168,0.047968000173568726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,1024,0.026724444495307073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,7168,0.16856800185309517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,6144,0.04520533482233683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,768,0.01723022262255351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,51200,8.316574096679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,3584,0.025076443950335186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,128,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,3072,0.007815111014578078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,256,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,768,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,2048,0.4067093266381158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,768,0.015917332635985482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,256,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,7168,0.010599111517270407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,2048,0.004181333300140169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,1024,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,7168,0.01872977779971229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,1024,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,3584,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,3584,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,3072,0.02847466535038418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,10240,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,3584,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,6144,0.03741600116093954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,1536,0.022728888524903193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,3584,0.04320444332228767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,6144,0.019546666079097323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,51200,5.31291749742296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,16384,0.11956532796223958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,64,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,768,0.06045066648059421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,1536,0.01861422260602315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,6144,0.030393779277801514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,128,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,768,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,128,0.006456888798210356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,8192,0.036335110664367676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,32,0.01075644459989336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,64,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,1024,0.03699377841419644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,3584,0.07767377959357367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,7168,0.029430223835839167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,3584,0.15082489119635686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,1024,0.004871110949251386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,6144,0.02183644473552704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,768,0.006113777971929974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,3584,0.16814400090111625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,64,0.02591466738118066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,4096,0.02084622283776601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,2560,0.007486222518814935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,2048,0.1014595561557346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,10240,0.012127111355463663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,64,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,3072,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,2560,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,1536,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,512,0.003491555651028951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,2048,0.01818933255142636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,768,0.024188444018363953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,7168,0.004865777575307422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,16384,0.031874666611353554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,3072,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,256,0.02351377738846673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,51200,0.05145066645410326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,7168,0.10144711203045315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,2560,0.09959110948774551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,256,0.008054222497675154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,64,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,16384,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,3072,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,2048,0.008035555481910706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,4096,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,1536,0.00619911112719112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,256,0.0053013331360287136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,1024,0.01587288909488254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,5120,0.14938399526807997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,6144,0.0499253339237637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,4096,0.029836446046829224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,768,0.002924444360865487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,3072,0.1879111131032308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,2560,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,8192,0.01850311126973894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,1536,0.11885955598619248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,768,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,2560,0.012895111408498553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,4096,0.04059911105367873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,4096,0.04917155702908834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,768,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,51200,13.028650071885851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,6144,0.011931555966536203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,4096,0.04552444484498766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,64,0.017288888494173687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,3584,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,10240,0.07026755809783936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,1024,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,2560,0.0074106670088238185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,3072,0.010471111370457543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,6144,0.009515555368529426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,1536,0.015980444020695157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,256,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,6144,0.9350106981065539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,3072,0.02100444502300686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,51200,0.5214604271782769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,3584,0.0667111078898112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,5120,0.11087644762463039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,128,0.01752355529202355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,1024,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,768,0.00563733321097162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,51200,0.44544442494710285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,512,0.05091377761628893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,16384,0.025091555383470323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,4096,0.08326844374338786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,3072,0.06844711303710938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,2048,0.015085332923465304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,4096,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,6144,0.05876977576149834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,128,0.016166221764352586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,256,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,2560,0.012352888782819113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,65536,0.19035555256737602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,256,0.006887110984987683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,4096,0.016919111212094624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,4096,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,32,0.03819200065400865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,128,0.005305777821275923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,65536,0.06316533353593615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,3072,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,64,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,10240,0.14730755488077799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,128,0.0053075556125905775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,256,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,3584,0.03678755627738105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,128,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,32,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,10240,0.07411555449167888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,768,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,256,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,512,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,8192,0.014274666706720987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,3584,0.015912888778580558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,8192,0.08098577790790133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,2048,0.01640888883007897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,10240,0.032032890452278986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,4096,0.0243848893377516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,128,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,3072,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,2048,0.02442399991883172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,32,0.03483555714289347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,3072,0.012808000048001608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,5120,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,768,0.119649781121148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,8192,0.03252799974547492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,3072,0.018777777751286823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,65536,0.2358293268415663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,4096,0.016328889462682936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,2048,0.020192889703644645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,3584,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,256,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,2560,0.6115253236558702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,16384,0.27383022838168675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,3584,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,7168,0.02165422174665663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,3584,0.0355333321624332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,65536,11.382916768391928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,128,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,16384,0.135843555132548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,256,0.025011556016074285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,4096,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,768,0.02200266718864441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,5120,0.024824889169798955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,512,0.029547555579079524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,8192,0.034690668185551964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,256,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,1024,0.038189331690470375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,16384,0.015822221835454304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,12288,0.30206667052374947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,64,0.0961386693848504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,7168,0.014555555250909595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,5120,0.009288888838556077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,32,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,51200,0.8033324347601997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,7168,0.01092266705301073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,2048,0.027383999692069158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,7168,0.06401155392328899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,8192,0.031868444548712835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,6144,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,2560,0.6056613392300075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,2048,0.011533333195580376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,32,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,1536,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,51200,0.06328711244795057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,65536,1.4670195049709742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,4096,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,1536,0.022962666220135156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,51200,3.176007164849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,2048,0.004513777792453766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,2048,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,2560,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,1536,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,16384,0.1512275536855062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,4096,0.01956977778010898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,4096,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,4096,0.026629333694775898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,51200,0.1055519978205363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,3072,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,5120,0.037997331884172224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,2048,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,7168,0.18181688255733916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,8192,0.024117334021462336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,1536,0.06458489100138347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,768,0.0032302221904198327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,3072,0.0883457793129815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,10240,1.7935617234971788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,51200,0.15346488687727186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,32,0.006681777950790193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,1024,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,128,0.01930488811598884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,10240,0.01720622181892395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,5120,0.009420444567998251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,16384,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,2048,0.04179822074042427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,32,0.005698666804366642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,512,0.024687111377716064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,3584,0.009581333233250512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,2560,0.03568622138765123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,12288,0.05338755581114027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,2048,0.00816177742348777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,3072,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,16384,0.023754666248957317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,3584,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,16384,5.109114752875434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,2048,0.00925155563486947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,2560,0.010949333508809408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,51200,14.552671644422743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,768,0.022470222579108343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,768,0.004963555683692296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,16384,1.1499733395046658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,3584,0.008787555827034844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,3584,0.007817777494589487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,4096,0.013551111022631327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,2048,0.07357688744862874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,3584,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,7168,0.43441955248514813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,3072,0.17070933183034262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,3072,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,512,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,128,0.004353777815898259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,512,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,1536,0.09070399734708999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,2560,0.05401510993639628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,65536,0.12080621719360352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,5120,0.5962186919318305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,2560,0.05685422155592176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,8192,0.20699466599358451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,768,0.17750310897827148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,8192,0.03312355611059401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,7168,0.03706489006678263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,256,0.022844443718592327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,3072,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,8192,0.06488711304134794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,768,0.007400000260935889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,8192,0.019127999742825825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,1024,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,1536,0.00812977800766627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,4096,0.01995822290579478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,4096,0.044621331824196704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,2048,0.01819022165404426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,8192,0.02460799945725335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,768,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,7168,0.020976000361972388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,51200,0.37997065650092227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,10240,0.03249955508444045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,51200,0.12499022483825684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,256,0.024076445235146418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,3584,0.02000711030430264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,32,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,1536,0.014367111855083041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,7168,0.18509156174129912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,10240,0.049125333627065025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,2560,0.03569066524505615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,128,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,256,0.015167999598715039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,51200,1.1599413553873699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,2560,0.1444071133931478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,2048,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,128,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,6144,0.012816888590653738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,65536,0.3947910997602675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,5120,0.05803910891215006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,12288,0.23024710019429526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,1024,0.010209778116808997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,768,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,8192,0.06150488720999824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,1024,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,512,0.016851555969980028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,4096,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,10240,0.015432000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,51200,0.15607466962602404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,7168,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,2560,0.20626489321390787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,7168,0.060560888714260526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,2560,0.05679111348258125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,4096,0.5284346474541558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,64,0.0019351111518012153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,512,0.014441778262456259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,2048,0.014576888746685453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,12288,4.00163099500868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,768,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,32,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,51200,0.07925155427720812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,4096,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,7168,0.3893875545925564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,2560,0.007497777541478475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,2560,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,8192,0.10553066598044501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,768,0.005475555443101459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,256,0.009684444301658207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,5120,0.03686666819784377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,4096,0.010647999743620554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,4096,0.12135288450453018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,16384,0.05186577637990316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,7168,0.03384533193376329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,256,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,3584,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,8192,0.045687112543318004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,7168,0.01961599952644772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,65536,3.3073946634928384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,3584,0.014011555247836642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,12288,0.0946542223294576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,768,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,2048,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,65536,0.3401795493231879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,3584,0.09877689017189874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,64,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,12288,0.07046222024493747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,6144,0.03151999910672506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,2560,0.028404444456100464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,768,0.054303109645843506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,12288,0.1911777787738376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,1536,0.40604445669386124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,3072,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,5120,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,12288,0.018623999423450895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,2048,0.012862222062216865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,768,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,512,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,2560,0.027211555176311072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,512,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,7168,0.040024889840020075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,1536,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,512,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,768,0.035535110367669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,3072,0.007367110914654202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,2560,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,4096,0.028443555037180584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,6144,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,2560,0.046632889244291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,10240,0.07095999850167169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,3584,0.07154399818844266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,6144,0.027706666125191584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,65536,0.34837155871921116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,1536,0.03499555587768555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,65536,1.3557075924343531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,10240,0.03926755653487311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,4096,0.012480888929631976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,6144,0.09344444672266643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,2048,0.04846400022506714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,10240,0.05529510974884033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,8192,0.012913777596420713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,8192,7.7700288560655375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,256,0.012685333689053854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,2048,0.029450668228997126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,512,0.030919111437267725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,12288,0.01239466667175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,1536,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,6144,0.014316444595654806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,16384,0.029013332393434312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,128,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,10240,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,7168,0.01964977714750502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,1536,0.003871110992299186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,768,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,6144,0.013162666724787818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,4096,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,3072,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,2048,0.009509333305888707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,51200,0.13615288999345568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,512,0.005423999908897612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,3072,0.2985031074947781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,65536,0.021746666895018682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,3072,0.03056710958480835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,4096,0.011045333411958484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,7168,0.012034667034943899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,512,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,16384,0.04495644569396973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,32,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,16384,0.13394399483998617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,7168,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,4096,0.007696888513035244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,1024,0.012140444583363004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,7168,0.034725334909227155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,1024,0.004584889031118817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,3072,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,2560,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,7168,0.15483200550079346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,4096,0.11526044209798177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,5120,0.039375109805001154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,65536,0.06137510803010729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,65536,0.062237335575951465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,65536,1.0674666298760307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,10240,0.05846133497026232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,128,0.006664000037643645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,16384,0.039157334301206805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,6144,0.049307554960250854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,16384,0.055254220962524414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,1024,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,2560,0.009642666412724389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,65536,3.6136695014105906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,7168,0.05213155680232578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,51200,0.10563111305236816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,8192,0.0580915543768141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,1536,0.003991999973853429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,3072,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,10240,0.01219733307758967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,10240,0.028247998820410833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,4096,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,1024,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,10240,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,6144,0.048158221774631076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,1536,0.012551999754375882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,256,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,1536,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,7168,0.07678755786683825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,1024,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,1024,0.03557866811752319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,3584,0.008469333251317343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,256,0.02162222233083513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,5120,0.11725778049892849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,32,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,3584,0.018253333038753934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,16384,0.03232266505559286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,1536,0.023352889551056757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,2560,0.028371555937661067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,5120,0.01552088889810774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,51200,0.22728888193766275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,65536,0.09233066770765518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,16384,0.04781244529618157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,65536,0.5488150914510092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,8192,0.011311111350854238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,1024,0.027569777435726587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,256,0.01722311145729489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,3072,0.011092444260915121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,768,0.005718222094906702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,6144,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,256,0.0052897776994440295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,3584,0.06370755698945788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,5120,0.01644444465637207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,3584,0.01982933282852173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,12288,0.02272355556488037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,7168,0.037328001525666975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,3584,0.1057902243402269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,1024,0.014440889159838358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,4096,0.38384532928466797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,10240,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,8192,0.11739288436041938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,10240,1.0900488959418404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,12288,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,2048,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,7168,0.015770667129092746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,1536,0.2102044423421224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,51200,0.0606479975912306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,1024,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,1024,0.029704888661702473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,512,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,8192,0.12276444170210098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,7168,0.030138668086793687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,768,0.006177777631415262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,512,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,65536,0.06262666649288602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,2048,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,7168,0.09140000078413223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,2560,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,12288,0.029659556017981634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,2048,0.024039111203617517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,128,0.034104890293545194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,51200,0.591867552863227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,51200,0.11197955078548855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,2560,0.024859555893474158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,256,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,51200,6.872778574625651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,3072,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,1024,0.13080710834927028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,2560,0.01887999971707662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,256,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,1536,0.013062222136391534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,5120,0.036186665296554565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,5120,0.015815110670195687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,5120,0.026820444398456152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,1536,0.03869422276814779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,5120,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,768,0.0026177778426143858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,51200,0.6362320052252876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,1024,0.007114667031500075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,51200,1.4714541965060766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,64,0.006009777800904379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,3072,0.06767200099097358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,6144,0.5389413303799099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,512,0.02160088883505927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,1024,0.053007112609015576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,6144,0.040603554911083646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,4096,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,7168,0.022446221775478784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,256,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,51200,0.06736266613006592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,1536,0.028992888000276353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,3072,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,32,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,128,0.017268444101015728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,64,0.1220053301917182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,256,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,1024,0.00850311087237464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,3584,0.03418577710787455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,7168,0.024099555280473497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,10240,0.07010933425691393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,4096,0.015542222393883599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,51200,0.1773973306020101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,2560,0.014708444476127625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,5120,0.02942311101489597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,32,0.011712888876597086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,1024,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,51200,0.061426666047838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,51200,13.238271077473959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,65536,0.253802670372857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,256,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,65536,0.3273422188229031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,6144,0.011194667054547204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,2560,0.014362666341993542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,2560,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,128,0.02130844526820713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,256,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,5120,0.16895111401875815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,6144,0.04203733470704821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,16384,0.089501334561242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,64,0.011697777443461947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,64,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,16384,0.10753689209620158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,32,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,6144,0.019506666395399306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,1536,0.05980177720387777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,12288,0.01830666595035129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,4096,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,3584,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,7168,0.03171199891302321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,10240,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,512,0.04719733198483785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,10240,0.014534221755133735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,12288,0.013264888690577613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,65536,0.09814577632480198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,3072,0.026829333768950567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,8192,0.20243377155727812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,65536,0.19447022014194063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,6144,0.13639644781748453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,5120,0.08846488926145767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,3584,0.013753777576817406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,16384,0.0357893341117435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,512,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,512,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,256,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,12288,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,768,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,256,0.008023111356629265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,128,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,128,0.053561776876449585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,5120,0.36516088909573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,7168,0.07888711161083646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,6144,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,2560,0.007342221836249034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,6144,0.01722400055991279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,32,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,1024,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,1024,0.16801955964830187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,256,0.02583377725548214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,8192,0.07088622119691637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,4096,0.03418933351834615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,3072,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,768,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,256,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,4096,0.03870844509866502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,3584,0.020591111646758187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,128,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,128,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,4096,0.17926043934292266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,7168,0.12116177876790364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,128,0.012794666820102267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,16384,0.02195555633968777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,51200,0.052589333719677396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,2048,0.026644445127911035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,256,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,1536,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,7168,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,256,0.018654222289721172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,4096,0.040780444939931236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,4096,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,1536,0.011929777761300405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,10240,0.013992889059914483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,51200,0.12292533450656468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,51200,0.14039111137390137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,5120,0.04878488845295376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,8192,0.29085156652662486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,3072,0.028408888313505385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,1536,0.05981155236562093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,10240,0.03560533457332187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,12288,0.16588000456492105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,3584,0.013798221945762634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,128,0.0021991110924217436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,16384,0.24477778540717232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,1536,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,16384,0.19316177897983125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,65536,0.05504444572660658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,8192,0.02536444365978241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,256,0.00904177791542477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,7168,0.005535111245181825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,51200,0.20302844047546387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,4096,0.019673777951134574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,65536,3.5070862240261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,10240,0.03864799936612447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,3072,0.013292444248994192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,7168,0.019680889116393197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,51200,0.21437156200408936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,256,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,256,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,5120,0.007052444749408298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,64,0.006509333435032103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,2048,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,256,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,3584,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,1024,0.02366133365366194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,3584,0.0886240005493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,12288,0.2824453247918023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,16384,0.037703110112084284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,256,0.005050666630268097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,128,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,2560,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,128,0.014441778262456259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,5120,0.9497493108113607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,5120,0.012028444144460889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,128,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,512,0.019202666150199044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,768,0.017539555827776592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,3584,0.020391111572583515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,10240,0.0658444431093004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,6144,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,65536,0.04833866821395027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,51200,0.04621333214971754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,128,0.0020826665891541377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,6144,0.07657955752478705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,128,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,6144,0.04669510987069872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,64,0.006143110907740063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,128,0.00408266650305854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,256,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,51200,0.09724266661538018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,512,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,8192,0.014597333139843412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,3072,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,2560,0.012784000072214337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,16384,0.058259553379482694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,7168,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,6144,0.010218666659461128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,1536,0.014303111367755465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,16384,0.025126222107145522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,3584,0.007520889242490132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,768,0.018553778529167175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,51200,0.22546309894985625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,12288,0.036157333188586764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,5120,0.23008267084757486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,8192,0.04763288961516487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,3072,0.021300445000330608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,4096,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,65536,3.3932497236463757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,256,0.05106400118933784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,32,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,4096,0.08004711071650188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,12288,0.08046755525800917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,32,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,4096,0.03583733240763346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,2560,0.010688888529936472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,10240,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,16384,0.4172933366563585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,8192,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,128,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,1024,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,4096,0.030067556434207495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,32,0.0021928888228204516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,8192,0.02769777841038174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,3072,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,2560,0.02160800000031789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,2560,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,10240,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,7168,0.027235555979940627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,128,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,1024,0.003559111307064692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,32,0.1538053353627523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,12288,0.046738667620552905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,7168,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,128,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,2560,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,2048,0.06390844451056586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,10240,0.02977333466211955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,10240,0.05617688761817085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,1024,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,1024,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,128,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,512,0.015452444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,2048,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,8192,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,4096,0.060066666867997914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,51200,0.4202337794833713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,32,0.12672800487942165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,8192,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,2048,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,768,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,6144,0.012011555333932241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,4096,0.07051110929912992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,2048,0.006729777902364731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,10240,0.013229333692126803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,5120,0.0170133329100079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,5120,0.02427733275625441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,8192,0.02377333409256405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,3072,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,65536,0.06810222069422404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,3072,0.05481333202785916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,256,0.04675199919276767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,16384,0.14061600632137722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,512,0.008107555409272512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,6144,0.015592000550693937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,128,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,2560,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,1024,0.047912001609802246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,3584,0.183333330684238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,2560,0.10853511095046997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,16384,0.2781742148929172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,1024,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,512,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,1024,0.0034328889515664843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,5120,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,768,0.020236444142129686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,1536,0.024062222904629175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,7168,0.11758577823638916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,256,0.007350222104125553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,6144,0.03408622079425388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,8192,1.8204941219753688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,12288,0.04179733329349094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,3072,0.38749154408772785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,512,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,64,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,65536,0.059248891141679555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,3584,0.10083199871910943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,1536,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,1536,0.008023999631404877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,16384,0.019897777173254225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,256,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,256,0.023503111468421087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,3584,0.6360461976793077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,2560,0.12602933247884116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,16384,0.029835555288526747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,512,0.057429333527882896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,4096,0.13639644781748453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,7168,0.14809600512186685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,1536,0.01759288873937395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,2560,0.013039111263222165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,6144,0.022664889693260193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,65536,0.05863644679387411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,3584,0.007827555967701806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,768,0.24544445673624674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,16384,0.044900443818834096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,7168,0.14151199658711752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,12288,0.027492443720499676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,64,0.013364444176355997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,1536,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,8192,0.03351466523276435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,4096,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,65536,0.16660355197058782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,10240,0.056051552295684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,64,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,1024,0.007393777370452881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,8192,0.4656960169474284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,8192,0.12534577316708034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,768,0.029382222228580054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,128,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,128,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,4096,0.02734666731622484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,3584,0.10171822044584487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,10240,0.03393066591686673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,16384,0.010289777484205034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,16384,0.03423466616206699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,65536,0.34382934040493435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,1024,0.04269955555597941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,8192,0.023728888895776536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,65536,0.35868000984191895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,51200,0.03976888789070977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,128,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,768,0.03127022253142463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,4096,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,1024,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,16384,0.16909421814812553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,6144,0.02399377855989668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,5120,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,2048,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,5120,0.0539537767569224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,256,0.01202755586968528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,3584,0.008392000363932716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,3072,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,10240,0.021631111701329548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,65536,0.12078577942318386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,2048,0.020264888803164165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,256,0.005459555735190709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,768,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,3072,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,10240,0.06015555726157295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,128,0.10924088954925537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,2560,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,65536,6.792917463514541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,64,0.005997333261701796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,64,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,2048,0.04678488771120707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,256,0.02907199992073907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,16384,0.025951999757024977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,7168,0.020564445190959506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,8192,0.1118862231572469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,5120,0.04941155513127645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,12288,0.04902844296561348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,16384,0.027319999204741582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,4096,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,7168,0.13241422176361084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,3072,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,128,0.015726221932305228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,2560,0.0784826676050822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,12288,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,4096,0.05063022176424662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,512,0.04696444339222378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,51200,0.0672960016462538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,12288,0.018187556001875136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,64,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,51200,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,65536,0.23942934142218697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,12288,0.1454533338546753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,2560,0.02370044423474206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,256,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,16384,0.04840444525082906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,512,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,256,0.015167999598715039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,2048,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,4096,0.018186666899257235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,7168,1.573579576280382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,3072,0.016343111793200176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,5120,0.013812444276279874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,2560,0.1836186647415161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,16384,0.016128889388508268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,4096,0.054273777537875705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,3584,0.03304800060060289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,512,0.007423111134105259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,5120,0.050514668226242065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,3072,0.10713689194785224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,8192,0.013551111022631327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,512,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,128,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,2048,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,5120,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,3584,0.013099555340078143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,65536,0.18046311537424722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,2048,0.05847644143634372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,10240,0.05452533231841194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,7168,0.021702221698231165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,1536,0.010400888820489248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,2560,0.01699022286468082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,256,0.011005333728260465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,10240,0.02916355596648322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,256,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,512,0.0024497778051429326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,65536,10.560496012369791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,2560,0.030015110969543457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,6144,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,1536,0.024166221419970196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,2560,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,12288,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,512,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,65536,0.06224089198642307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,65536,0.08401688602235581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,51200,0.1035333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,4096,0.0923608872625563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,10240,0.042747557163238525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,2048,0.015579554769727917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,5120,0.2676195568508572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,3584,0.007932444413503012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,64,0.07026577658123441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,32,0.004986666556861666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,5120,0.01921066641807556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,3072,0.02495288848876953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,51200,0.09550044271681045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,512,0.002971555623743269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,5120,0.1513528823852539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,2560,0.02713955607679155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,256,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,6144,0.02204888893498315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,256,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,3072,0.02902311086654663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,6144,0.02984444300333659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,6144,0.05239822136031257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,8192,0.03968177901373969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,8192,0.2269982231987847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,6144,0.06538133488761054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,128,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,768,0.004204444587230682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,128,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,3072,0.007676444119877285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,1024,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,1536,0.02889066603448656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,10240,0.19376533561282686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,51200,0.9021502600775825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,12288,0.07724444733725654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,256,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,12288,0.02424888809521993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,12288,0.1568631066216363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,3584,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,8192,0.0499528878264957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,768,0.019370666808552213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,65536,0.13067377938164607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,3584,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,6144,0.027225777506828308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,16384,0.023923556009928387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,3584,0.8495013448927137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,6144,0.018583999739752877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,1024,0.01885600056913164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,8192,0.03056710958480835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,3072,0.017483555608325534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,65536,1.2152212990654838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,65536,0.29948798815409344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,16384,0.04628444380230374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,10240,0.10525155729717678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,768,0.005731555736727185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,12288,0.03374844458368089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,3584,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,128,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,6144,0.0369466675652398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,16384,0.37005066871643066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,128,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,768,0.015180443723996481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,2560,0.014294221997261047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,256,0.008059555457697975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,65536,0.046909332275390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,65536,0.26128265592787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,3072,0.021911111142900255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,3584,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,4096,0.0159608887301551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,3584,0.011354666617181567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,512,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,16384,0.03572533196873135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,4096,0.11294399367438422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,128,0.04932622114817301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,7168,0.011413333316644033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,10240,0.07253333595063952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,64,0.01867288847764333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,3584,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,16384,0.01608444419172075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,32,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,6144,0.02681511143843333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,5120,0.03525333272086249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,12288,0.05489066574308607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,12288,0.018946665856573317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,10240,0.06723199950324164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,256,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,16384,0.08874399794472589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,6144,0.02159022291501363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,768,0.005519999812046687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,64,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,512,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,1024,0.022368000613318548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,128,0.009164444274372524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,12288,0.3000008794996473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,8192,0.01292799992693795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,16384,0.042657779322730176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,16384,0.09495911333296035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,16384,0.3654408984714084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,3584,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,3072,0.010481778118345471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,7168,0.014696000350846184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,256,0.003585777762863371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,6144,0.018181333939234417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,5120,0.019926221834288705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,6144,0.26471734046936035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,3584,0.017525333497259352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,16384,0.1663235558403863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,51200,0.019542222221692402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,1536,0.018188445104493033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,65536,1.0131119622124565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,2560,0.027429333991474573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,3584,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,2560,3.4381226433648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,16384,0.017332444588343304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,64,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,2560,0.05196711089875963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,7168,0.03843999902407328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,1536,0.076292441950904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,12288,0.05875200033187866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,1536,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,3584,0.01293333371480306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,512,0.013097777962684631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,2560,0.0196079992585712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,64,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,64,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,7168,0.2656879954867893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,5120,0.03206400076548258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,65536,1.1377236048380535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,2048,0.01405244403415256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,3072,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,6144,0.024108444650967915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,768,0.03366044494840834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,1024,0.05995111332999336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,128,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,2048,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,12288,0.06660888592402141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,1536,0.0177484436167611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,2560,0.061234666241539844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,768,0.05191911260286967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,128,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,2048,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,8192,0.05235022306442261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,65536,0.21619289451175264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,5120,0.017239999439981248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,512,0.004334222111437055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,10240,0.027453333139419556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,512,0.013145777914259167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,12288,0.02593155536386702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,7168,0.12059822347429063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,12288,0.02238666680124071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,10240,0.015375110838148328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,4096,0.02623555560906728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,3584,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,4096,0.047232889466815524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,8192,0.08415644698672825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,1024,0.02714577813943227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,12288,0.1608417828877767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,65536,0.03587377733654446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,8192,0.17169155014885795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,16384,0.044585777653588184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,5120,0.009642666412724389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,5120,0.031594667169782854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,1024,0.007476444045702617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,51200,0.06853155295054118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,4096,0.02581866747803158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,3584,0.012674666941165924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,16384,0.1423919995625814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,4096,0.02293333411216736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,10240,0.0266951107316547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,512,0.015136000182893542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,4096,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,768,0.017866666118303936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,10240,0.02682311170630985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,5120,0.9683208465576172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,8192,0.040463109811147056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,7168,0.3310240109761556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,512,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,7168,0.030559112628300984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,7168,0.010048888623714447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,3072,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,2560,0.07048710849550036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,8192,0.018060444129837882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,1536,0.07650933663050334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,256,0.018526222970750596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,12288,0.18882666693793404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,12288,0.15909777747260198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,65536,0.12969244851006403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,128,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,7168,0.011396444506115384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,512,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,1024,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,3072,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,128,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,1024,0.027913779020309448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,4096,0.020448888341585796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,8192,0.01129155523247189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,8192,0.9241083992852105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,1024,0.00959644466638565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,256,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,256,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,3072,0.014865777558750577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,6144,0.17763911353217232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,7168,0.16058222452799478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,65536,0.2793360021379259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,2048,0.014552889598740472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,8192,0.04191022117932638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,1536,0.031017776992585924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,128,0.019703111714786954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,65536,0.07031200329462688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,1024,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,3072,0.05205066667662727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,768,0.034492443005243935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,6144,0.021328889661365088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,3584,0.02496355606449975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,6144,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,6144,0.06877866718504164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,2048,0.024057777391539678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,1536,0.007394666473070781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,7168,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,3072,0.025395555628670588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,16384,0.03593600127432082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,3072,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,8192,0.01777777738041348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,8192,0.12227822674645318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,16384,0.03183822168244256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,2048,0.196760892868042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,6144,0.03555555476082696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,256,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,1024,0.00925155563486947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,2048,0.021261332763565913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,512,0.018183110488785636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,51200,0.05102044343948364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,3072,0.08569422033098008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,1024,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,1536,0.005371555685997009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,1024,0.01553244392077128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,3584,0.1091102229224311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,12288,0.23871289359198677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,64,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,256,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,51200,0.11254399352603489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,4096,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,3072,0.07019911209742229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,6144,0.0691359970304701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,768,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,51200,0.4238293435838487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,8192,0.02222844461599986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,10240,0.021698666943444148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,4096,0.025331555141343012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,8192,0.19801066981421578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,768,0.04007110993067423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,8192,0.015204444527626038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,2560,0.01884888940387302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,5120,0.11060355769263373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,2560,0.013335111240545908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,7168,0.43486489189995664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,2560,6.2403717041015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,7168,0.07260533173878987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,2048,0.012542222109105853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,51200,2.7955805460611978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,5120,0.09653866953319973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,512,0.024321777953041926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,65536,0.07259377506044176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,65536,0.12348088953230117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,1536,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,1536,0.014886221951908536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,4096,0.021851556168662176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,256,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,10240,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,10240,0.3025511105855306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,12288,0.028743998871909246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,2560,0.06799911128150092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,512,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,128,0.013612444202105204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,1536,0.012879999975363413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,768,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,768,0.019592000378502738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,2048,0.01921955578856998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,1536,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,10240,0.014169777433077494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,10240,0.04627911249796549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,3584,0.022657778528001573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,3072,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,8192,0.18300978342692056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,2560,0.46474398507012266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,3584,0.027718222803539697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,7168,0.13937600453694662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,4096,0.04172177778349983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,512,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,2048,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,7168,0.033039109574423894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,64,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,16384,0.3578017817603217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,3072,0.07767733600404528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,4096,0.020119110743204754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,16384,0.022427555587556627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,768,0.013952000273598565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,3072,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,16384,0.04246755441029867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,3584,0.1552435557047526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,3584,0.05890311135186089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,1024,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,128,0.014321777555677625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,16384,0.23585245344373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,256,0.008540444076061249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,2048,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,16384,0.08629955848058064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,10240,0.028821332587136164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,65536,1.731214205423991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,3072,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,3072,0.025674665967623394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,51200,0.29767465591430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,64,0.00245066670080026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,128,0.020601777566803824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,1024,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,32,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,768,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,12288,0.025826666090223525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,10240,0.0499271121289995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,1536,0.6171413527594672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,2560,0.14236711131201849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,3072,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,32,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,7168,0.12586400243971083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,51200,0.15676266617245144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,3072,0.022287999590237934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,10240,0.022053332792388067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,1024,0.01625422139962514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,51200,0.016724444097942777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,1536,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,2048,0.09743733538521661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,16384,0.27811821301778156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,7168,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,7168,0.02311644454797109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,4096,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,12288,0.08393333355585735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,1536,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,65536,0.06314400169584486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,256,0.08147555589675903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,2048,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,65536,0.06389777527915107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,2560,0.5199617809719509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,768,0.02437066700723436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,2048,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,16384,0.0402124457889133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,12288,0.018590221802393597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,512,0.015344889627562629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,6144,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,512,0.01253600004646513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,65536,0.021847999758190576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,16384,0.17469333277808297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,12288,0.2156133386823866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,65536,0.09756977690590753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,51200,0.6237306594848633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,2048,0.01184266640080346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,51200,0.05793866846296522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,7168,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,3584,0.13370045026143393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,3584,0.035424000687069364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,128,0.023018666439586218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,512,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,2048,0.014302222265137566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,128,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,7168,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,128,0.044547554519441396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,2560,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,6144,0.012061333490742577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,16384,0.027808000644048054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,16384,0.5705111291673448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,16384,0.032249778509140015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,16384,0.047006222936842174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,8192,0.03888444436921014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,128,0.06540799803204007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,3584,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,65536,0.10150400135252212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,16384,0.05424977673424614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,2560,0.016201777590645682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,4096,0.048199110560946994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,16384,4.84901597764757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,51200,0.05229511194758945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,12288,0.29817954699198407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,512,0.0028728888266616394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,2048,0.04271289043956333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,768,0.017822222577200997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,51200,0.23696798748440215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,2560,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,16384,0.032497776879204646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,6144,0.011481777661376528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,1536,0.02419822249147627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,16384,0.11366844177246094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,3072,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,3072,0.22844976849026152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,256,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,65536,1.873793707953559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,1536,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,64,0.009736888938479954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,12288,3.297945870293511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,7168,0.05214755402670967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,768,0.16322488254970974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,10240,0.4550648795233832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,1536,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,4096,0.12003910541534424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,2048,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,32,0.3676906691657172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,3072,0.03854755560557047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,768,0.10126577483283149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,128,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,8192,0.11249600516425239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,7168,0.12685867150624594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,3072,0.016876444220542908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,2560,0.06447022491031222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,3072,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,16384,0.18650666872660318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,10240,0.03532800078392029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,256,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,2048,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,12288,0.02328799996111128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,16384,6.095857832166884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,64,0.012867555850081973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,256,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,65536,0.0773199995358785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,7168,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,2560,0.09726666741900974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,128,0.022682666778564453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,2048,0.023740443918440077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,512,0.039192888471815325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,16384,0.11048177878061931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,768,0.0466195543607076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,2560,0.003896000070704354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,2048,0.06316710842980279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,1024,0.004556444370084339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,1024,0.03172711200184292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,128,0.0053671110007498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,5120,0.017264000243610807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,512,0.0032640000184377036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,65536,0.1500693294737074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,512,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,64,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,1536,0.013815999858909182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,3584,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,12288,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,2560,0.26692265934414333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,6144,0.024511999554104272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,32,0.015895111693276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,3584,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,2048,0.00499199993080563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,7168,13.740384419759115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,1024,0.01991911066903008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,2048,0.03331111205948724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,256,0.12402933173709446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,16384,0.08204622401131524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,10240,0.09253155522876316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,3072,0.03685333331425985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,2560,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,1024,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,256,0.08698044220606486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,4096,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,256,0.03671111000908746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,5120,0.010478222535716163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,64,0.1719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,2560,0.02369244396686554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,256,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,7168,0.13469421863555908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,12288,0.05697688791486952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,10240,0.08842044406467015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,3072,0.08399822314580281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,6144,0.06388266881306966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,16384,0.1781066656112671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,32,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,6144,0.019535111056433786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,4096,0.24298222859700522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,2560,0.08095288938946195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,64,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,6144,0.03711466656790839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,1536,0.0033297776761982176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,8192,0.021293333835071985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,2560,0.02881866693496704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,4096,0.0627795524067349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,256,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,7168,0.03983555568589105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,16384,0.9977493286132812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,8192,0.011379555695586734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,768,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,3584,0.005680000202523337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,65536,0.3676186667548285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,51200,0.07936177651087443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,2048,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,8192,0.08652622169918484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,3584,0.030212445391549006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,1024,0.014125333891974555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,5120,0.05638933181762695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,51200,0.13110489315456814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,5120,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,65536,0.7059475580851237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,10240,0.0465555555290646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,3584,0.0396382212638855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,768,0.007919111185603673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,256,0.05600711372163561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,768,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,3584,0.1367324458228217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,1536,0.010407110883129967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,12288,0.3322880003187391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,512,0.05700800153944227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,3584,0.045752889580196805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,1536,0.0258942229880227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,2048,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,7168,0.02437955637772878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,12288,0.056828445858425565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,7168,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,2048,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,3584,0.02940000096956889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,65536,2.3502089182535806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,51200,0.16471821731991237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,6144,0.048531555467181735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,5120,0.00904088881280687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,768,0.012240000069141388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,51200,9.382069057888454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,1536,0.013391111459996967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,51200,0.05134222242567274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,16384,0.09325155284669663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,1024,0.03996444410747952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,2048,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,64,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,768,0.19568978415595162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,6144,0.03289244572321574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,51200,0.20907733175489637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,2560,0.06206133630540636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,5120,0.0466195543607076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,128,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,128,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,8192,0.02569688856601715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,256,0.01682577696111467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,64,0.005434666656785541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,1536,0.2973742220136854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,128,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,3072,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,32,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,768,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,3584,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,16384,1.4886052873399522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,51200,9.260078430175781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,256,0.0637271139356825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,4096,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,51200,0.0638151102595859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,2048,0.619600878821479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,3072,0.041831112570232816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,6144,0.011587555209795633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,8192,0.017646221650971305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,1536,0.08487555715772842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,12288,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,1536,0.004601777841647466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,2560,0.015659555792808533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,65536,0.07796977625952826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,256,0.014837332897716098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,2560,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,4096,0.07798666424221463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,12288,0.19804178343878853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,65536,0.43628621101379395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,256,0.012069332930776807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,12288,0.019464888506465487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,2560,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,51200,0.10638133022520278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,2560,0.07576800054974027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,4096,0.010307555397351583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,256,0.0053484443989064955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,3584,0.025320000118679468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,51200,0.981911129421658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,512,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,1024,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,8192,0.011420444481902652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,51200,0.484191099802653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,8192,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,128,0.005491555564933353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,12288,0.12011555830637614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,768,0.018681777848137748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,3072,0.7509466807047526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,256,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,7168,0.019671999745898776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,12288,0.03136355678240458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,3072,0.007585778004593319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,5120,0.37401866912841797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,5120,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,3584,0.02385422256257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,128,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,128,0.01681688924630483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,32,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,1024,0.016545777519543965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,12288,0.013325333595275879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,64,0.014691554837756686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,7168,1.0959786309136283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,3072,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,3584,0.027116444375779893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,2560,0.037010666396882795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,1536,0.017881777551439073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,3072,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,32,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,7168,1.3133083979288738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,2048,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,16384,0.026543110609054565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,2560,0.03225866622394986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,51200,0.06700444221496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,2560,0.05588799715042114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,2048,0.024519110719362896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,2560,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,51200,0.050148444043265454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,32,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,5120,0.13849511411454943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,51200,0.4241644541422526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,32,0.00638488887084855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,64,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,5120,1.4065404468112523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,4096,0.01718488832314809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,2560,0.01444533301724328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,4096,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,2560,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,3072,0.009757333331637913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,64,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,128,0.019036443697081674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,3072,0.012879999975363413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,64,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,5120,0.00453955555955569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,1536,0.014767999450365702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,256,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,7168,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,16384,0.171834667523702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,512,0.004473777694834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,3072,0.06930222113927205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,768,0.008746667040718926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,7168,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,4096,0.008023111356629265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,65536,0.4637466536627875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,3584,0.007877333296669854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,10240,0.023003555006451074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,1536,0.016190222567982145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,5120,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,4096,0.9104933208889432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,4096,0.010775110787815519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,768,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,51200,0.8649866845872668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,12288,0.11062222056918675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,12288,1.855227576361762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,7168,0.03330399923854404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,256,0.01145066652033064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,512,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,7168,0.045132444964514844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,8192,0.05379111236996121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,512,0.010297777752081553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,512,0.006212444355090459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,1536,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,5120,0.03639288743336996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,2560,0.00721599989467197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,16384,0.025432000557581585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,64,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,2048,0.019267555740144517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,7168,0.03085600005255805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,16384,0.3589635425143772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,128,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,256,0.0022577777918842104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,7168,0.05598400036493937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,2560,0.018596443865034316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,3584,0.009729777773221334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,10240,0.03856977820396423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,10240,0.06719822353786893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,7168,0.07811911238564385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,16384,0.07861155271530151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,4096,0.03859377900759379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,4096,0.031847112708621554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,3072,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,128,0.011446221835083432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,8192,0.029112001260121662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,1536,0.004293333325121137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,12288,0.011674666570292579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,2048,0.03380266825358073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,51200,0.5639484193589952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,5120,0.060011552439795605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,8192,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,128,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,2048,0.00906666699382994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,2560,0.04320799973275927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,2048,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,5120,0.02015999952952067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,512,0.00997688869635264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,256,0.012470222181744046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,256,0.010620444185203977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,7168,0.9320568508572049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,5120,0.8468409114413791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,768,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,128,0.009357333183288574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,128,0.025055110454559326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,1024,0.0032186667538351486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,3584,0.02964444292916192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,8192,0.028453333510292902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,5120,0.064282668961419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,5120,0.022080888350804646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,256,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,32,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,12288,0.01680355601840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,8192,0.022270222504933674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,2560,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,10240,0.03143999973932902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,512,0.01737155516942342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,768,0.04100888967514038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,8192,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,3584,0.039919111463758684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,16384,0.08828977743784587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,3584,0.05337511168585884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,4096,0.040954666005240545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,512,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,6144,0.024113777610990737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,65536,0.2124248875512017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,2560,0.008397333323955536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,512,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,1024,0.05327111151483324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,3072,0.007233777807818518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,2560,0.10942400164074367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,1024,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,51200,0.04892977741029528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,1024,0.26724444495307076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,6144,0.058635552724202476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,256,0.015307555596033732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,6144,0.16431021690368652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,32,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,65536,0.8808328840467664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,51200,0.060467554463280566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,7168,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,2560,0.043531556924184166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,16384,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,3072,0.005392888767851724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,2560,0.02366399930583106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,6144,0.05828266673617893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,1536,0.019359999232821994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,5120,0.13467110527886286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,3072,1.1131395763821073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,2560,0.01608711149957445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,8192,0.05299644337760078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,7168,0.03353955679469638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,4096,0.01923644377125634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,65536,0.05337599913279215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,1024,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,128,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,7168,0.023038221730126276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,51200,7.1611726548936625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,65536,0.831901338365343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,64,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,2560,0.03218133250872294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,51200,0.03386044502258301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,1024,0.004395555704832077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,51200,0.4811306529574924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,3584,0.037754668129815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,4096,0.03398399882846408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,32,0.009881777895821465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,4096,0.03706399930848016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,65536,0.24020356602138945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,768,0.003916444463862313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,4096,0.015488889482286243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,4096,0.019800888167487252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,768,0.04775377776887682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,3584,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,6144,0.1351999971601698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,51200,0.34051201078626847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,3072,0.02562933332390255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,64,0.0021991110924217436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,10240,0.22155466344621447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,4096,0.01773955590195126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,2560,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,12288,0.016917333006858826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,51200,0.08405422502093846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,4096,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,51200,0.07653066847059461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,16384,0.06578133503595988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,128,0.014115555418862237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,2048,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,10240,0.012134222520722283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,5120,0.01866933372285631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,3584,0.09016178051630656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,4096,0.017549332645204332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,64,0.057589332262674965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,3072,0.038591110044055514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,10240,0.12842488288879395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,64,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,2560,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,10240,0.02913244565327962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,6144,0.01143200033240848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,2048,0.08072621954811944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,5120,0.13621155420939127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,7168,1.3804462220933702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,7168,0.03878844446606106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,512,0.005721777677536011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,1536,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,512,0.02368444369898902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,64,0.004535111288229625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,16384,0.08164355489942762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,32,0.020040889581044514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,8192,0.022062222162882488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,2560,0.014426666829321118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,16384,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,2048,0.007139555282062954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,768,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,10240,0.044080889887279935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,768,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,128,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,4096,0.010769777827792697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,12288,0.1526026725769043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,12288,0.018391999933454726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,7168,0.08373688989215428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,768,0.01687733332316081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,256,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,2048,0.05361688799328274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,51200,0.3792595598432753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,2560,0.04941511154174805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,8192,0.06722844309277005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,128,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,65536,0.7561679946051704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,1024,0.007361777954631382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,16384,0.021783999270863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,512,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,5120,0.1404915518230862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,12288,0.06772533390257093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,65536,0.05502666698561775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,256,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,768,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,128,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,1536,0.05878133243984646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,1536,0.46880801518758136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,256,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,12288,0.6159040133158366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,16384,0.02204977803760105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,8192,0.011239111423492432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,16384,0.03828355669975281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,1024,0.08994488583670722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,768,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,1536,0.01962311069170634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,7168,0.014588443769348992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,5120,0.029819554752773706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,10240,0.011751111182901593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,7168,0.029857777886920508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,12288,0.025408888856569927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,6144,0.1963111162185669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,10240,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,2048,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,64,0.00518222235971027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,8192,0.11187556054857041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,64,0.0163137780295478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,8192,0.04069066709942288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,1024,0.057851552963256836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,512,0.030785779158274334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,1536,0.017301332619455125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,16384,0.026214222113291424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,16384,0.24491643905639648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,16384,0.08549866411421035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,10240,0.08609244558546279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,3072,0.030518223841985066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,7168,0.020272889071040683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,3584,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,64,0.03426577647527059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,6144,0.015287111202875773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,7168,0.08128977484173246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,7168,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,5120,0.014715555641386243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,1024,0.030679110023710463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,6144,0.02686311139000787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,3584,0.05277155505286323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,256,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,1024,0.005984000033802456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,32,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,65536,2.6523990631103516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,1024,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,512,0.0064088888466358185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,4096,0.09331288602617051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,512,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,64,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,12288,0.1062684456507365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,256,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,12288,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,128,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,1024,0.015955555770132277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,4096,0.04787733488612705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,32,0.007017778025733099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,1536,0.053159111075931124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,51200,0.4913937780592177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,7168,0.051057779126697116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,6144,0.010847110715177325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,2048,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,2048,0.1094684468375312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,512,0.058840003278520375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,1024,0.02683288852373759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,12288,0.02455377744303809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,1024,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,51200,0.1376711130142212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,5120,0.017551110850440133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,512,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,4096,0.035495112339655556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,2048,0.03657689028316074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,5120,0.023221333821614582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,1536,0.02257688840230306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,3584,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,16384,0.026260443859630164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,16384,0.04807466599676344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,256,0.017606221967273288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,10240,0.1808613273832533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,51200,0.8072382079230415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,5120,0.022833777798546687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,12288,0.01867022282547421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,51200,0.12394489182366265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,1536,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,2048,0.03786133395300971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,16384,3.239039103190104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,1536,0.026878222823143005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,128,0.010775999890433418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,2048,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,6144,0.014225777652528552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,6144,0.044290665123197764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,6144,0.03602666656176249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,128,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,2048,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,8192,0.04310133390956455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,16384,0.41272444195217556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,12288,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,2048,0.005048888838953442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,16384,0.23901955286661783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,2560,0.011714666253990598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,1024,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,1024,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,2048,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,7168,1.1881999969482422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,4096,0.02271822260485755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,4096,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,10240,0.011771555576059552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,6144,0.0281644430425432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,256,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,3584,0.014494222071435718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,2048,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,4096,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,51200,0.046911110480626426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,512,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,7168,0.014402666025691561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,12288,0.13606845008002386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,128,0.026570666167471144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,51200,0.33646843168470597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,5120,0.23992623223198783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,768,0.019109333554903667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,512,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,10240,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,5120,0.04775644342104594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,8192,0.45038578245374894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,4096,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,2560,0.028184887435701158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,8192,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,3584,0.0302471121152242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,12288,0.03737511237462362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,5120,0.011397333608733283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,4096,0.10638311174180771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,3072,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,1536,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,128,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,256,0.014913777510325113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,8192,0.013298666311634911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,128,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,5120,0.13103911611768934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,3584,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,10240,0.1653991142908732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,1024,0.028612444798151653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,6144,0.04283200038803948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,65536,0.0654933320151435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,512,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,1536,0.006315555423498154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,3072,0.00905777762333552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,2048,0.019997333486874897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,1536,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,1024,0.014879110786649915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,12288,0.15607555707295737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,2560,0.009269333548016017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,51200,0.10143199894163345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,10240,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,2048,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,512,0.015836444165971544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,5120,0.05981600284576416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,256,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,16384,0.0714142190085517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,51200,0.10775644249386257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,1024,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,4096,0.03110577662785848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,256,0.005535999933878581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,1024,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,2048,0.01825777855184343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,4096,0.09734844499164158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,2560,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,768,0.01236088905069563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,10240,0.03703466720051236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,16384,0.02887466549873352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,2048,0.033207999335394964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,10240,1.2095333735148113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,5120,0.08690577745437622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,10240,0.15561510456932917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,512,0.008265777594513362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,16384,0.0322728885544671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,65536,0.18752177556355795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,12288,0.017879999346203275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,256,0.01517955627706316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,8192,0.014064000712500678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,7168,0.17957688702477348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,32,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,8192,0.1118311087290446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,3072,0.04103199972046746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,7168,0.03625955515437656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,16384,0.14718843830956355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,16384,0.01956977778010898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,2560,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,2560,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,6144,0.018332444959216647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,10240,0.04525244567129347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,4096,0.022647110952271357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,65536,3.722476535373264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,128,0.024711999628278945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,6144,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,3072,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,2560,0.012055111428101858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,256,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,8192,0.15916889243655735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,5120,0.11416888236999512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,512,0.16769422425164116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,2560,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,12288,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,2048,0.03153333399030898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,2560,0.02702222267786662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,128,0.009453333086437648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,5120,0.012248888611793518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,512,0.016569778323173523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,4096,0.04452711012628344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,6144,0.34201600816514754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,3072,0.20119733280605742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,2048,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,2048,0.014911111858155994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,8192,0.49366044998168945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,8192,0.022986667023764715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,128,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,10240,0.028389334678649902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,32,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,512,0.014783999986118741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,64,0.018403554956118267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,65536,0.09984711143705581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,5120,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,7168,0.012831110921170978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,2560,0.013918222652541267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,2560,0.043748444981045194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,4096,0.00926577796538671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,7168,0.14991022480858698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,1024,0.051651553975211255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,2560,0.03401066528426276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,512,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,6144,0.025997334056430395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,8192,0.011959111524952782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,32,0.006077333456940121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,51200,0.0499582224422031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,10240,0.02438666588730282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,2560,0.01938222183121575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,10240,0.11619467205471462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,128,0.009373333719041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,512,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,512,0.015115555789735583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,4096,0.038022223446104265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,5120,0.3331342273288303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,256,0.003043555551105075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,3072,0.03911822372012668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,65536,0.07902755339940389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,8192,0.13984621895684138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,128,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,2048,0.011109333071443768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,51200,4.96036614312066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,1536,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,65536,0.05698844459321764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,5120,0.03903733359442817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,2048,0.03821333249409994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,64,0.06896177927652995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,10240,0.04475111100408766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,65536,0.0766506658660041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,4096,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,65536,0.18841866652170816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,128,0.013824000126785703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,256,0.010408889088365767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,3584,0.010596444209416708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,3072,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,16384,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,2560,0.032631112469567194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,2560,0.032251556714375816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,512,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,16384,0.015777778294351365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,10240,0.01870755520131853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,16384,0.39484087626139325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,8192,0.05779555771085951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,5120,0.10821777582168579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,512,0.01748533381356133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,12288,0.14080177413092718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,256,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,12288,0.029110223054885864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,5120,0.019215111931165058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,16384,0.009519110951158736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,256,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,10240,0.0443520016140408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,16384,0.14825599723392063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,768,0.024104000793562994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,51200,0.2836284372541639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,5120,0.10724266370137532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,2560,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,4096,0.05121422145101759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,4096,0.0730266637272305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,5120,0.009646221995353699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,256,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,2560,0.06332000096638997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,3584,0.022950222094853718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,10240,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,65536,1.4572683970133464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,51200,0.33620532353719074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,768,0.01441333360142178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,4096,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,65536,0.18052889241112602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,10240,0.24195644590589735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,4096,0.0678222245640225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,768,0.003107555624511507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,12288,0.2212488916185167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,51200,0.13090044922298855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,2560,0.016498666670587327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,3072,0.12750666671329075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,256,0.009768889182143742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,12288,0.032235556178622775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,6144,0.03644266724586487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,32,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,256,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,10240,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,8192,0.03607733382119073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,8192,0.07989688714345296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,12288,0.004930666751331753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,1024,0.017553778158293832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,1536,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,3584,0.16318310631646052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,65536,0.07846666706932916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,3072,0.043353776137034096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,16384,0.08247377475102742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,128,0.12220088640848796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,51200,0.0159288893143336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,4096,0.07177155547671847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,512,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,3584,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,65536,1.0214595794677734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,768,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,32,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,4096,0.028198222319285076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,2560,0.008628444539176093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,768,0.01201600001917945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,6144,0.019880889190567862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,65536,0.19704711437225342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,128,0.014120888378885059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,51200,0.41472533014085555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,768,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,64,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,3072,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,64,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,10240,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,8192,0.026359111070632935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,2560,0.2189191049999661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,128,0.0203146669599745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,65536,10.769866095648872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,512,0.00700711127784517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,6144,0.044406221972571484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,2560,0.010217777556843227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,3584,0.028697778781255085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,3072,0.052389333645502724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,256,0.022273777259720698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,128,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,2048,0.012484444512261285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,32,0.0018942221585247253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,256,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,256,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,32,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,1536,0.03931022352642483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,10240,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,1536,0.020954666866196528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,768,0.014021333720948962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,65536,0.07722577783796522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,4096,0.06105244159698486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,2560,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,51200,0.05620533227920532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,12288,0.024033778243594702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,256,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,51200,0.40155643886990017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,16384,0.08745511372884114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,3072,0.026716444227430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,2048,0.47508976194593644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,512,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,6144,0.004899555610285865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,32,0.004534222185611725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,7168,0.045200887653562755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,7168,0.02499733368555705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,512,0.014138667119873894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,4096,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,51200,0.6221155590481228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,16384,0.04829688866933187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,256,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,32,0.02450399928622776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,512,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,4096,0.01960177719593048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,10240,0.04418755571047465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,10240,0.02662311163213518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,1536,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,16384,0.016523554921150208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,3584,0.08879288699891831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,10240,0.06820622417661878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,3072,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,16384,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,6144,0.062168002128601074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,3072,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,3584,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,2560,0.016122667325867545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,256,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,1024,0.03743733300103082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,3584,0.0648897753821479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,10240,0.012170666621790992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,10240,0.04016977879736159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,1536,0.05609777900907728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,128,0.017902221944597032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,10240,0.015241778559154935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,256,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,3072,0.0722364452150133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,2048,2.019661373562283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,8192,0.02678311036692725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,1024,0.023944000403086346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,10240,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,8192,0.028374221589830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,16384,0.12360178099738227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,128,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,8192,0.02551288902759552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,65536,0.05647555324766371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,6144,0.027338667048348322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,2560,0.008212444682916006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,16384,0.1360657744937473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,3584,0.0440631111462911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,12288,1.8397706349690754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,65536,0.15035288863711888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,4096,0.04792622062895033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,65536,0.13794488377041286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,768,0.004295999805132548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,128,0.021387555532985266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,3072,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,6144,0.124889784389072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,32,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,10240,0.03035199973318312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,12288,0.005500444521506627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,256,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,768,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,32,0.005643555687533484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,4096,0.017866666118303936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,65536,39.087256537543404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,2048,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,65536,0.168250666724311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,768,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,16384,0.888677332136366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,256,0.008445333275530074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,1024,0.03371644351217482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,16384,0.23208533393012154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,7168,0.021793777743975323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,3072,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,6144,0.09590666823916966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,65536,0.09442310863071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,2048,0.07504266500473022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,7168,0.013768889009952545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,3584,0.052878220876057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,128,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,3584,0.031088001198238794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,8192,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,32,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,3584,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,51200,0.21828356054094103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,8192,0.04840177628729078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,12288,0.10237599743737115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,3584,0.03459822138150533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,5120,0.06305688619613647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,3072,0.03113422128889296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,768,0.003109333415826162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,64,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,4096,0.14316977394951716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,256,0.027816888358857896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,12288,0.04867022236188253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,65536,0.24306400616963705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,64,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,1024,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,12288,0.02642311155796051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,1536,0.012315555579132505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,12288,0.07756177584330241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,16384,0.027106665902667578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,12288,0.021659556362364028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,256,0.002114666625857353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,16384,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,3072,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,10240,0.0258568889564938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,2048,0.033231112692091204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,2560,0.0314302212662167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,51200,0.05117422342300415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,12288,0.04475288920932346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,512,0.012375999655988483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,12288,0.01889333294497596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,2560,0.008515555825498369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,6144,0.02347733411524031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,2560,0.049752887752321034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,12288,0.09814933273527358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,2560,0.03232977787653605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,16384,0.06063377857208252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,6144,0.04565511147181193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,3072,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,128,0.0030906666070222855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,768,0.0038248888320393036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,16384,0.01834844383928511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,65536,1.183743158976237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,16384,0.03394844465785556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,32,0.0123831108212471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,768,0.051648888323042125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,768,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,10240,0.010088000032636855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,6144,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,7168,0.028488887680901423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,768,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,7168,0.035263111193974815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,51200,0.038388444317711726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,2048,0.014928888943460254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,8192,0.2081857787238227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,768,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,5120,0.009752888646390703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,8192,0.27121244536505806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,6144,0.10503733158111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,768,0.007393777370452881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,7168,0.030992888742023047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,65536,0.05145600106981066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,512,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,128,0.03491377830505371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,5120,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,5120,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,768,0.02602400051222907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,7168,0.005492444253630108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,2560,0.04174400038189358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,12288,0.10462755627102321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,16384,0.16456354988945857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,1024,0.013807999591032663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,7168,0.05086400111516317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,32,0.014727110664049784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,10240,19.906404283311634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,10240,0.1096080011791653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,16384,0.01777066621515486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,768,0.05455288953251309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,2048,0.020287111401557922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,32,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,51200,2.2005625830756292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,512,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,2048,0.006666666517655055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,8192,0.016537777251667447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,64,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,64,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,2048,0.055155555407206215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,5120,0.028929776615566675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,16384,0.05453155438105265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,256,0.02035199933581882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,65536,0.06512888934877184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,7168,0.013073777986897362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,3072,3.7889573838975696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,3072,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,5120,0.053936001327302724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,768,0.036024000909593366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,4096,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,3584,0.03683999843067593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,1024,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,6144,0.02438844409253862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,1536,0.04457333352830675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,4096,0.008265777594513362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,2048,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,128,0.008252444366614023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,12288,0.03436177637841966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,1536,0.020382222202089097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,1024,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,7168,0.009424000150627559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,4096,0.021747555997636583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,2560,0.004865777575307422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,1536,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,256,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,128,0.01902844508488973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,64,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,2048,0.02612177696492937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,256,0.011920000116030375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,4096,0.09365422195858425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,128,0.015495111544926962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,10240,0.05487111210823059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,64,0.005201777650250329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,2560,0.02313066687848833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,512,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,8192,0.023203555080625746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,10240,0.06580711073345609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,3584,0.6785235404968262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,12288,0.0595102243953281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,2560,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,6144,0.03504800134234958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,512,0.010436444646782346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,1024,0.055072890387641064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,2048,0.007086222370465596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,2048,0.02495288848876953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,64,0.004110222061475118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,12288,0.019291554888089497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,128,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,2048,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,12288,0.09568177991443211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,7168,0.04036622246106466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,5120,0.06674310896131727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,51200,0.09550488657421535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,3584,0.02996888756752014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,512,0.005683555371231503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,10240,0.017877333694034152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,1536,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,3584,0.017518222332000732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,3584,0.03256533212131924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,1536,0.027078222897317674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,12288,0.19674400488535562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,768,0.018082666728231642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,512,0.019609777463806998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,12288,0.029121776421864826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,2560,0.03181688984235128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,128,0.005653333332803514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,12288,0.04703911145528158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,64,0.001855111163523462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,768,0.0032266666077905228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,51200,0.1326542165544298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,2048,0.029653333955340918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,3584,0.019893333315849304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,128,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,65536,0.16273421711391872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,32,0.006360000206364526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,51200,0.05189777745140923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,16384,0.023358222511079576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,51200,0.33780622482299805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,256,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,5120,0.07117866807513766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,1024,0.23166312111748588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,65536,0.07857866419686212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,65536,0.42293421427408856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,4096,0.008104888929261101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,3072,0.015792888071801927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,12288,0.01477688882086012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,256,0.11169066694047715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,2560,0.006780444333950679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,5120,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,16384,0.7877475420633951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,10240,0.04996711015701294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,4096,0.01200711147652732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,1024,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,2560,0.014819555812411837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,128,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,10240,0.02334311107794444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,10240,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,768,0.013507555756303998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,3584,0.8625049061245389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,1536,0.02424266603257921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,16384,0.2965199947357178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,128,0.003907555507289039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,4096,0.031586666901906334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,3584,0.010793777803579966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,6144,0.022594667143291895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,512,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,10240,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,4096,0.019117333822780185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,1536,0.013375110924243927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,12288,0.19748000303904215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,256,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,1024,0.02371999952528212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,5120,0.015527110960748462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,12288,0.2256408797370063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,128,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,1024,0.005701333284378052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,768,0.2645760112338596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,3072,0.039247112141715154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,7168,0.01015022231472863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,8192,0.030813333061006334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,5120,0.11921244197421604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,768,0.01408088869518704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,10240,0.24296087688869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,3072,0.03352088729540507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,2048,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,16384,0.02869422237078349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,1024,0.02791022260983785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,128,0.0294773346847958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,12288,19.89093017578125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,10240,0.42859554290771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,2560,0.12615022394392225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,16384,0.06024266613854302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,1536,0.009469333622190688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,65536,0.1568737824757894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,65536,1.6722515953911676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,128,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,16384,0.38510221905178493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,128,0.002798222212327851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,8192,0.012875555290116204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,2048,0.004561777744028303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,1536,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,768,0.014309333430396186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,10240,0.02885422110557556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,6144,0.014749333262443542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,3072,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,16384,0.0613937775293986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,128,0.023408000667889912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,256,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,256,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,16384,0.10471022129058838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,10240,0.24766667683919272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,256,0.014095111025704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,4096,0.1194888883166843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,8192,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,5120,0.03366311060057746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,768,0.024471110767788355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,1024,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,12288,0.05955377552244398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,2048,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,3072,0.2658657762739393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,10240,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,10240,0.300144010119968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,12288,0.29103467199537486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,6144,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,2048,0.007002666592597961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,1024,0.006735111276308696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,2048,0.007341333561473423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,5120,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,16384,0.20700977908240425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,2048,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,5120,0.08732799688975017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,256,0.0053119998839166425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,51200,0.06241599718729655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,256,0.010378667049937779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,128,0.006866666591829724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,1536,0.02961955467859904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,3584,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,16384,0.11322132746378581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,2560,0.021987555755509272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,1024,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,64,0.025847999585999384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,12288,0.020446222689416673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,51200,0.13051288657718235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,3584,0.028861333926518757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,1536,0.01614488826857673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,51200,0.25447021590338814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,65536,0.11440266503228082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,4096,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,6144,0.030971556901931763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,6144,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,256,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,1536,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,2560,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,2048,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,1536,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,64,0.05424266722467211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,16384,0.03418577710787455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,6144,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,1536,0.02034222251839108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,7168,0.09483822186787923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,8192,0.5958791308932835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,7168,0.17445955011579725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,2048,0.012470222181744046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,7168,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,256,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,12288,0.030626667870415583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,8192,0.024487111303541396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,6144,0.04092444313897027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,768,0.05137777659628126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,51200,0.5688346756829156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,3072,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,10240,0.10972711112764146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,1024,0.02808177802297804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,51200,0.07349066601859199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,1024,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,10240,0.15324532985687256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,51200,0.023205333285861548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,1024,0.015823110938072205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,512,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,51200,0.16885156101650664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,32,0.0029973333908451926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,10240,0.23616443739997017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,6144,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,1024,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,32,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,256,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,1536,0.005758222192525864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,3584,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,768,0.034635557068718806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,65536,0.2652595573001438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,6144,0.06973422235912748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,65536,0.9255439970228406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,512,0.052448888619740806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,3584,0.04384088847372267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,512,0.004292444636424382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,256,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,1024,0.02680177821053399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,65536,0.11447200510236953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,768,0.003661333272854487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,6144,0.06298044655058119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,3584,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,1536,0.3414453400505914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,7168,3.3458071814643016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,512,0.02028888960679372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,1024,0.09803999794854058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,8192,0.03733955489264594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,65536,0.27836354573567706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,3584,0.06354577673806085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,8192,0.02944444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,512,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,1536,0.007537778053018782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,12288,0.02889066603448656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,256,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,1024,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,2048,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,768,0.022131555610232886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,768,0.015563555889659457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,512,0.02424088948302799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,8192,0.017139555679427255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,2048,0.08124533626768324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,768,0.13092266188727483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,1536,0.0038373333712418876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,512,0.01372977760103014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,51200,0.9617404937744141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,32,0.02446311049991184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,256,0.06111288732952542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,3584,0.032163557079103254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,256,0.022981334063741896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,2560,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,1536,0.06352800130844116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,256,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,2048,0.017582222819328308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,12288,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,5120,0.012302222351233164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,128,0.015298666225539314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,1536,0.0038737778862317405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,128,0.002201777779393726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,65536,0.7672693464491102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,8192,0.042444444364971586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,6144,0.07394044266806708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,256,0.02461155586772495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,768,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,6144,0.021850667066044275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,8192,0.28819288147820366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,8192,0.024246222443050806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,12288,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,10240,0.05690755446751913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,3072,0.08299733532799615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,4096,0.006042666733264923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,65536,9.153163486056858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,4096,0.028845333390765723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,5120,0.01368533323208491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,65536,0.04972711205482483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,256,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,8192,0.016731555263201397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,12288,0.012076444096035428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,3072,0.018276444739765592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,768,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,3072,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,6144,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,51200,0.38692710134718156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,512,0.09761244720882839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,4096,0.08468800120883518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,4096,0.7803200085957845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,8192,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,768,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,4096,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,3584,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,8192,0.02575911084810893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,512,0.010819555156760745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,2560,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,12288,0.10305244392818874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,4096,0.5296550856696235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,3584,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,3584,0.3486435678270128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,6144,0.009296000003814697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,12288,0.05229155553711785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,8192,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,16384,0.07144711414972942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,16384,1.7383066813151042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,128,0.047424889273113675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,8192,0.04822133315934075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,5120,0.045569777488708496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,64,0.006196444647179709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,65536,0.01794311073091295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,10240,0.1960915591981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,6144,0.036278221342298723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,10240,0.11096888780593872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,8192,0.015303111738628812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,2560,0.0068400001360310455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,51200,0.06350666946834989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,2560,0.007703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,2560,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,4096,0.09267733494440715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,768,0.016869333055284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,768,0.018188445104493033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,5120,0.004304000073009067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,12288,0.12936622566647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,10240,0.017324444320466783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,51200,0.09573600027296279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,10240,0.04281244344181485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,512,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,10240,0.011762667033407422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,12288,0.05621600151062012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,4096,0.04664444592263964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,6144,0.0675084458457099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,3072,0.08365155590905084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,12288,0.026410667432679072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,512,0.02551377813021342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,8192,0.050366222858428955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,2048,0.27651111284891766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,4096,0.10718578100204468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,4096,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,6144,0.012881778180599213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,16384,0.19068800078497994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,10240,0.4730106459723578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,16384,0.0247999992635515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,6144,0.01812533371978336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,6144,0.045251554912990995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,5120,0.009424000150627559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,2048,0.037666665183173284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,1536,0.006509333435032103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,51200,0.040456000301573015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,5120,0.2547493245866564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,2048,0.05170933405558268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,7168,0.20222844017876518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,3584,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,65536,0.0755937761730618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,12288,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,2048,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,3072,0.012114666402339935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,16384,0.02525866693920559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,3584,0.04370933439996508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,10240,0.5225093099806044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,1024,0.01698755555682712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,65536,0.6404435369703505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,5120,0.04406755500369602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,3584,0.060995552274915904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,4096,0.02905511193805271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,3584,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,1536,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,3584,0.011864888999197217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,65536,0.09558400180604722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,2560,0.2506817711724175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,512,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,8192,0.013307555682129331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,16384,0.14458133114708796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,5120,0.009104000197516548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,4096,0.03039111031426324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,128,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,10240,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,64,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,5120,0.03471022182040744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,1536,0.003939555750952827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,768,0.004211555752489301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,1536,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,7168,0.004954666727119022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,2560,0.027815110153622095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,2560,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,4096,0.14203466309441462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,7168,0.0944711102379693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,64,0.0118231111102634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,256,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,256,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,512,0.0038248888320393036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,512,0.01740800009833442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,5120,0.030093332131703694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,2560,0.03331644336382548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,16384,0.4627608723110623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,3584,0.02219999995496538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,8192,0.06259822183185153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,512,0.010852444503042432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,1024,0.019899555378490023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,1024,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,12288,0.08349244462119208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,2560,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,32,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,1024,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,16384,0.03893688983387417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,16384,0.07069244649675158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,4096,0.00959644466638565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,6144,0.15895111031002468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,2048,0.024215110474162634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,512,0.015878222054905362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,12288,0.039806223577923246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,4096,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,7168,0.02795999911096361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,6144,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,32,0.00500266667869356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,768,0.1486133337020874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,8192,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,6144,2.34789064195421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,5120,0.10915555556615193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,7168,0.0373599992858039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,6144,0.250054226981269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,2560,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,16384,0.29793956544664174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,10240,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,7168,0.08518844180636936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,128,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,4096,0.22638845443725586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,16384,0.043988443083233304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,3072,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,4096,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,65536,0.5425804456075033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,32,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,768,2.1634303198920355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,65536,2.7886596255832248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,6144,0.015311110350820752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,3072,0.039268443981806435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,8192,0.03432088759210374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,128,0.016650666793187458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,51200,0.05872888697518242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,16384,0.031167109807332356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,51200,0.3658933374616835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,64,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,2048,0.04753333330154419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,10240,1.1519386503431532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,256,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,1536,0.021617778473430212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,3072,0.010954666468832227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,512,0.008913777768611908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,6144,0.06411733229955037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,8192,0.023565333750512864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,2048,0.13569422562917074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,12288,0.04907200071546766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,7168,0.012094222009181976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,65536,0.08465066883299087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,5120,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,512,0.015513777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,10240,0.03291822142071194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,512,0.013912000589900546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,512,0.021698666943444148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,32,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,3072,0.017279111676745944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,3584,0.08443733056386311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,12288,3.060707516140408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,256,0.06382666693793403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,1024,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,6144,0.04468889037768046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,7168,0.008854221966531541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,256,0.05559644434187147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,12288,0.3314613236321343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,1536,0.12826311588287354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,1536,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,12288,1.7897377014160156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,64,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,10240,0.04102755586306254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,8192,0.4040168921152751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,8192,0.08142044146855672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,512,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,8192,0.03240888979699876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,51200,0.04950755503442553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,5120,0.046893331739637584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,128,0.01572088897228241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,51200,1.2503830591837566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,4096,0.029867556360032823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,64,0.017529777354664274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,768,0.009931555224789513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,2560,0.04753422074847751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,32,0.007099555598364935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,10240,0.08010311259163751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,768,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,768,0.012666666673289405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,8192,0.19647378391689727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,512,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,1536,0.016823111308945548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,51200,0.03655022382736206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,2560,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,51200,4.440217759874132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,64,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,2560,0.029119110769695703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,256,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,512,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,10240,0.036143110858069524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,51200,0.08113688892788358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,8192,0.03649955656793382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,128,0.014551111393504672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,256,0.6885306570265027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,6144,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,1536,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,2048,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,5120,0.026704000102149114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,16384,0.02165866725974613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,12288,0.08042755391862658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,51200,0.15279377831353083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,3072,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,64,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,128,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,1536,0.006799110935793982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,16384,0.019070222973823547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,7168,0.01014311114947001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,32,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,6144,0.02167644434505039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,8192,0.030619555049472388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,6144,0.11470399962531196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,7168,0.06721422407362196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,1536,0.023744889431529578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,256,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,5120,0.23161511951022676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,32,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,768,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,4096,0.007329777710967594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,768,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,10240,0.17352355851067436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,3584,0.05018666717741224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,1024,0.017091555727852713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,7168,0.0509111119641198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,16384,0.43082578976949054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,6144,0.4257626798417833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,6144,0.2387831211090088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,3584,0.13465955522325304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,2560,0.04786666565471225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,1024,0.015447111593352424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,8192,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,1536,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,3584,0.01607466737429301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,32,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,10240,0.12631466653611925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,1536,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,2560,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,128,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,3072,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,6144,0.03138044476509094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,64,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,1024,0.0063235556913746735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,768,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,32,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,768,0.01996799972322252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,5120,0.37907732857598203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,1024,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,2560,0.028805332051383123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,7168,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,768,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,128,0.028607110182444256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,16384,0.06939733028411865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,1536,0.01903555625014835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,512,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,7168,0.022491556074884202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,65536,0.03219555483924018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,128,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,12288,0.03166755702760484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,128,0.014966222974989148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,3072,0.05109599894947476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,16384,0.04320533408059014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,1536,0.021029333273569744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,3072,0.02336799932850732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,128,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,51200,0.049871110253863864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,3072,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,51200,18.272986518012154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,7168,0.01255022237698237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,12288,0.07931733131408691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,4096,0.056968887646993004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,10240,0.017383999294704862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,16384,0.017224889662530687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,3584,0.09084444575839573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,1536,0.006079999936951532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,12288,0.17365511258443198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,1536,0.0546231104267968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,65536,0.16052533520592585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,3584,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,32,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,128,0.0018737777653667664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,2048,0.004358222087224324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,128,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,3584,0.01680888897842831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,4096,0.023353776997990076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,16384,0.0882835586865743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,16384,0.4475706418355306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,768,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,10240,0.014296889305114746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,5120,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,256,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,2560,0.030347555875778198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,7168,0.06382399797439575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,4096,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,768,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,3072,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,1536,0.004804444395833545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,1536,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,768,0.008809777597586313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,10240,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,8192,0.16950133111741808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,64,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,51200,4.701527913411458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,12288,0.025996444953812495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,32,0.07085955805248685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,8192,0.015640000502268474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,1536,0.009357333183288574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,768,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,7168,0.0630213353368971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,256,0.010952889091438718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,512,0.01889333294497596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,256,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,3072,0.01536355581548479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,6144,0.05593511131074694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,16384,0.01647644407219357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,3072,0.014737778239780001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,65536,0.15995644198523626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,6144,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,2560,0.00674044465025266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,10240,0.061584002441830106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,8192,0.10831022262573242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,768,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,3072,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,3584,0.0351235568523407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,1536,0.02200266718864441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,10240,0.07703200313780043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,65536,0.0789688891834683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,128,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,6144,0.014944000376595391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,768,0.2053040001127455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,512,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,3584,0.016958221793174744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,8192,0.011499555574523078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,256,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,5120,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,3072,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,12288,0.22980621125963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,2560,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,256,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,12288,0.028022223048739966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,1536,0.04564977685610453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,512,0.01591466698381636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,6144,0.07766133546829224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,16384,0.022291556000709534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,768,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,768,0.011872889267073737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,2560,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,1536,0.011807111402352652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,65536,2.158172395494249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,768,0.02230311102337307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,7168,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,8192,0.08194844590293036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,2560,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,128,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,4096,0.023570666710535686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,512,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,768,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,6144,0.019550222489568923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,16384,0.18648355536990693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,3584,0.041687998506757945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,768,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,5120,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,1024,0.1212044424480862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,512,0.011191111471917896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,51200,0.9481493631998698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,768,0.05800266398323906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,512,0.003855111284388436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,16384,0.057533330387539335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,1536,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,2048,0.17008444997999403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,32,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,4096,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,768,0.034062223301993474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,1536,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,1536,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,768,0.025496888491842482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,8192,0.056403557459513344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,65536,0.09499466419219971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,12288,0.020073778099483915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,1024,0.02371911042266422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,256,0.1534079975552029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,4096,0.03952622082498338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,10240,0.05335644549793667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,5120,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,2048,0.033089776833852134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,4096,0.024336889386177063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,3072,0.01846577723821004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,16384,0.14266578356424967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,2560,0.08249155680338542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,32,0.004324444466167026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,7168,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,1536,0.026741334133678015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,3584,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,8192,0.08283199866612752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,10240,0.9999556011623807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,8192,0.03979911075698005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,12288,0.04269244604640537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,8192,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,2048,0.017430222696728177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,10240,6.0294371710883246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,512,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,32,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,3072,0.10530310869216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,3072,0.02730400032467312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,2560,0.08257333437601726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,5120,0.04373244444529215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,32,0.008556444611814287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,6144,0.1865262190500895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,65536,15.626666598849825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,4096,0.0940435528755188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,65536,2.1985128190782333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,1024,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,65536,1.3776497311062281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,256,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,3584,0.010961777634090848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,16384,2.982257843017578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,512,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,4096,0.249034669664171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,1024,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,1536,0.01941777765750885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,51200,0.17925422721438936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,768,0.022896000080638464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,3584,0.022662222385406494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,3072,1.8911885155571833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,7168,1.3690790600246852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,12288,0.022317333353890315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,2048,0.03940977652867635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,4096,0.10621600018607245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,3584,0.01753155555990007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,128,0.00673333348499404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,8192,0.01603555513752831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,6144,0.07786933581034343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,3072,0.007344889144102733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,8192,0.09284355243047078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,2048,0.01310755560795466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,7168,0.05212710963355171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,8192,2.6195218827989364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,3072,0.011729777687125735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,65536,1.3510168923272028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,128,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,3584,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,5120,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,4096,0.013086222112178802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,256,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,512,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,64,0.005442666510740916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,512,0.014828445182906257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,2048,0.02653955585426754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,2560,0.03046222196684943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,5120,0.017215111189418368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,51200,0.4054311116536458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,7168,0.037985778517193265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,51200,0.045528001255459256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,4096,0.009045333498054082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,1024,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,1536,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,1536,0.018604444132910836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,6144,0.04040088918473985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,2048,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,128,0.01144177797767851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,1024,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,8192,0.026904000176323786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,512,0.012469333079126147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,1024,0.01554844445652432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,256,0.005395555661784277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,65536,0.0976675550142924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,1536,0.029475556479560003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,8192,0.05685244666205513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,8192,0.04018311036957635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,64,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,5120,0.08917866812811957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,51200,0.04300711221165127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,32,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,32,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,12288,0.1394080056084527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,3072,0.0913155542479621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,256,0.02438755498992072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,2048,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,2560,0.025919111238585577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,128,0.021020443903075323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,768,0.019166222876972623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,128,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,256,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,256,0.00254844439526399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,4096,0.02602488961484697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,3072,0.010098666780524785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,6144,0.025261332591374714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,256,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,8192,0.07841510905159844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,16384,0.05110933383305868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,256,0.024879111184014216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,7168,0.045349333021375865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,10240,0.21185778246985543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,8192,0.06869866450627644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,4096,0.01741333305835724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,5120,0.01571644511487749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,3072,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,4096,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,12288,0.02124711043304867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,8192,0.0498524440659417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,8192,0.03984266519546509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,512,0.02723377777470483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,16384,0.010974222587214576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,6144,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,51200,0.259061336517334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,4096,0.010519111322032081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,128,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,1024,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,12288,0.2540648778279622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,8192,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,8192,0.36089422967698837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,256,0.018253333038753934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,6144,0.012641777594884237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,7168,0.017997332745128207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,8192,0.025232000483406916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,12288,0.01590222285853492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,5120,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,51200,0.7904186778598361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,6144,0.0942115585009257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,2560,0.005992888990375731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,5120,0.017267554998397827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,3072,0.007903111477692923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,1536,0.019832889238993328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,51200,0.7066933314005533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,1536,0.020991999242040846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,128,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,12288,0.3118151028951009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,16384,0.12988711727990046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,1536,0.19265600045522055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,65536,0.23629956775241426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,8192,0.14338489373524985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,2560,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,1536,0.014494222071435718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,7168,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,512,0.020398222737842135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,128,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,16384,0.020616888999938965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,12288,0.035516444179746844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,65536,0.2819306585523817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,3072,0.018263111511866253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,64,0.002201777779393726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,7168,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,3072,0.017619555195172627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,512,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,3584,0.04882133338186476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,1024,0.038430223862330117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,12288,0.009517333573765224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,16384,0.06085155407587687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,512,0.022742221752802532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,4096,0.01146044416560067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,8192,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,128,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,12288,0.013435555828942193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,2560,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,10240,0.0181022220187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,2560,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,10240,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,5120,0.026554667287402686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,7168,0.018404444058736164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,12288,0.2103253338072035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,10240,0.12101244926452637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,256,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,3072,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,2560,0.03222222129503886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,768,0.47359821531507706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,4096,0.008649778034951951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,3072,0.012736000120639801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,768,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,10240,0.013730666703648038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,768,0.022113778524928625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,2560,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,256,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,3072,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,7168,0.019857777489556205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,64,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,7168,0.08993244171142578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,128,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,1024,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,4096,0.023910222782029047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,7168,0.012995555169052549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,16384,0.18005066447787813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,1024,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,768,0.03053155541419983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,65536,0.10897066858079697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,256,0.01832799944612715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,1536,0.0035137778355015647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,128,0.01519555515713162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,12288,0.03645688957638211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,51200,1.5773653454250758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,12288,0.02818044357829624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,12288,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,3072,0.009950222240553962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,256,0.010113777385817634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,1024,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,5120,0.024682665864626568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,768,0.010825778047243753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,768,0.008832889298597971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,10240,0.018217777212460835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,65536,0.26537778642442494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,1024,0.016413332687483895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,65536,0.08112177583906385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,16384,0.7811057832505969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,256,0.015277332729763456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,128,0.06134400102827284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,4096,0.021243555678261652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,5120,0.04619822237226698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,3584,0.021310221817758348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,768,0.01881511178281572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,3072,0.01346577786737018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,2560,0.021983999345037673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,2048,0.00980799976322386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,256,0.076401776737637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,51200,0.16868889331817627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,3072,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,128,0.010703110860453712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,6144,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,2048,0.02012888921631707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,7168,0.12361599339379205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,2048,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,12288,0.08160799741744995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,7168,0.034882667991850115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,8192,0.014340443743599786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,128,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,1536,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,8192,0.04378933376736111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,12288,0.03493422269821167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,10240,0.012488000094890594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,32,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,6144,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,4096,0.09038133091396755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,3584,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,4096,0.04881511131922404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,7168,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,5120,0.016122667325867545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,128,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,8192,0.01090222183201048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,7168,0.02324177821477254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,51200,0.06405422422620985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,2560,0.4692862298753526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,6144,0.45596355862087673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,512,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,12288,0.17554222212897408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,65536,0.22926221953497994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,5120,0.012119111087587146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,6144,0.009188444250159794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,2560,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,2048,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,512,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,12288,0.04072711202833388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,32,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,2560,0.011680000358157687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,256,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,8192,0.009933333430025313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,768,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,3584,0.027120888233184814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,3072,0.055402666330337524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,5120,0.03894222113821242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,16384,0.026446221603287592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,512,0.0025928889711697898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,8192,0.08162666691674127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,8192,0.015107555521859063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,65536,0.0744382209248013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,4096,0.04049688908788893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,16384,0.04071288969781663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,2560,0.09011999766031902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,32,0.0028853333658642242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,2048,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,512,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,16384,0.04703022374047173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,51200,0.08054400152630277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,1536,0.014720888601409065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,2560,0.006849777781301075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,512,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,7168,0.01365066650840971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,32,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,2048,0.07240266932381524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,64,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,512,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,768,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,12288,0.027179555760489568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,12288,0.1847884390089247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,8192,0.027144889036814373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,512,0.03896977835231357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,1024,0.028424888849258423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,1024,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,12288,0.03390844331847297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,51200,3.0205582512749563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,512,0.03750222259097629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,10240,0.11671111318800186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,65536,0.034126222133636475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,6144,0.021169778373506334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,8192,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,3584,0.019126221537590027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,32,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,7168,0.019838222199016146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,512,0.00618577789929178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,5120,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,768,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,64,0.024041778511471216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,64,0.002543111021320025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,7168,0.07151555352740817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,10240,0.041851556963390775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,5120,0.024452444579866197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,1024,0.0131217779384719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,1536,0.048244443204667836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,64,0.003208000005947219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,3584,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,8192,0.06115555763244629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,51200,0.06961511241065131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,51200,0.13600533538394505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,2560,0.29273777537875706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,8192,0.7211919890509711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,32,0.020206222931543987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,6144,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,768,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,51200,2.587079154120551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,1536,0.07343822055392794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,768,0.011007111105653973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,6144,0.12891911135779485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,7168,0.034324444002575345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,4096,0.03289688958062066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,128,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,10240,0.04503555430306328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,4096,0.050035556157430015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,128,0.026081777281231348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,16384,0.05372977919048733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,64,0.08738044235441421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,7168,0.1650142272313436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,7168,0.7017679744296603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,16384,0.5212506718105739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,7168,0.008409777449236976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,768,0.014468444718254937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,3072,0.061759114265441895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,1536,0.03305777907371521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,5120,0.07691555553012423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,5120,0.01201777739657296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,51200,61.618679470486114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,1536,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,128,0.03256888853179084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,2560,0.009635555247465769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,3072,0.010112000008424124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,16384,0.1672524478700426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,65536,0.132968889342414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,4096,0.025235555238193933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,3072,0.007126222054163615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,7168,0.8890737957424588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,256,0.026485333840052288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,7168,17.829332139756946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,65536,2.720738728841146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,51200,0.11291910542382134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,10240,0.11975467205047607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,32,0.008265777594513362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,12288,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,6144,0.038733333349227905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,4096,0.021391999390390184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,7168,0.09747999906539917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,12288,0.05281688769658407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,10240,0.005744888964626525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,3072,0.031930665175120033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,2048,0.01826400061448415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,64,0.002549333290921317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,65536,0.08945688936445449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,65536,0.718840863969591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,10240,0.036327110396491155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,5120,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,2048,0.008692444198661381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,1536,0.0284222231970893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,512,0.010806221928861407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,4096,0.022950222094853718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,3584,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,51200,0.07054666678110759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,768,0.0034835553831524322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,4096,0.023248000277413264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,8192,0.04247377647293938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,1024,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,6144,0.011117333339320289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,6144,0.11090399821599324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,4096,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,2048,0.042619556188583374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,3584,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,2048,0.007848000360859765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,7168,0.02056799994574653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,5120,0.0902328888575236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,8192,0.03403110967742072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,6144,0.05683110819922554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,4096,0.03866666555404663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,2048,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,6144,0.12537866168551975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,128,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,51200,0.01183111137813992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,2560,0.03971199856864081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,512,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,3584,0.0236826671494378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,51200,0.09506666660308838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,51200,0.2082293298509386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,8192,0.03811644514401754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,256,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,8192,0.029470221863852605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,4096,0.3480079968770345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,12288,0.04721777637799581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,51200,0.09761333465576172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,6144,0.020187555087937247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,64,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,3072,0.015471110741297403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,3584,0.029888888200124104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,1024,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,2560,0.15643821822272405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,3072,0.04792266752984789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,128,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,7168,0.06474222077263726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,64,0.01793955597612593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,65536,0.14918400181664362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,51200,0.19235200352138945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,4096,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,3584,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,3584,0.13928000132242838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,8192,0.01832177738348643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,1024,0.07477777534061007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,1536,0.02281066609753503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,7168,0.16364178392622206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,128,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,32,0.009076444639099969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,65536,0.06720978021621704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,12288,0.0195448895295461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,3584,0.027095999982621934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,768,0.037676443656285606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,512,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,512,0.025376889440748427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,768,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,2560,0.02941333254178365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,51200,0.12054488393995498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,5120,0.01649777756796943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,32,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,3072,0.01315199997689989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,3072,0.004543111142185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,12288,0.017983999517228868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,7168,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,8192,0.25496888160705566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,4096,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,16384,0.2527048852708605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,4096,0.013964444398880005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,65536,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,12288,0.024139554964171514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,65536,0.06843911276923285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,512,0.013227555486891003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,7168,0.06804977522956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,768,0.003552000141806073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,2560,0.007366221812036302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,65536,0.06544088655047946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,10240,0.17289777596791586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,65536,0.061653335889180504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,1024,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,128,0.011686222420798408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,128,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,5120,0.04387377699216207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,12288,0.03010666701528761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,12288,0.12612799803415933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,3584,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,3584,0.07520888911353217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,5120,0.01773066653145684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,8192,0.03140177660518222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,512,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,3584,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,512,0.005173333403136996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,3584,0.03205600049760606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,5120,0.1353315512339274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,12288,0.03964889049530029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,1536,0.24624623192681205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,256,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,3584,0.013307555682129331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,6144,0.01181333346499337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,8192,0.032724443409177996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,51200,0.0380364457766215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,2560,0.04972088999218411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,7168,0.010265777508417765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,4096,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,4096,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,6144,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,1536,0.017118222183651395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,3584,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,512,0.0026488889836602737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,1024,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,256,0.01755288905567593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,10240,0.024608888559871252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,128,0.017283555534150865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,10240,0.028947555356555518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,1024,0.024317334095637005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,16384,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,5120,0.03228266702757941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,8192,0.029849777619043987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,768,0.013002666334311167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,7168,0.020794666475719877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,6144,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,7168,0.09597333272298177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,1024,0.030449777841567993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,65536,0.24941865603129068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,65536,0.12179556157853867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,4096,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,64,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,7168,0.04867911007669237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,5120,0.09004800187216865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,256,0.012607999973826937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,512,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,512,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,512,0.003539555602603488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,12288,0.7158497704399956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,4096,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,7168,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,12288,0.053820444477929004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,3072,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,1536,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,1536,0.19736799928877088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,2560,0.015633778439627755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,6144,0.11050933599472046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,2560,0.017159110969967313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,6144,0.025504888759719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,256,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,10240,0.016379555066426594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,12288,0.02425777746571435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,4096,0.01800355480776893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,3584,0.11187644799550374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,16384,0.11005689038170709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,16384,0.04044355617629157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,256,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,1536,0.12950755490197077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,2560,0.01312800000111262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,512,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,5120,0.02442311081621382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,512,0.0057795556883017225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,4096,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,64,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,256,0.012110222544935016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,1024,0.0058613332609335584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,7168,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,32,0.013408888545301227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,256,0.007119999991522894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,8192,0.02156266735659705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,1536,0.007228444847795699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,8192,0.023176888624827068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,10240,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,4096,0.015097777048746744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,768,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,128,0.06509333186679415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,12288,0.035558223724365234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,2560,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,51200,0.07658400138219197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,12288,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,4096,0.018248889181349013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,51200,0.1978062258826362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,2560,0.08427821927600437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,2048,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,2560,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,64,0.002005333287848367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,51200,0.04780711068047417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,16384,0.4550924301147461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,768,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,768,0.00962311112218433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,6144,0.025605332520272996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,512,0.8122088644239637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,512,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,3584,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,1536,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,12288,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,6144,0.01036444471942054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,8192,0.04014044337802463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,51200,0.08360799815919663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,768,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,2560,0.0351804428630405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,4096,0.019440889358520508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,10240,0.035637332333458796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,10240,0.01273155543539259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,768,0.0124195557501581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,16384,0.27188266648186576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,16384,0.5226604673597548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,8192,0.5473386446634928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,64,0.04803466796875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,32,0.03814222084151374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,10240,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,6144,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,51200,0.07327200306786431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,32,0.0069057775868309875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,3072,0.036664889918433294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,7168,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,1536,0.010442666709423065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,6144,0.03195733494228787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,4096,0.0392524434460534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,1024,0.05162488751941257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,51200,0.054937776592042714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,128,0.0034933334423436057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,512,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,128,0.006538666784763336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,512,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,8192,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,768,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,1024,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,8192,0.04553777641720242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,10240,0.04289688997798496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,2048,0.06006044149398804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,5120,0.058892442120446094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,512,0.05263466636339823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,6144,0.031609776947233416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,256,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,3072,0.013062222136391534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,65536,1.8374809688991969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,10240,0.5393946435716417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,3072,0.037772443559434675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,3584,0.008857777549160851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,1536,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,2048,0.16413954893747965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,256,0.008443555898136562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,6144,0.06705600023269653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,256,0.018170666363504197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,51200,0.02091644373204973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,128,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,2560,0.024738666084077623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,128,0.02162666618824005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,768,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,6144,0.25062666998969185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,10240,0.0123831108212471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,1536,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,8192,0.011138666835096149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,256,0.021643555826610986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,32,0.0023111111174027124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,1536,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,4096,0.027855111493004694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,3072,0.031178666485680476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,10240,0.09039822551939224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,3584,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,512,0.023760888311598036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,8192,0.1542044480641683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,12288,0.021358221769332886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,4096,0.06932000319163005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,1024,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,2560,0.012725333372751871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,10240,0.014318222800890604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,4096,0.36109601126776797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,65536,0.07726933558781941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,2560,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,16384,0.06390488810009427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,3584,0.04320799973275927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,128,0.029834667841593426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,65536,0.03836888737148709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,128,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,16384,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,1536,0.016406221522225272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,3584,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,2560,0.01292266696691513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,65536,2.0948738521999783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,51200,0.056167112456427686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,1024,0.0037831109431054858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,16384,0.06207999918195936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,256,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,1536,0.004705777598751916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,256,0.04934222168392605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,5120,12.916605631510416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,2560,0.01586844523747762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,6144,0.06649600134955512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,5120,0.3236871030595568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,4096,0.014075555735164218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,8192,0.2634630997975667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,4096,0.19823911454942492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,4096,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,2048,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,5120,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,51200,0.059279998143514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,51200,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,3584,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,768,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,8192,0.015264888604482016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,6144,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,2560,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,3584,0.010836444795131683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,1024,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,51200,0.15447644392649332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,6144,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,7168,0.088009774684906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,768,0.16080978181627062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,10240,0.06359644730885823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,32,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,16384,0.047389331791136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,10240,0.013602666556835175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,7168,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,6144,0.0736044446627299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,65536,0.13004977173275417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,7168,0.30318133036295575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,768,0.013546666337384118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,16384,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,1024,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,512,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,4096,0.03221244282192654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,51200,0.04004177782270644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,10240,0.043975998957951866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,10240,0.06906221972571479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,512,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,1536,0.008927110996511247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,7168,0.0366231103738149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,3584,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,768,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,12288,0.02295911146534814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,51200,0.10682666964001125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,128,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,7168,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,5120,0.012332444389661154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,2560,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,3584,0.041570666763517596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,2560,0.03345155715942383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,128,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,2560,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,6144,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,2048,0.02676888969209459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,8192,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,12288,0.46845245361328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,3584,0.18564621607462564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,16384,0.021987555755509272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,3072,0.012247111234400006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,65536,0.09572444359461467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,2560,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,8192,0.02169688873820835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,3584,0.051635556750827365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,2048,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,51200,0.7174640231662326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,512,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,256,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,128,0.021524444222450256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,512,0.12400978141360813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,51200,0.043615112702051796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,8192,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,12288,0.20390755600399443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,128,0.02485688858562046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,10240,0.012895111408498553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,1536,0.04385688900947571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,1024,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,1024,0.07635733154084948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,512,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,32,0.05987377961476644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,32,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,512,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,8192,0.13457866509755453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,256,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,1024,0.009836444424258338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,512,0.012659555508030785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,6144,0.015987555185953777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,16384,0.27799643410576713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,256,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,768,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,512,0.027631999717818365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,8192,0.11628266175587972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,2048,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,8192,0.04914933443069458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,2048,0.026941332552168105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,128,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,256,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,65536,0.04987377921740214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,128,0.016198222835858662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,10240,0.02642222245534261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,7168,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,5120,0.22244887881808809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,65536,0.07328444719314575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,512,0.008571555217107138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,8192,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,768,0.03198400139808655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,12288,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,64,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,51200,0.05040355523427328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,2560,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,3584,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,51200,1.4266657299465602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,768,0.007147555549939473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,4096,0.042119999726613365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,7168,0.004632888982693355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,256,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,4096,0.1723191075854831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,12288,0.4063279893663194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,32,0.003638222399685118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,3072,0.015182221929232279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,32,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,12288,1.1765911314222548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,32,0.011051555474599203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,51200,1.7372062471177843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,4096,0.05478577812512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,10240,1.267929818895128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,2560,0.16718578338623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,16384,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,2560,0.12621422608693442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,2560,0.011300444602966309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,512,0.023962666591008503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,51200,0.030294222964180842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,5120,0.02289955483542548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,128,0.01572444372706943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,51200,1.598306655883789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,1024,0.013560000393125745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,65536,0.01961511042382982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,3072,0.07288089063432482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,65536,0.36857954661051434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,16384,0.32398488786485463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,3072,0.03424355718824599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,768,0.017885333961910672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,1024,0.014702222413486905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,1536,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,3072,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,65536,0.27748089366488987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,256,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,768,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,5120,0.012359999948077731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,6144,0.7913537555270724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,10240,0.05316266748640272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,1024,0.05257066753175524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,4096,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,256,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,12288,0.01402933398882548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,51200,0.0549813343418969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,3072,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,256,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,2048,0.02942311101489597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,4096,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,128,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,10240,0.01142133358452055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,16384,0.0410497784614563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,6144,1.453354623582628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,65536,0.6839795642428927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,6144,0.03682488865322537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,2048,0.005528000079923206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,512,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,1024,0.018907555275493197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,16384,0.0819937785466512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,7168,0.026777777406904433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,2560,0.17004888587527803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,7168,1.5869021945529516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,51200,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,51200,0.5476329061720107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,64,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,32,0.0034871109657817413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,512,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,16384,0.019692444139056735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,1024,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,10240,0.08543199963039821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,3584,0.5020960172017416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,32,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,2048,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,1536,0.021127111381954614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,3584,0.011510222322411008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,1024,0.003294222263826264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,65536,0.06646311283111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,2048,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,512,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,128,0.16776622666252983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,6144,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,5120,0.09744266668955485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,5120,0.05008622341685825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,4096,0.16313066747453478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,2560,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,12288,0.5114284621344672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,1024,0.02126844392882453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,256,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,65536,0.03968088825543722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,65536,0.3744648827446832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,256,0.023194667365815904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,2048,0.07859555880228679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,32,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,1536,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,16384,0.016940444707870483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,16384,0.03600000010596381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,5120,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,512,0.1099662250942654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,12288,0.29451823234558105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,3072,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,65536,0.3903822104136149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,2048,0.018114666144053142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,512,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,7168,0.10174044635560776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,3072,0.052589333719677396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,8192,0.01588622232278188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,1024,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,65536,0.24496621555752227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,6144,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,7168,0.03481066558096144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,4096,0.5189021958245171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,128,0.005190222213665645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,3584,0.0383840004603068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,128,0.02807733416557312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,12288,0.23645332124498156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,12288,0.5217804378933376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,65536,0.13764354917738172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,3072,0.016960889101028442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,128,0.0532897777027554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,2048,0.027102222045262653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,3584,0.06491200129191081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,51200,0.2481386661529541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,2560,0.021704889006084863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,6144,0.011184888581434885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,7168,0.016688888271649677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,128,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,5120,0.013910222384664746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,2560,0.11601511637369792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,3584,0.025016888976097107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,8192,0.03806666533152262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,3584,0.3464035458034939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,2048,0.01904533306757609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,1024,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,7168,0.0940435528755188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,3072,0.1448168887032403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,6144,0.037646220790015325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,8192,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,3584,0.014936889211336771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,5120,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,32,1.2039431465996635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,512,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,1024,0.020212444994184706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,6144,0.09966933064990574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,8192,0.2816248999701606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,65536,0.23036445511711967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,7168,0.025215999947653875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,6144,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,6144,0.010954666468832227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,512,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,51200,0.06996977991527982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,7168,0.04022488991419474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,6144,0.015621332658661736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,10240,0.10509244600931804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,6144,0.3153759903377957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,4096,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,65536,0.20087999767727324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,64,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,8192,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,3072,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,2048,0.023412444525294836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,256,0.009531555904282464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,6144,0.004822222308980094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,256,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,10240,0.012376888758606382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,3072,0.06642311149173312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,1536,0.030127998855378892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,7168,0.08781955639521281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,7168,0.0243857784403695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,6144,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,51200,0.04967733224232992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,3072,0.006895111252864202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,128,0.01482311056719886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,6144,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,10240,0.05760000149408976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,4096,0.16142488850487605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,768,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,51200,0.04965155654483371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,5120,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,7168,0.037471999724706016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,32,0.010427555276287926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,12288,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,6144,0.018043556147151522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,8192,0.016840888394249808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,128,0.03392533461252848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,3072,0.01735377808411916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,3072,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,7168,0.013696889082590738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,10240,1.4246817694769964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,64,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,5120,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,256,0.029545777373843726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,1536,0.1949297719531589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,768,0.013537777794731988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,65536,0.07504000266393025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,10240,0.01871200071440803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,6144,0.03579644362131754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,16384,0.050204445918401085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,16384,0.6008577876620823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,3584,0.010000000397364298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,16384,0.755564424726698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,3584,0.03269155489073859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,6144,0.017823111679818895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,7168,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,2560,0.05605066484875149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,256,0.024713777833514746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,32,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,512,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,768,0.0038488888078265717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,2560,0.015855999456511605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,10240,0.026368000441127356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,51200,0.052926222483317055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,256,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,4096,0.04833777745564779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,51200,0.03971555497911241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,2048,0.008361777497662438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,16384,0.06969955232408312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,4096,0.008784000244405534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,256,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,2048,0.022088888618681166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,6144,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,51200,0.07276444302664863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,4096,0.07066844569312201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,65536,0.0890515579117669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,4096,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,3584,0.023797333240509033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,51200,0.42238132158915204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,256,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,6144,0.02250755495495266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,64,0.003847111016511917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,65536,0.14851644304063585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,5120,0.0233297778500451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,3072,0.22722578048706055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,16384,0.19196711646185982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,10240,0.04184177849027845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,768,0.022587555978033278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,768,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,8192,0.03536533315976461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,3584,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,3584,0.014337778091430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,10240,0.02111377815405528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,4096,0.05605955256356133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,3584,0.5402142206827799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,12288,0.016623111234770883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,5120,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,256,0.018920888503392536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,128,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,1024,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,2560,0.06595110893249512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,2560,0.033306668202082314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,65536,0.0769253306918674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,10240,4.632602691650391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,512,0.022340445054901972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,8192,0.05204266640875074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,512,0.005208888815508948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,2048,0.00620888877246115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,2048,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,4096,0.007917332980367873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,768,0.038704888688193426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,3584,0.020534222324689228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,7168,0.090137779712677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,7168,0.060692442788018115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,7168,0.14619110690222845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,7168,0.025921778546439275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,12288,0.5302471054924859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,128,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,768,0.04087466663784451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,1536,0.014304889572991265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,3072,0.0919991135597229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,16384,0.1365991168551975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,128,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,512,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,65536,1.861787584092882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,3584,0.02397955622937944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,12288,0.04615555538071526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,2048,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,6144,0.08824621968799168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,128,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,2560,0.00792711145348019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,768,0.01702222228050232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,51200,0.3970079951816135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,5120,0.014504000544548035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,128,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,6144,15.393218994140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,16384,0.03098755412631565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,51200,0.5392124387953017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,768,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,65536,0.18673511346181235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,256,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,768,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,7168,0.03555377655559116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,6144,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,4096,0.0107342220014996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,768,0.01906133360332913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,4096,0.03854488995340135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,65536,1.3186302185058594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,3072,0.011790222591824003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,3584,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,256,0.025748444928063288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,2560,0.07005777623918322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,2048,0.02938666608598497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,32,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,1024,0.009374221993817223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,16384,1.9369547102186415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,4096,0.1133599943584866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,1536,0.003896000070704354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,768,0.05207733313242594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,2560,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,768,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,8192,0.07904444138209026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,65536,0.08272977670033772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,768,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,128,0.020067556036843192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,7168,0.014332445131407844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,256,0.11847821871439616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,10240,0.05096177922354805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,512,0.014581332604090372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,128,0.08409511380725437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,2560,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,4096,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,128,0.005680888891220093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,12288,0.018532445033391316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,256,0.013435555828942193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,51200,0.05303111010127597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,10240,0.03675555520587497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,2560,0.057393776045905225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,7168,0.7554204199049207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,3072,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,51200,1.211935149298774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,512,0.02648177742958069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,8192,1.991312026977539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,3072,0.5529555744594997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,768,0.014641778336630927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,6144,0.0806844433148702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,1024,0.010312889185216691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,3584,0.010062221851613786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,3072,0.014196443888876172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,2048,0.025192888246642217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,8192,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,768,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,2560,0.007400888535711501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,8192,0.04503644506136576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,6144,0.4369617833031549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,1024,0.02050044470363193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,5120,0.3185173405541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,2560,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,10240,0.10229155752393936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,12288,0.1328479978773329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,8192,0.030191110240088567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,1024,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,128,0.007100444700982835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,5120,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,2560,0.024101333485709295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,1536,0.04031199879116482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,12288,0.05515733361244202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,2048,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,512,0.029608001311620075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,8192,0.15313243865966797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,2048,0.02095288866096073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,256,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,16384,6.911871168348525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,1536,0.18209155400594076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,768,0.019303111566437613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,8192,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,16384,0.0440977778699663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,128,0.01777688827779558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,2560,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,12288,0.019276445110638935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,6144,0.03161244591077169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,12288,0.10497955481211345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,6144,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,7168,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,51200,1.6958728366427953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,4096,0.06006311045752632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,768,0.008812444077597724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,8192,0.03680888811747233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,6144,0.12253688441382514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,4096,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,65536,0.059141331248813205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,51200,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,4096,0.020430222153663635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,7168,0.02068800065252516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,1536,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,65536,0.27411288685268825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,12288,0.01748533381356133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,6144,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,1024,0.011801777614487542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,3584,0.01124622258875105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,51200,0.1986231141620212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,6144,0.2746257781982422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,3584,0.02698755595419142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,128,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,128,0.02014755540423923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,128,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,12288,0.019650666250122916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,65536,0.08072621954811944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,51200,1.5711547003851996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,16384,0.8618479834662544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,65536,0.4637689060635037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,2560,0.1700355609258016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,1536,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,10240,0.01763288842307197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,64,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,256,0.009935110807418823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,12288,0.02038666605949402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,8192,0.06210844384299385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,768,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,128,0.035840001371171736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,65536,0.11126044061448838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,12288,0.10949155357148911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,10240,0.1511626640955607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,16384,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,3584,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,12288,0.021524444222450256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,3584,0.037626667155159846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,6144,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,1536,0.025444444682863023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,3584,0.040236443281173706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,6144,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,128,0.01552888916598426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,7168,0.1668942239549425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,4096,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,3072,0.021703999903466966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,4096,0.7472044626871744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,5120,0.03802400165134006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,3584,0.008179555336634317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,10240,0.023349333140585158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,3584,0.02920088834232754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,3584,0.029422223567962646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,2048,0.020057777563730877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,12288,0.04679377873738607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,12288,0.06822755601671007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,51200,0.06336000230577257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,51200,1.3040835062662761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,4096,0.01036533299419615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,2048,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,7168,0.23792621824476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,8192,0.01555999947918786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,8192,0.038006222910351224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,5120,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,16384,0.2139564487669203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,8192,0.16145866447024876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,6144,0.06995555427339342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,65536,0.39290489090813535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,512,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,1024,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,256,0.017920888132519193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,128,0.13007022274865046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,768,0.10743466350767349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,51200,0.34229066636827254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,8192,0.11712177594502766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,16384,0.3219839996761746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,5120,0.1269715494579739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,12288,0.39157599873012966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,768,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,10240,0.054541332854164966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,16384,0.020586666133668687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,3584,0.00870222184393141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,3584,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,1536,0.017808000246683758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,7168,0.04336177640491062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,768,0.007780444290902879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,65536,0.21353688504960802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,512,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,1024,0.0032026666320032547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,1024,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,3584,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,2560,0.8015519777933756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,16384,0.060170663727654346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,512,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,6144,0.03636622097757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,5120,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,51200,0.0742044448852539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,2560,0.03500266538725959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,7168,0.04760266674889458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,12288,0.028949333561791316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,3584,0.04486044579082065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,5120,0.025408888856569927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,1024,0.06829688946406047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,5120,0.060418665409088135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,32,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,12288,0.03218133250872294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,256,0.025194666451878015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,7168,0.07071822219424777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,51200,1.0555946562025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,256,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,768,0.003297777846455574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,2560,0.004934222333961063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,256,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,32,0.025358221597141687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,5120,0.00904622260067198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,7168,0.011471110913488599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,1536,0.013198222551080914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,8192,0.07235644261042277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,51200,0.10817244317796494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,512,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,6144,0.01275199982855055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,5120,0.023941333095232647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,10240,0.021480000681347314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,1024,0.017044444878896076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,12288,1.3770204120212133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,65536,0.06344266732533772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,1536,0.01407199932469262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,128,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,8192,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,51200,0.055804444683922656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,1536,0.025091555383470323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,128,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,12288,0.3338008986579047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,256,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,8192,0.18032533592647979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,2560,0.027190221680535212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,6144,0.017948443690935772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,6144,0.03171733352873061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,4096,0.01110311100880305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,256,0.005405333307054307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,768,0.026787555880016748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,256,0.0063822223908371395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,1536,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,256,0.008406221866607666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,5120,0.01668355531162686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,65536,0.12522933218214247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,5120,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,512,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,32,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,12288,0.10945510864257812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,2560,0.27627110481262207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,5120,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,2048,0.015842666228612263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,3072,0.10513866609997219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,10240,0.017383999294704862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,2048,0.017954667409261067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,768,0.013528889252079858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,256,0.011689778003427716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,10240,0.38653333981831867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,512,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,51200,0.11224533451928033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,256,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,2048,0.006719111154476802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,768,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,512,0.014524444937705994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,6144,4.661560906304254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,16384,0.04503022299872505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,65536,0.17421332995096842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,51200,6.833779652913411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,256,0.07366755273607042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,16384,0.27346934212578666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,512,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,1536,0.019840000404251944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,3584,0.010191111101044549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,4096,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,32,0.01053155544731352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,2560,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,128,0.023706666297382776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,16384,0.02499288817246755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,4096,0.07245777712927924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,65536,1.0703378253512912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,1024,0.021312889125612047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,2048,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,1536,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,7168,0.3240328894721137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,12288,0.010572444233629437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,256,0.014853333433469137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,2048,0.07992444435755412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,5120,0.03408977720472548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,4096,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,256,0.1357484393649631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,10240,0.053768889771567456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,768,0.020287111401557922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,8192,0.01551555593808492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,2560,0.4669066535101996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,12288,0.014428445034556918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,512,0.019555555449591745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,4096,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,51200,0.07897155814700656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,10240,0.012808889150619507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,1536,0.5020177629258897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,6144,0.023159111539522808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,6144,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,10240,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,512,0.01591199967596266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,5120,0.05254488852288988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,5120,0.04622222317589653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,2048,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,65536,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,2048,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,3072,0.05524177683724297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,256,0.011704888608720569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,5120,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,5120,0.008479110896587372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,6144,0.014069333672523499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,3072,0.02097333305411869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,5120,0.12217156092325847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,7168,0.023166222704781428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,2560,0.02132266594303979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,16384,0.07292266686757405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,6144,0.03759288787841797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,3584,0.10983733336130778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,10240,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,65536,0.07647289170159234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,32,0.02972444560792711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,7168,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,8192,0.010751999914646149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,128,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,64,0.0029608888758553397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,8192,0.005677333308590784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,2048,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,5120,0.06476622157626681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,65536,0.06984800100326538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,2560,0.039493332306543984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,2048,0.02377333409256405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,12288,0.041335999965667725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,4096,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,1024,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,51200,0.6934675640530057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,1024,0.013983999689420065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,4096,0.30291112263997394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,512,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,6144,0.05010222064124214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,8192,0.051650666528277926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,16384,0.04627822173966301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,16384,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,3072,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,10240,0.17380266719394258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,16384,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,768,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,64,0.009240888886981541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,1536,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,512,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,128,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,4096,0.01849244369400872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,256,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,10240,0.2996257676018609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,16384,0.14232533507876927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,32,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,51200,0.5293368763393825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,10240,0.044228444496790566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,128,0.022640888889630634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,16384,0.021969777014520433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,8192,0.06600266695022583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,2048,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,3072,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,16384,0.2511742115020752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,51200,0.17653155326843262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,16384,0.02369511127471924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,3584,0.008008889026112026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,768,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,7168,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,768,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,6144,0.10810933510462443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,1536,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,3584,0.007409777906205919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,10240,0.005453333258628845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,256,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,7168,0.02477688921822442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,768,0.04927022258440653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,128,0.0017991111510329777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,7168,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,128,0.012812444733248817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,10240,1.8988338046603734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,4096,0.13726045025719538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,7168,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,8192,0.02530577778816223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,5120,0.04812888966666328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,4096,0.02347466680738661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,1536,0.008447999755541483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,16384,0.09799555275175308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,2048,0.04264800084961785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,1024,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,256,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,65536,0.146689772605896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,7168,0.02437955637772878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,51200,0.2202835612826877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,64,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,1024,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,16384,0.02256533337963952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,6144,0.028252445989184912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,16384,0.06006044149398804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,7168,0.023676445086797077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,32,0.008416888614495596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,7168,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,128,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,51200,0.40584977467854816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,1536,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,64,0.0028844444702068963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,256,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,4096,0.01536977787812551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,12288,0.01369511087735494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,51200,0.4544960127936469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,512,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,128,0.011693333586057028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,128,0.006823110911581252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,3584,0.008443555898136562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,8192,0.020998222960366145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,4096,0.015336889359686108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,64,0.005139555368158552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,51200,21.764351738823787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,5120,0.013272888958454132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,51200,0.041972445117102734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,8192,0.030566222137875024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,768,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,768,0.01235466698805491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,512,0.014755555325084262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,768,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,256,0.017847110827763874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,1024,0.006360000206364526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,512,0.006717333363162146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,768,0.014803555276658801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,1536,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,10240,0.00885511106914944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,7168,0.2410044405195448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,3584,0.022498667240142822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,64,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,65536,0.1354737811618381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,6144,0.11758132775624593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,16384,0.03348355491956075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,6144,0.02125866711139679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,1024,0.10599022441440159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,5120,0.30721955829196507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,8192,0.3055653307172987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,6144,0.009538667069541084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,3072,0.01615111033121745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,1536,0.007761778102980719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,3584,0.014681778020328946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,3072,0.017645332548353408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,4096,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,7168,0.03615199857287937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,1024,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,2560,0.15748000144958496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,6144,0.014456000592973499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,51200,0.7470817565917969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,3072,0.02225333286656274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,10240,0.30183998743693036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,16384,0.020861332615216572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,1024,0.02257866660753886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,5120,0.14243111345503065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,512,0.03623822331428528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,10240,0.28038043446011013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,1024,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,128,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,1024,0.026036444637510512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,8192,0.08614933490753174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,128,0.022964444425370958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,1536,0.012404444317022959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,512,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,16384,0.022462222311231825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,12288,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,3072,0.15579377280341253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,2048,0.020689777202076383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,3584,0.03757066528002421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,2048,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,12288,0.5244560241699219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,64,0.1531546645694309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,2048,0.03503911031617059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,12288,0.03682755430539449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,2560,0.01924622224436866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,768,0.023881778120994568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,2048,0.029666665527555678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,8192,0.042760888735453285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,1536,0.03767111235194736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,32,0.06052355633841621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,6144,0.01939733326435089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,8192,0.02204622162712945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,5120,0.011575111084514193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,2048,0.5229448742336696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,5120,0.005049777941571342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,3584,0.02782400117980109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,12288,0.04212711254755656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,256,0.024844444460339014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,128,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,32,0.007357333269384172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,2560,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,1024,0.017954667409261067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,6144,0.03446311089727614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,7168,0.027341332700517442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,10240,0.05012177758746677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,64,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,16384,0.036702222294277616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,7168,0.01891999940077464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,256,0.030877331892649334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,1536,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,8192,0.08857333660125732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,51200,0.8126968807644315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,64,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,512,0.013247110777431063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,2048,0.01901333365175459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,6144,0.021683555510309007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,5120,0.12173155943552653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,7168,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,7168,0.11306488513946533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,4096,0.1133413314819336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,3072,0.04087911049524943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,1024,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,5120,0.09757066435284084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,2560,0.04912977748446994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,1024,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,1536,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,12288,0.01902933253182305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,32,0.36059289508395725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,3072,0.5774933497111002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,2560,0.03090133269627889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,6144,0.020030222005314298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,65536,0.12325955761803521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,5120,0.00535022219022115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,2560,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,3584,0.021781333618693884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,10240,0.01959022217326694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,6144,0.08864177597893609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,12288,0.23835378222995332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,3072,0.03830755419201321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,64,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,3072,0.00851644410027398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,10240,0.08538222312927246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,3584,0.02036177780893114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,5120,0.011382222175598145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,64,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,51200,0.06257155868742201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,12288,11.059343973795572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,1536,0.03594844539960226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,768,0.022634666826989915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,2560,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,2560,0.00907733374171787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,51200,0.41136799918280703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,7168,0.06695910957124498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,6144,0.02149244480662876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,2048,0.05418577790260315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,2560,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,4096,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,3584,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,4096,0.07670311133066814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,51200,11.248859829372831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,65536,0.08251911401748657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,65536,1.1751111348470051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,8192,0.051515556044048734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,1536,0.019914666811625164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,51200,0.6544755299886068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,768,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,12288,0.16148800320095485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,3584,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,2048,0.024255999260478552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,10240,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,768,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,8192,0.11178044478098552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,3584,0.03509599963823954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,16384,0.01926577753490872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,768,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,32,0.04164088765780131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,8192,0.022120889690187242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,3072,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,1024,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,51200,0.05071466830041674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,256,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,2560,0.003926222109132343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,128,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,51200,1.6130861706203883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,2560,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,2560,0.016511110795868766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,1024,0.01576622161600325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,3072,0.024104000793562994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,3072,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,128,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,6144,0.27749421861436635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,51200,0.04614311125543383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,128,0.01648533344268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,1024,0.049231112003326416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,3584,0.011223110887739392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,7168,0.03802666730350918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,6144,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,6144,0.05395822061432732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,4096,0.010301333334710863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,5120,0.021299555897712708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,3584,0.12143554952409531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,7168,0.026191999514897663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,1024,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,7168,0.12182399961683486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,1536,0.018930666976504855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,1024,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,3072,0.03341866532961527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,2560,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,128,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,1536,0.054094221856859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,8192,0.02887644370396932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,1536,0.006121777825885349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,6144,0.16936533980899385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,768,0.03895644346872965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,6144,0.055592887931399874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,64,0.006688000013430913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,65536,0.08298666609658135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,2560,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,3584,0.04927911029921638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,51200,0.13944444391462538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,51200,0.06560444169574313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,3584,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,16384,0.017490666773584154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,12288,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,512,0.02892177634769016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,4096,0.010303999814722273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,2048,0.03269955515861511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,51200,0.046053333414925464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,8192,0.09602310922410752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,5120,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,256,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,8192,0.011707555916574268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,7168,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,1536,0.01594577729701996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,512,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,6144,0.0188773340649075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,2560,0.044396443499459155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,1024,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,51200,0.1390320062637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,6144,0.02148355543613434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,2560,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,128,0.03700266612900628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,256,0.015785778562227886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,64,0.11213689380221897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,65536,0.5531911320156521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,4096,0.028020444843504164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,1024,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,6144,0.10635377963383992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,64,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,2560,0.08481955528259277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,10240,0.5781422191196018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,8192,0.0789288878440857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,512,0.05582844548755222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,51200,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,1024,0.06728355752097236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,1536,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,1536,0.07118844323688082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,4096,0.09858666525946723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,3584,0.0340302222304874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,7168,0.031573332018322416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,51200,0.4843146536085341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,3072,0.14877155092027453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,51200,0.30495643615722656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,3072,0.01757688820362091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,51200,0.18880177868737114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,4096,0.024720888998773363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,8192,0.2534160084194607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,64,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,12288,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,4096,0.03269955515861511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,7168,0.022097777989175584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,3584,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,51200,0.01904266741540697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,512,0.03698488738801744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,512,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,1024,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,5120,0.015535111228624979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,512,0.022156443860795762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,5120,0.02997244397799174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,768,0.02915466825167338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,3072,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,3072,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,7168,0.0218577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,51200,4.941433800591363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,3072,0.05200533403290642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,128,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,3584,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,16384,0.037976887491014265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,3072,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,16384,0.29945243729485405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,2560,0.012015110916561551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,768,0.0027884443600972495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,2048,0.03398399882846408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,2048,0.014095111025704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,65536,0.06159822146097819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,1024,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,256,0.01144444445768992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,7168,0.08628800180223252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,2560,0.00983911090426975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,7168,0.36030133565266925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,10240,0.16041155656178793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,3072,0.02994577752219306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,4096,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,3584,0.11532266934712727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,10240,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,7168,0.023000889354281958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,65536,1.7979395124647353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,1024,0.01890222231547038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,768,0.05935555696487427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,2048,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,65536,0.4246826701694065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,8192,0.18443377812703451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,8192,0.038253333833482527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,7168,0.17556356059180367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,768,0.008033778104517195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,51200,0.539905760023329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,10240,0.27835289637247723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,8192,0.2317670981089274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,8192,0.030590222941504583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,512,0.02748799986309475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,7168,0.051085333029429116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,128,0.014757333530320061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,2560,0.9879840215047201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,3584,0.031141334109836157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,1024,0.011362666885058085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,1536,0.06269244352976482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,6144,0.05829422341452705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,5120,0.027199111051029626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,2560,0.018552000323931377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,2560,0.05078755484686958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,16384,0.5584701961941189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,7168,0.17404266198476157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,5120,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,51200,0.03484977616204156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,12288,0.20576088958316377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,512,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,4096,0.026075555218590632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,51200,3.910558912489149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,6144,0.02808177802297804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,51200,0.17341421710120308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,5120,0.05138488941722446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,1024,1.5204986996120875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,5120,0.03087555699878269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,128,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,2560,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,1024,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,4096,0.0322017769018809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,4096,0.04952977763281929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,256,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,51200,0.027133334014150832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,32,0.00535111129283905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,5120,0.21228533320956758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,3072,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,512,0.021958221991856892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,65536,0.4307733376820882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,256,0.030840890275107488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,128,0.00941955546538035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,16384,0.02240088913175795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,3072,0.06506133079528809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,4096,0.06310400035646227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,5120,0.013090666797426013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,8192,0.016308445069524977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,768,0.003335111050142182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,6144,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,256,0.20616888999938965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,1024,0.009451555709044138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,7168,0.026437333888477747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,1024,0.0032275555034478507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,51200,0.46791553497314453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,1024,0.21225778261820474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,2048,0.013092444174819522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,3072,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,256,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,512,0.012784000072214337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,10240,0.010031110710567897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,12288,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,3072,0.7386435402764214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,1024,0.004448889030350579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,4096,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,128,0.0022088889446523455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,65536,0.09180355734295315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,12288,0.03392622205946181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,6144,0.020647999313142564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,6144,0.038576887713538274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,16384,0.05640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,4096,0.026847999956872728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,512,0.1229119963116116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,32,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,51200,0.7649911244710287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,64,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,1024,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,3584,0.011180444724029966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,3584,0.055121776130464345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,1536,0.11361689037746853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,3072,0.034959110948774554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,2048,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,12288,0.10897688733206855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,10240,0.17761422528160942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,3072,0.06882666879230075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,1024,0.003212444484233856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,2560,0.029084444046020508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,1024,0.007140444384680853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,2048,0.03871999846564399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,2560,0.017274666163656447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,7168,0.03592977921168009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,10240,0.04308177696333992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,1536,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,64,0.005758222192525864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,1536,0.016551999582184684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,16384,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,64,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,51200,0.050388445456822716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,768,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,8192,0.04010399844911363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,3072,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,10240,0.24521154827541777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,32,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,12288,0.08514489067925347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,16384,0.057815108034345836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,7168,0.012700444294346703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,768,0.030738668309317693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,51200,0.5233599874708388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,8192,0.0326533317565918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,65536,0.08638577991061741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,32,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,4096,0.02325155503220028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,3584,0.015221332510312399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,16384,0.037984887758890785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,768,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,1536,0.004452444612979889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,16384,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,7168,0.37142133712768555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,8192,0.044616887966791786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,16384,0.03532177872127957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,768,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,256,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,51200,0.32371287875705296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,8192,0.059396445751190186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,2048,0.036815110180113055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,1536,0.007038222418891058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,6144,0.020610666937298246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,6144,0.03051466743151347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,2048,0.003912000192536248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,16384,0.02093155516518487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,12288,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,3584,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,64,0.010098666780524785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,16384,0.15522844261593288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,12288,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,5120,0.010688888529936472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,6144,0.021672889590263367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,8192,0.04200444287723965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,1536,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,65536,0.14726221561431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,12288,0.04115644428465102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,128,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,1536,0.09807288646697998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,768,0.018961777289708454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,12288,0.187553776635064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,1024,0.02237333357334137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,16384,0.8867342207166883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,768,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,8192,0.02753244505988227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,4096,0.08735022279951309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,768,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,7168,0.08734044763776992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,2048,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,4096,0.07229955328835382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,51200,0.7197697957356771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,65536,0.07254133621851604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,2560,0.042675554752349854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,8192,0.035437332259284124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,12288,0.053968889845742114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,1536,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,256,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,7168,0.009754666851626502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,2560,0.04749600092569987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,8192,0.02999911043379042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,32,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,2560,0.12399200598398845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,65536,0.3496684498257107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,64,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,8192,0.1382746696472168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,8192,0.057555556297302246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,3072,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,768,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,6144,0.013582222163677216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,1536,0.022820444570647344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,2048,0.029115554359224107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,512,0.016140444411171805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,2560,0.008022222254011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,768,0.009552888572216034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,1536,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,12288,0.19024977419111463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,1024,0.12844267156389025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,10240,0.05343288845486111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,768,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,10240,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,7168,0.11497333314683701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,3072,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,6144,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,4096,0.020727111233605277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,1024,0.0070471109615431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,51200,0.28682756423950195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,3584,0.018018666240904067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,256,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,3584,0.013754666679435305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,8192,0.19049955738915336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,256,0.023014222582181294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,10240,0.02510844502184126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,128,0.010207999911573198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,5120,0.01611911091539595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,512,0.01719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,5120,0.24247911241319445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,8192,0.06907822026146783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,768,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,1536,0.03854399919509888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,5120,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,10240,0.026449778013759192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,65536,0.0662826697031657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,1024,0.015582222077581616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,4096,0.02805688977241516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,6144,0.02443733314673106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,7168,0.0758533345328437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,512,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,6144,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,1024,0.036834667126337685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,512,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,65536,0.03702222307523092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,7168,0.024821332759327356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,5120,0.09904888603422378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,2560,0.1102000011338128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,4096,0.00814577771557702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,1536,0.012464000119103326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,51200,6.495551215277778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,512,0.034272889296213783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,512,0.015212444795502557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,1536,0.023901333411534626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,16384,0.0617191129260593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,1024,0.0063777777055899305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,128,0.02351733379893833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,4096,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,3584,0.019161777363883126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,2560,0.025557332568698462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,1024,0.022091555926534865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,6144,0.013365333278973898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,64,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,5120,0.0090524446633127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,1024,0.010613333019945355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,5120,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,5120,0.0196142229768965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,512,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,12288,0.6940915319654676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,1536,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,51200,0.21629067262013754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,3072,0.012260444462299347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,5120,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,2560,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,1536,0.0262080000506507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,6144,0.011215111447705163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,65536,0.05267644590801663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,2048,0.013591999808947245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,256,0.06209422482384575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,1536,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,16384,0.029692444536421034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,6144,0.015611555841233997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,32,0.013995556367768182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,10240,0.053567998939090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,7168,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,3072,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,12288,0.05110133356518216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,512,0.019280888968043856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,5120,0.05031911200947232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,1024,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,16384,0.034669333034091525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,65536,0.07532799906200834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,6144,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,12288,0.04664177695910136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,3072,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,8192,0.05122311247719658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,1536,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,1536,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,5120,0.038443555434544883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,16384,0.021715554926130507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,65536,0.8081208864847819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,2560,0.008416888614495596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,64,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,2560,0.013904000322024027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,10240,0.10156889094246759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,3072,0.023950222465727065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,6144,0.03596977723969354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,12288,0.013639999760521783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,3584,0.08107822471194796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,51200,0.11577510833740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,4096,0.0325217776828342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,2048,0.00452533322903845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,16384,2.884376949734158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,51200,0.08745777606964111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,65536,0.22619199752807617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,12288,0.06993955373764038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,16384,0.020899555749363367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,5120,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,7168,0.0355004436439938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,4096,0.034295111894607544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,32,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,12288,0.017197334104114108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,768,0.020436444216304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,51200,0.21655821800231934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,12288,0.05301422211858961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,768,0.009544889132181803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,65536,0.04655111167165968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,16384,0.015770667129092746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,64,0.00508266646001074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,1024,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,2560,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,16384,0.03348799877696567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,6144,0.6577617857191298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,768,0.04049777653482225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,2048,0.07516888777414958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,2560,0.10633599758148193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,2560,0.01611199975013733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,8192,0.044347554445266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,5120,1.8168409135606554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,65536,0.33609334627787274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,32,0.008389333056079017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,8192,0.04016177852948507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,2560,0.005700444595681296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,51200,0.05847644143634372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,65536,0.06430755721198188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,768,0.29593510097927517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,3072,0.03529955612288581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,128,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,2560,0.027854220734702215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,512,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,3584,0.3015635543399387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,5120,0.013267555170589022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,6144,0.00908622228437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,3584,0.06634666522343953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,3072,0.07181600067350599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,8192,0.15898756186167398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,256,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,8192,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,16384,0.8185440169440376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,2048,0.09578577677408855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,3072,0.00812533332241906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,1024,0.0225591113169988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,16384,0.07470399803585477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,1024,0.015190222197108798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,4096,0.07630311118231879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,12288,0.42201688554551864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,8192,0.026183999247021143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,7168,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,2048,0.005980444451173146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,12288,0.018888889087571036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,16384,0.026125333375400964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,16384,0.02277155551645491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,64,0.0029191111938820947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,3584,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,3072,0.01052177780204349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,7168,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,16384,0.039817776944902204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,512,0.002883555574549569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,768,0.02181688944498698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,12288,0.08874666690826416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,5120,0.02128355536195967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,128,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,1024,0.019277332557572257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,10240,0.024746666351954143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,2560,0.0125937776433097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,2560,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,512,0.059227552678849965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,3072,0.010434666441546546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,10240,0.02688177757793003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,6144,0.03937155670589871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,8192,0.023404444257418316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,128,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,3072,0.007471111085679796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,12288,0.17335289054446748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,3072,0.024285333024130926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,256,0.02872266703181797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,12288,0.058456891112857394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,16384,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,128,0.02243822150760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,2048,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,768,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,2048,0.009582222335868413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,128,0.013265777793195514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,768,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,10240,0.019692444139056735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,3072,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,8192,0.2251439889272054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,3584,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,65536,0.08503733078638713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,1536,0.006319111006127463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,6144,0.015769778026474845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,8192,0.03787822193569607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,128,0.0030293334275484085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,256,0.00637511122557852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,3072,0.06851822137832642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,5120,0.05933155616124471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,3584,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,4096,0.011560888753996955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,4096,0.02474133339193132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,512,0.015543111496501498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,128,0.01625333395269182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,256,0.019326221611764696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,64,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,256,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,1536,0.008270222279760573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,3584,0.015743111570676167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,16384,0.060079998440212674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,3072,0.00962755580743154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,65536,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,5120,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,128,0.011015999648306104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,1536,0.04127555423312717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,3072,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,3072,0.47725243038601345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,5120,0.011523555550310345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,256,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,6144,0.21918043825361463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,3584,0.049309333165486656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,3072,0.027641778190930683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,4096,0.0075137780772315125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,128,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,3584,0.033597333563698664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,256,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,8192,0.01576622161600325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,512,0.008050666915045844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,256,0.023569777607917786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,128,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,3072,0.042711999681260854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,4096,0.02055911057525211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,3584,0.10336622264650132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,4096,0.013946667313575745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,5120,0.029493331909179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,5120,0.0617920027838813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,2048,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,3584,0.027633777923054163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,7168,0.3615439997778998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,51200,1.339712036980523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,1024,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,16384,0.2788275612725152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,1536,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,6144,0.031488001346588135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,10240,0.07848799890942044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,128,0.22059644593132866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,6144,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,1536,0.07039288679758708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,512,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,1536,0.08318755361768934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,1536,0.00424888895617591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,6144,0.019096000327004325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,256,0.00416088890698221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,3072,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,128,0.004876444323195351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,256,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,65536,0.3017164336310493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,768,0.005744888964626525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,1024,0.020299555526839364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,128,0.018215111560291715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,4096,0.05683288971583048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,2560,0.030751112434599135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,10240,0.4167528947194417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,7168,0.028619554307725694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,5120,0.01827111177974277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,65536,0.06650844547483656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,8192,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,3584,0.027814222706688776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,12288,0.030744890371958416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,51200,2.1113929748535156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,6144,1.275686264038086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,1024,0.030127998855378892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,5120,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,1024,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,12288,0.1870622237523397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,3584,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,3584,0.061583108372158475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,16384,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,7168,0.02641422218746609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,256,0.08337599701351589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,8192,0.027237332529491846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,2560,0.025791111919615004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,8192,0.057533330387539335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,3072,0.01050222251150343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,65536,0.12299022409651016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,16384,3.866668277316623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,2048,0.021372444099850122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,32,0.0026933333526055017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,16384,3.2771244049072266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,51200,0.8509377903408475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,51200,0.1365742286046346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,5120,0.004315555509593752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,1536,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,512,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,4096,0.03164622187614441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,8192,0.011389333340856763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,51200,0.024493333366182115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,1024,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,64,0.08568622006310357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,51200,0.3237911065419515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,6144,0.024681778417693243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,256,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,1024,0.032588445478015475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,128,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,1024,0.013628444737858243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,7168,0.022409778502252366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,1024,0.3640408780839708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,10240,0.030756443738937378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,256,0.003257777748836411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,65536,0.08888533380296494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,512,0.1515608893500434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,128,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,6144,0.027076444692081873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,2048,0.01886311173439026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,1024,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,2560,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,4096,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,65536,0.9768853717380099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,512,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,8192,0.026378666361172993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,768,0.004295999805132548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,65536,0.022778666681713525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,3072,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,10240,0.12031466431087917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,32,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,7168,0.026479111777411565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,51200,0.3227653238508436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,65536,1.1141022576226127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,65536,0.18166311581929526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,3072,0.010075555907355415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,2560,0.025012445118692186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,32,0.005379555539952383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,1024,0.040264000495274864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,7168,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,51200,0.08152888880835639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,3584,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,6144,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,10240,0.21503822008768717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,2560,0.01926666663752662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,5120,0.04871111114819845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,1024,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,1024,0.0032133333798911837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,65536,0.8487448692321777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,3584,0.27655643886990017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,6144,0.019847111569510568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,4096,0.016688000824716356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,10240,0.025682666235499915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,6144,0.01924888955222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,16384,0.021573333276642695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,5120,0.07419021924336751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,3072,0.044569777117835156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,128,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,1024,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,3584,0.015808888607554965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,65536,1.6078231599595811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,2560,0.02584533393383026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,10240,0.02049066623051961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,32,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,12288,0.17161156071556938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,51200,0.06307378080156115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,2560,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,3072,0.009543110926946005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,256,0.01796800063716041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,256,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,12288,0.12245066960652669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,2048,0.01350488927629259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,2560,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,256,0.03159288896454705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,768,0.02295822236273024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,1536,0.050408889849980675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,10240,0.02553155521551768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,1536,0.04518577787611219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,5120,0.01757777730623881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,7168,0.05737600061628553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,2048,0.010497777826256223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,7168,0.09188889132605658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,6144,0.27451467514038086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,10240,0.28204533788892955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,10240,0.07083911365932889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,512,0.03426400158140395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,64,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,51200,0.1602168877919515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,51200,0.06596622202131483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,128,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,6144,1.2475449244181316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,6144,0.2751591205596924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,256,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,16384,0.2619724538591173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,1536,0.006090666684839461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,4096,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,128,0.002307555534773403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,3584,0.032371554109785296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,1536,0.00639111093348927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,2048,0.019335110982259113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,2048,0.011981333295504252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,4096,0.08728977706697251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,2560,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,128,0.009511999785900116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,2048,0.031968000862333514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,768,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,3072,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,5120,0.020177778270509508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,256,0.011173333558771344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,128,0.015660444895426433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,7168,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,1536,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,512,0.02145511077509986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,4096,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,51200,0.05036977926890055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,12288,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,1536,0.4983661969502767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,8192,0.024041778511471216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,12288,0.021872000561820135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,8192,1.1532346937391493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,1536,0.005192888693677055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,16384,0.030919111437267725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,2048,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,12288,0.029114666912290785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,512,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,8192,0.04509688748253716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,1024,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,128,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,51200,0.676886240641276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,8192,0.03903733359442817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,8192,0.0779306689898173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,4096,0.08535733487870957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,6144,0.1146355602476332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,10240,0.012048000262843238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,10240,1.5466666751437719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,128,0.007913777397738563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,12288,0.3572399881150987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,32,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,3584,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,3072,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,1536,0.01386044422785441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,2048,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,10240,0.9175457424587674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,12288,0.014426666829321118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,12288,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,10240,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,65536,0.046909332275390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,51200,0.5382186571756998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,12288,0.014039110806253223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,51200,0.05306488937801785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,10240,0.0466177761554718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,1024,0.12465866406758626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,65536,3.6613252427842884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,10240,1.8445538414849176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,10240,0.036329779360029436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,4096,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,512,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,256,0.005489777773618698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,128,0.016167999969588388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,128,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,51200,0.050669332345326744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,8192,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,65536,0.07821955283482869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,10240,0.025248888466093276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,65536,3.806342230902778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,128,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,512,0.032842665910720825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,65536,0.6523706648084853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,2560,0.012198222180207571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,65536,0.376909335454305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,7168,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,3584,0.6354702313741049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,4096,0.05545777744717068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,2560,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,256,0.018135999639829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,2048,0.04350577791531881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,768,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,3072,0.060544888178507485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,2048,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,768,0.02921688887808058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,1536,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,8192,13.653199089898003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,16384,1.5022399690416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,1024,0.008361777497662438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,4096,0.012651555240154266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,1024,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,8192,0.039834668238957725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,3584,0.007720888488822513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,1536,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,6144,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,256,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,2048,0.014175110393100314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,2048,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,2560,0.09102221992280747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,3584,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,7168,0.05069244570202298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,256,0.0029306666304667792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,16384,0.11040800147586399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,10240,0.01625955601533254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,3072,0.02310488952530755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,8192,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,3584,0.10398399829864502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,10240,0.02055288851261139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,5120,0.009451555709044138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,768,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,6144,2.329031202528212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,51200,0.7944995562235514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,12288,0.01663733356528812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,8192,0.05053955647680494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,4096,0.07003555695215861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,6144,0.2621244324578179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,3072,0.011363555159833698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,256,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,2048,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,128,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,512,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,1024,0.03636444608370463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,5120,0.021672889590263367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,1536,0.0188044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,4096,0.08463999960157607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,12288,0.020289777053727042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,64,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,768,0.0063662222690052455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,6144,0.028723554478751287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,8192,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,7168,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,1536,0.018886221779717337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,1024,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,6144,2.8250702752007375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,256,0.017571555243598092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,256,0.0031537777847713898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,16384,1.5265057881673176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,128,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,3584,0.04127644499142965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,8192,0.022092445029152766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,1024,0.045548445648617215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,256,0.00792266676823298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,2048,0.044475555419921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,2560,0.02456888887617323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,256,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,12288,0.1601573361290826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,2048,0.12100533644358318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,4096,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,768,0.01568177839120229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,5120,0.0196426659822464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,65536,0.14775555663638645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,2560,0.011807111402352652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,12288,0.2897057798173693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,256,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,1536,0.022292445103327434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,32,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,12288,0.053723557127846606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,2560,0.006672888994216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,10240,0.029951112137900457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,7168,0.00489155575633049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,4096,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,1536,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,2560,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,2560,0.0752426650789049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,8192,0.023988443944189284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,12288,0.024756444825066462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,5120,0.15429777569240993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,768,0.028568890359666612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,1024,0.21858489513397217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,3584,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,2048,0.0074035558435651995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,5120,0.04331466555595398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,2560,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,65536,0.1476951175265842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,1024,0.10030311346054077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,6144,0.029480000336964924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,8192,0.021760889225535925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,4096,0.010123555858929953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,16384,0.03494222296608819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,10240,1.9080986446804469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,3584,0.03435733252101474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,6144,0.012981333666377597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,2048,0.09208089113235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,8192,0.01536622146765391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,51200,0.5342995325724283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,1024,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,16384,0.2600319915347629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,128,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,256,0.02124622298611535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,10240,0.011679111255539788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,64,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,2560,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,3584,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,65536,0.06490577591790093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,1536,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,2560,0.04035288757748074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,1536,0.016583111551072862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,5120,0.02485688858562046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,128,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,4096,0.22239910231696236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,32,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,51200,0.11614844534132217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,3584,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,16384,0.14090043968624538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,65536,0.07071022192637126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,6144,0.007679999702506595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,3072,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,16384,0.02276355524857839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,2048,0.01334755536582735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,1536,0.011337777806652917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,4096,0.22186666064792207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,1024,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,512,0.07219822539223565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,65536,5.210476345486112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,7168,0.02012711101108127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,65536,0.2029715511533949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,3584,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,51200,0.841209782494439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,768,0.019384000036451552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,12288,0.07564800315433078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,51200,0.4789395332336426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,10240,0.03326755430963304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,16384,0.037217779292000666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,256,0.005315555466545953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,16384,0.23880177074008516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,1536,0.03916888766818576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,1024,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,7168,0.028040889236662123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,512,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,128,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,128,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,8192,0.026548445224761963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,5120,0.011531555818186866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,3072,0.11983022424909805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,1024,0.021847999758190576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,3072,0.015217777755525378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,16384,2.2184151543511286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,16384,0.025075554847717285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,3072,0.01633066601223416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,10240,0.04821155468622843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,16384,1.0114427142673068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,8192,0.05134310987260607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,512,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,512,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,12288,0.03200622068511115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,64,0.014206222361988492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,10240,0.05458844370312161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,3584,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,51200,1.0938249164157443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,2560,0.02384800049993727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,6144,0.008746667040718926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,16384,0.03316977620124817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,3584,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,512,0.0030382221771611106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,6144,0.022278222772810195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,768,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,10240,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,1024,0.007080000307824876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,2048,0.0068693334857622785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,128,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,2560,0.030256887276967365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,2560,0.07938755883110894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,128,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,10240,0.015410666664441427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,12288,0.2576826678382026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,1536,0.00665155549844106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,1536,0.008397333323955536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,16384,0.43261154492696124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,4096,0.0934728913837009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,256,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,32,0.0018666666001081467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,2560,0.0480640000767178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,1536,0.01475644442770216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,256,0.08023644155926175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,65536,1.0024329291449654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,3584,0.15117511484358045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,1024,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,64,0.003847111016511917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,4096,0.02860977914598253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,768,0.011567110816637674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,12288,0.039975110027525164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,3072,0.018280888597170513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,1536,0.015415110521846347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,32,0.0018764444523387486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,7168,0.03614755471547445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,1536,0.017980444762441847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,512,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,7168,1.0704986784193251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,512,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,1536,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,3584,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,10240,0.05003822180959913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,8192,0.045422222879197865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,65536,0.06409688790639241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,65536,0.805024888780382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,4096,0.04717333449257744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,768,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,2048,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,5120,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,1024,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,6144,0.31339555316501194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,10240,0.025382222400771245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,8192,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,64,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,7168,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,768,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,12288,0.04052088989151849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,1536,0.12066400051116943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,256,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,6144,0.018608889646000333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,512,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,3584,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,256,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,3072,0.007431999676757389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,4096,0.08017066452238295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,2560,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,51200,0.09906399912304348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,512,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,6144,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,6144,0.06737155384487577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,1536,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,512,0.019898666275872123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,65536,1.479337798224555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,10240,0.14469156000349256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,5120,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,3072,0.02458400030930837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,8192,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,65536,1.6400221718682184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,51200,20.090282864040798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,512,0.10370400216844346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,2048,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,51200,0.025920000341203477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,12288,0.05165955424308777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,3584,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,1536,0.015221332510312399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,51200,0.06793777810202704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,5120,0.011644444531864591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,12288,0.20900000466240776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,12288,0.13195021947224936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,65536,0.7120604515075684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,3072,0.00601244428091579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,768,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,2048,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,3584,0.03604711095492045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,6144,0.1101271112759908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,5120,0.04567733407020569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,6144,0.017795556121402316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,7168,0.011903111305501727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,128,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,51200,0.05673066775004069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,4096,0.028196444114049275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,3072,0.04453066653675503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,32,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,1024,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,5120,0.017320000463061862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,8192,0.023369777533743117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,1536,0.03376177615589566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,65536,1.0053457683987088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,2048,0.023344000180562336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,16384,0.5351324611239964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,128,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,4096,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,256,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,32,0.0024560000747442245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,8192,2.681499481201172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,3072,0.04122311207983229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,16384,0.042912890513738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,2048,0.025735111700163946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,768,0.013463999662134381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,2560,0.014299554957283868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,12288,0.02535022298494975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,3584,0.058992001745435924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,12288,0.031494223409228854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,256,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,1024,0.022316444251272414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,1536,0.010815111299355825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,2048,0.028766221470303003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,10240,0.10184888707266913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,6144,0.15057599544525146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,256,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,16384,0.04393066631423103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,16384,0.0866417752371894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,64,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,64,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,51200,0.0239982224173016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,8192,0.036601778533723615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,6144,0.0382008883688185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,3072,0.013552889227867126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,5120,0.17283822430504692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,512,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,768,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,64,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,2560,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,16384,0.06855199734369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,256,0.0032302221904198327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,2048,0.031770666440327965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,1536,0.029324445459577773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,2560,0.004853333450025982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,512,0.1278417772716946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,128,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,8192,0.15738399823506674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,10240,0.031974222924974226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,10240,0.05829777982499865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,3584,0.00776800016562144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,2560,0.03042844600147671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,1536,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,3584,0.028534223635991413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,4096,0.016388444436921012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,7168,0.06331911351945665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,8192,0.01993688941001892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,4096,0.025223111112912495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,10240,0.08341688579983181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,12288,0.10666488938861424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,7168,0.1564426687028673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,1024,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,1024,0.004763555609517627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,2560,0.02001688877741496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,2048,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,2048,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,2560,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,2560,0.011612444288200803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,3072,0.01000088867213991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,5120,0.017277333471510146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,5120,0.01789955629242791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,768,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,10240,0.1941279967625936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,512,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,51200,0.7959298027886285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,1536,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,65536,0.3963510990142822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,128,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,256,0.01905066602759891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,7168,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,51200,0.11656533347235785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,3072,0.009102221992280748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,5120,0.017733333839310538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,768,0.00812266684240765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,512,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,5120,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,2560,0.015541333291265698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,768,0.157331559393141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,3584,0.01182666669289271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,8192,0.01590399940808614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,10240,0.023997333314683702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,256,0.02169955604606205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,2560,0.026405332816971674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,3584,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,2048,0.02032888929049174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,256,0.007505777809354994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,16384,0.055832001898023814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,3072,0.03869333200984531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,64,0.0021031111892726687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,768,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,2048,0.07401600148942736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,16384,0.14486577775743273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,5120,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,5120,0.03894666830698649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,64,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,128,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,32,0.010408889088365767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,1024,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,8192,0.029157333903842505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,12288,0.25735288196139866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,2560,0.023016000787417095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,32,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,128,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,10240,0.009963555468453301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,16384,0.01332711097266939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,8192,0.031141334109836157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,16384,0.08066222402784559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,3072,0.03768800033463372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,4096,0.07378044393327501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,128,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,1536,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,256,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,32,0.007448888487286038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,4096,0.013978666729397245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,1536,0.07692622476153903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,1024,0.01945066617594825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,3072,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,65536,0.1386373307969835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,1536,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,16384,0.059135999944474965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,768,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,128,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,2560,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,64,0.7126426696777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,1024,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,2048,0.1342195537355211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,12288,0.06784266895718045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,512,0.013471999930010902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,16384,0.06883822547064887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,16384,0.21624355845981172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,5120,0.0485208895471361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,7168,0.15204444196489122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,16384,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,4096,0.05224800109863281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,2560,0.033519999848471746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,2048,0.036622222926881574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,512,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,64,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,256,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,12288,0.3032160070207384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,2048,0.45124175813462997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,8192,0.6428293122185601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,256,0.009626666704813639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,256,0.0288702216413286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,8192,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,10240,0.020793777373101976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,512,0.015110222829712762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,2560,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,1536,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,51200,0.04603200157483419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,3584,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,3584,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,5120,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,5120,0.044235554006364614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,3584,0.028652442826165095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,3072,0.0339937773015764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,64,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,12288,0.0376160012351142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,12288,0.12643110752105713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,1024,0.04083022144105699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,4096,0.035014222065607704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,51200,0.7901520199245877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,256,0.015582222077581616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,256,0.004899555610285865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,256,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,2048,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,2560,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,2560,0.02782400117980109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,1024,0.09399021996392144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,2048,0.10783911413616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,2560,0.007087111473083496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,1536,0.010767111347781287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,768,0.028170665105183918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,16384,0.01885600056913164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,12288,0.019898666275872123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,51200,0.41643643379211426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,16384,0.0855697790781657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,5120,0.017828444639841717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,8192,0.18409600522783068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,256,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,16384,0.040491554472181536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,512,0.03648533423741659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,16384,1.3006426493326824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,32,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,7168,0.022276444567574397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,128,0.008351999852392409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,3072,0.032256888018714056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,768,0.03467377689149644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,65536,2.4762017991807723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,1536,0.02053244411945343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,768,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,64,0.0024053334361977046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,3584,0.3381928867763943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,8192,10.382860819498697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,1536,0.009076444639099969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,512,0.18503733476003012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,512,0.011874666644467248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,2048,0.03630222214592828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,10240,0.027291556199391682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,5120,0.09596088859770034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,768,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,128,0.0053680001033677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,65536,0.6434053315056695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,4096,0.5464595688713921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,7168,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,2560,0.009280888570679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,1024,0.025719111164410908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,65536,0.7022604412502713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,4096,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,51200,0.07554044326146443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,4096,0.056448002656300865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,512,0.021352888809310064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,512,0.0820666684044732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,65536,0.0373253325621287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,6144,0.09283555878533258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,3072,0.01815111107296414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,4096,0.015533333023389181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,1536,0.025050666597154405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,12288,0.08113155762354533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,65536,10.781731499565971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,51200,0.19858755005730522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,12288,0.036561777194341026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,16384,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,5120,0.03393955694304572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,5120,0.01773155563407474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,32,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,8192,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,1024,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,3072,0.0437102218468984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,2048,0.015232000086042615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,65536,2.030674616495768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,10240,0.038945777548684016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,8192,0.021406221720907424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,3072,0.09317244423760308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,64,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,65536,0.19059822294447157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,16384,0.06642133659786649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,128,0.015433778365453085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,768,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,5120,0.0236462222205268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,8192,0.1497973336113824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,64,0.006720888945791457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,4096,0.03882755504714118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,16384,0.020899555749363367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,51200,0.03820000092188517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,2048,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,2560,0.04123555620511373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,128,0.00226488895714283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,3072,0.020251555575264823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,1536,0.02660888930161794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,768,0.012074666718641916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,768,0.006108444597986009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,128,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,2048,0.004388444539573458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,6144,0.051911999781926475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,3072,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,32,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,32,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,7168,0.15482044219970703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,768,0.02056799994574653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,512,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,512,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,2048,0.03843555516666836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,10240,2.109972423977322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,2560,0.1913226710425483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,51200,0.8762453397115072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,128,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,128,0.017879999346203275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,2048,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,8192,0.14603732691870794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,2048,0.01921333372592926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,2560,0.06321955389446683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,5120,0.019323555959595572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,1024,0.003906666818592283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,5120,0.007071110937330458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,8192,0.02401777770784166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,768,0.023444443941116333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,16384,0.043939554029040866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,12288,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,6144,0.02012266715367635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,1536,0.025951999757024977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,51200,0.308680878745185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,3584,0.05799733267890083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,5120,0.03460888730155097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,16384,0.022435555855433147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,6144,0.07605599694781832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,1024,0.03279022375742594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,65536,3.409969753689236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,1024,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,1536,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,1024,0.033562666840023465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,7168,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,1536,0.07555199993981256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,128,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,3072,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,2048,0.02234755622016059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,2560,0.04051377707057529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,3072,0.060469335979885526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,65536,0.06173066960440742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,4096,0.03490488727887472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,64,0.0047582222355736625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,128,0.013197333448463015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,65536,0.5769182311164008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,65536,0.07164088884989421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,32,0.00573511090543535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,256,0.012485333614879184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,3072,0.07099022467931111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,12288,1.5257395638359919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,1536,0.020591111646758187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,2560,0.019366222951147292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,4096,0.02146133283774058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,1536,0.05449155635303921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,32,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,51200,0.16166933377583823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,32,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,10240,0.020690666304694284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,12288,0.03047910994953579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,5120,0.00887288898229599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,16384,0.028604444530275133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,8192,0.023072888453801472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,256,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,12288,0.06865421930948894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,12288,0.44112088945176864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,16384,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,65536,0.025436444414986506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,51200,0.10642133156458537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,16384,0.021677333447668288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,10240,0.3685164451599121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,65536,0.9076693852742513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,32,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,6144,0.045888887511359326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,3072,0.0484791100025177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,2048,0.018416888184017606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,128,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,5120,0.03763733307520548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,7168,0.027484445108307734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,2048,0.1013724472787645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,8192,0.04714044597413805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,32,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,512,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,3072,0.04158400164710151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,3072,0.12141688664754231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,10240,0.031143999762005273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,1024,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,2048,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,65536,0.04742666747834948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,2048,0.1837537818484836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,128,0.030173334810468886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,1024,0.03728444377581278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,3072,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,7168,0.05588089095221626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,4096,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,12288,0.4062995645734999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,1536,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,256,0.02221422228548262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,1024,0.016203555795881484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,1536,2.034063975016276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,8192,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,256,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,32,0.009331555830107795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,51200,0.5704328748914931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,12288,1.9541333516438801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,65536,0.10564977592892116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,8192,0.04508622156249153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,5120,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,3072,0.060630222161610924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,512,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,1024,0.016398222910033334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,3584,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,1536,0.027830223242441814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,7168,0.06851910882525973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,256,0.00254488881263468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,64,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,4096,0.12245599428812663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,51200,0.020040889581044514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,3072,0.008049777812427944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,2560,0.012798222402731577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,10240,0.05176000131501091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,1536,0.007450666692521837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,2048,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,2560,0.015844444433848064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,8192,0.11161600218878852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,10240,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,3584,0.018329777651362948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,256,0.003363555504216088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,5120,0.19444621933831108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,65536,1.169097794426812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,2560,0.038078222009870745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,7168,0.006999111009968652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,256,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,4096,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,768,0.02929511004024082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,512,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,12288,0.030699554416868422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,256,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,12288,0.025411556164423626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,1024,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,3584,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,7168,0.01254755589697096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,3584,0.029655112160576716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,128,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,51200,0.10023822387059529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,16384,0.02472266720400916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,3072,0.008207110895050896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,128,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,7168,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,2048,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,1536,0.02810844447877672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,3584,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,2048,0.07510933611128065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,10240,0.12795733080969915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,4096,0.025380444195535447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,5120,0.10841688844892715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,7168,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,128,0.014893333117167154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,128,0.024685333172480266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,512,0.047129776742723256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,7168,0.07657866345511542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,8192,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,12288,0.08631822135713364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,6144,0.01054933336046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,512,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,2560,0.015792888071801927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,7168,0.2509048779805501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,16384,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,65536,2.5858311123318143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,51200,0.1317733393775092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,5120,0.016333333320087857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,256,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,7168,0.010547555155224271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,3584,0.08476710981792873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,5120,0.14882043997446695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,65536,0.09823111030790542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,7168,0.026893332600593567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,4096,0.021740444832377966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,256,0.026770666241645813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,2048,0.030008001459969416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,10240,0.3813110987345378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,10240,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,128,0.04702044526735941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,512,0.005533333453867171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,10240,0.0247013337082333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,12288,0.15304711129930285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,3072,0.007567110988828871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,12288,0.2300959957970513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,2048,0.00739911115831799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,10240,0.016338666280110676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,4096,0.37178932295905215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,3584,0.007886221839321984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,10240,0.040376888381110296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,8192,0.1743368837568495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,512,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,10240,0.513780434926351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,51200,0.05243022243181864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,6144,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,6144,0.00777599960565567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,5120,0.275872892803616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,1024,0.025263110796610515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,10240,0.017902221944597032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,8192,0.5750337706671821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,2048,0.373656882180108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,768,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,10240,0.05299555593066745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,3072,0.03046222196684943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,32,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,6144,0.02259822189807892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,12288,0.01965777741538154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,2048,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,1024,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,16384,0.18099821938408744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,3584,0.09904978010389541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,512,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,768,0.02494755552874671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,2048,0.004238222208287981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,6144,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,7168,0.05986933575736152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,128,0.03153955605294969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,3584,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,64,0.01680266691578759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,65536,0.11780444780985515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,3072,0.03809422254562378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,3584,0.03966222206751505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,3584,0.008929777476522658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,256,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,16384,0.018544000056054857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,64,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,1024,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,3584,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,8192,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,4096,0.061400002903408475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,16384,0.02221244408024682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,3584,0.018203554881943595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,2560,0.023802666200531855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,7168,0.03143289022975498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,2560,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,5120,0.02871733241611057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,4096,0.016175111134847004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,512,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,2560,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,1536,0.013949332965744866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,51200,0.2342160012986925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,6144,0.009522666533788046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,4096,0.015273777974976433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,512,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,4096,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,51200,1.4529715643988714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,12288,0.013169777890046438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,16384,0.03136266602410211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,10240,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,2048,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,5120,0.026830222871568467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,128,0.0019164443429973389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,2048,0.4044177797105577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,3584,0.035651554663976036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,1536,0.006839111033413146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,64,0.009043555292818282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,768,0.18331289291381836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,2048,0.011794666449228922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,5120,0.20748089419470894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,2560,0.031073778867721558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,2048,0.027952889601389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,3072,0.02105955613984002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,2560,0.10974666807386611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,7168,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,4096,0.03323910964859857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,65536,0.1039502223332723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,512,0.002549333290921317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,1536,0.03250133328967624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,3584,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,1024,0.0032026666320032547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,4096,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,2048,0.08735377920998467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,10240,0.03395466672049628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,512,0.04055644406212701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,3072,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,10240,0.21319466167026094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,3584,0.012999110751681857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,128,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,10240,0.03314133485158285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,4096,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,51200,0.26917865541246205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,51200,3.893293168809679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,6144,0.020076443751653034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,7168,0.03433777888615926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,8192,0.05488888753785027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,1024,0.01791111131509145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,512,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,2048,0.020234665936893888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,6144,0.12776711252000597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,12288,0.04121422105365329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,16384,0.3709751235114203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,3072,0.007684444387753804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,12288,0.03886133432388306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,2048,0.006501333167155583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,16384,0.2978595627678765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,4096,0.4809102482265896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,1024,0.024633778466118708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,128,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,2560,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,3072,0.04599377844068739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,16384,0.28044178750779897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,6144,0.13259288999769422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,65536,0.1522231101989746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,7168,0.40475111537509495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,7168,0.1157804462644789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,1536,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,51200,0.020210666788948905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,3584,0.03499733408292135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,128,0.022656889425383672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,128,0.03253066539764404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,7168,0.017030222548378837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,1024,0.07320266962051392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,64,0.006780444333950679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,512,0.018768888380792405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,2560,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,768,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,5120,0.038683556848102145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,2048,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,512,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,64,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,10240,0.02960977620548672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,10240,0.006505777852402792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,64,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,768,0.08383555544747247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,7168,0.029623998536003962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,1536,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,6144,0.02590399980545044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,8192,0.014602666099866232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,12288,0.02879377868440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,768,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,3072,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,7168,0.02146133283774058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,51200,0.05403288867738512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,7168,0.041336887412601046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,2560,0.012489777472284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,768,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,2560,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,51200,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,3072,0.025067556235525344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,7168,0.14127822717030844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,8192,0.010990222295125326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,8192,0.018571555614471436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,3072,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,6144,0.02982399861017863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,128,0.005708444449636672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,1536,0.0034862222770849862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,51200,0.3215297857920329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,64,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,256,0.014106666048367819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,4096,0.01908533275127411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,1536,0.009542221824328104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,65536,0.2534595595465766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,4096,0.15823466247982448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,256,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,3072,0.039206223355399236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,512,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,768,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,8192,0.034332444270451866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,2560,0.00777422222826216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,2560,0.5399093098110622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,3072,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,256,0.017479999197853934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,32,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,512,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,6144,0.025895110434956018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,10240,0.00962311112218433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,51200,0.05294488867123922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,6144,0.21222221851348877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,1024,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,12288,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,512,0.031932443380355835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,16384,0.1868213282691108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,1536,0.022310222188631695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,16384,0.09104533327950372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,16384,0.028247998820410833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,5120,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,3072,0.013948443863126965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,65536,0.17159199714660645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,64,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,10240,0.31099645296732586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,1024,0.0131217779384719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,128,0.010710222025712332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,16384,0.03187377916442023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,128,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,10240,0.0355662206808726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,512,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,64,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,2560,0.019722667005327012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,128,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,7168,0.0657679968410068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,6144,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,51200,0.4642684194776747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,6144,0.06902399990293714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,2560,0.007340444458855524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,1024,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,6144,0.14414666758643255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,3072,0.05717421902550591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,768,0.017861333158281114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,10240,0.01959288948112064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,5120,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,512,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,5120,0.048546665244632296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,128,0.013903111219406128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,51200,0.05429866578843859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,16384,0.024336889386177063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,32,0.10462133089701335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,64,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,128,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,256,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,12288,0.23203823301527235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,2560,0.020632000433074105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,3072,0.01258399999803967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,51200,0.10799555646048652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,1536,0.0035137778355015647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,128,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,768,0.2071911096572876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,8192,0.05578133132722643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,256,0.02568977740075853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,51200,2.8398443857828775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,6144,0.041951109965642296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,6144,0.025014221668243408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,1024,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,1536,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,12288,0.014595554934607612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,8192,0.0340648889541626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,64,0.10307733217875163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,51200,0.07935555775960286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,5120,0.03651377889845107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,7168,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,1536,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,1024,0.008045333127180735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,16384,0.043191111750072904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,1536,0.005423999908897612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,16384,0.14294221666124132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,8192,0.10018044710159302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,2560,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,2048,0.02778488728735182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,8192,0.013550221920013428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,1024,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,51200,0.28803110122680664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,32,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,512,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,16384,0.9192399978637695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,65536,0.1075635552406311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,51200,0.11518755224015977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,4096,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,6144,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,3584,0.016572443975342643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,3072,0.01756977703836229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,3584,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,6144,0.2841920057932536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,3072,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,128,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,3072,0.013025778035322825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,5120,0.0825422207514445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,6144,0.004229333251714706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,128,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,12288,0.11286311679416233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,65536,0.0976675550142924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,128,0.0029333333174387612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,768,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,6144,0.11991022692786323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,512,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,512,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,768,0.011358222199810876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,8192,0.011119999819331698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,768,0.019645333290100098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,5120,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,256,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,1536,0.007582222421964009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,128,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,2048,0.09237333138783772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,256,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,16384,1.5312968360053167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,16384,0.7702319886949328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,5120,0.018957333432303537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,128,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,10240,0.11714666419559056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,2048,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,2048,0.057780444622039795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,64,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,1024,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,2560,0.019076445036464267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,2048,0.02174844510025448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,256,0.025973333252800837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,8192,0.011383111278216044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,256,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,4096,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,3072,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,2048,0.035240888595581055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,12288,0.08390666378868951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,2048,0.022608889473809138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,1024,0.00423200014564726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,16384,0.22570133209228516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,65536,0.6644915474785699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,128,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,12288,0.013804444836245643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,5120,0.010813333094120026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,1024,0.018404444058736164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,2560,0.04382222228580051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,3072,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,51200,0.055416001213921436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,2048,0.007095110913117726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,128,0.005096000101831225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,65536,1.305140495300293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,512,0.003871110992299186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,256,0.012411555482281579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,512,0.012069332930776807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,768,0.010968888799349466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,6144,0.013194666968451606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,128,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,51200,0.6327493455674913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,128,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,128,0.024940444363488093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,6144,0.0990426672829522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,12288,0.015032889114485847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,3072,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,8192,0.03790844480196635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,16384,0.17491732703314888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,51200,0.016557332542207506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,2560,0.09300888909233941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,768,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,8192,0.052108443445629544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,8192,0.10260177983178033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,768,0.011339555184046427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,1024,0.015200000670221118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,3072,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,256,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,3072,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,128,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,256,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,3072,0.03939200109905667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,3072,0.5902061992221409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,51200,0.03013155526585049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,3072,0.03789244426621331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,5120,0.01628533336851332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,5120,0.014463111758232117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,12288,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,128,0.016228444046444364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,2560,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,12288,0.017235555582576327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,768,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,64,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,12288,0.025826666090223525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,3072,0.03195377853181627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,7168,0.9261804156833224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,51200,0.12043378088209365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,8192,0.043395555681652494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,10240,0.012820444173283048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,2048,0.03213511241806878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,256,0.02386133372783661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,6144,0.242287105984158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,12288,0.01622577839427524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,51200,0.05190222130881416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,3072,0.013893332746293811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,256,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,65536,0.13704710536532932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,3584,0.042113777663972646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,6144,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,6144,0.021998221675554912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,32,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,128,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,5120,0.10485777590009902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,8192,0.036526223023732506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,32,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,64,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,256,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,12288,0.03931644558906555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,51200,0.09249777926339044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,6144,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,51200,0.06036000119315254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,16384,0.25396090083652073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,8192,0.030577778816223145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,128,0.02537600033813053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,65536,0.7821413146124946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,128,0.0018648888087934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,1024,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,768,0.03664977682961358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,256,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,5120,0.05115288827154371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,5120,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,5120,0.028734223710166082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,6144,0.030917333232031927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,2560,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,16384,0.02093244426780277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,51200,0.4560782114664714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,51200,0.1162266731262207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,7168,0.013854222165213691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,5120,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,65536,0.7709964646233453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,768,0.02131822208563487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,65536,7.736905415852864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,8192,0.03681955403751797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,2048,0.009489778015348647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,16384,0.08920711278915405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,128,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,4096,0.018609777092933655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,8192,0.045560889773898654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,1536,0.020836444364653692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,2048,0.021570665968788996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,3072,0.01348000019788742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,6144,0.050042665667004056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,768,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,51200,0.06357510884602864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,768,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,65536,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,5120,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,1536,0.02639466689692603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,1536,0.13960177368587917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,64,0.0018275556051068837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,512,0.30947555436028373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,2560,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,2560,0.02239466706911723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,10240,0.020961778031455144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,12288,0.9220515357123481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,7168,0.17298400402069092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,5120,0.1481706698735555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,6144,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,10240,0.291869322458903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,512,0.039121776819229126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,5120,0.012800888882742988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,10240,0.09828355577256943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,10240,0.09570844305886163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,16384,0.11635200182596843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,12288,0.12078044149610732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,1536,0.02106222179200914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,8192,0.004221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,1536,0.05033422178692288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,65536,0.0643813345167372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,16384,1.17866759830051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,6144,0.18486666679382324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,2560,0.06661778026156955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,10240,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,10240,0.028763555818133887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,5120,0.07270222240024142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,51200,0.1542488866382175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,12288,0.16832533147599962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,7168,0.09502577781677246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,3584,0.0958719982041253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,512,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,64,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,6144,0.030055999755859375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,7168,0.021167111065652635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,256,0.018239999810854595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,10240,0.03239555491341485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,256,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,3584,0.026909333136346605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,1536,0.017249777913093567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,7168,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,3584,0.3266506724887424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,2560,0.022486221459176805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,512,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,128,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,4096,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,768,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,4096,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,4096,0.023984000086784363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,10240,0.5941928757561578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,51200,0.04621866676542494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,3584,0.5751591258578831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,5120,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,256,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,256,0.026037333740128413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,10240,0.08745777606964111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,51200,0.6651608679029677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,512,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,7168,1.515875604417589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,7168,0.03252622154023912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,2048,0.014918223023414612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,10240,0.02615822189384037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,12288,0.007920000288221572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,2048,0.04875377813975016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,2048,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,5120,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,5120,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,4096,0.17585333188374838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,6144,0.8723200162251791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,16384,0.2498373455471463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,3584,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,512,0.00300177786913183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,512,0.017737777696715463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,3072,0.05198311143451267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,512,0.015788444214397006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,2048,0.39481510056389707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,16384,0.07891288730833265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,12288,0.022950222094853718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,65536,0.3132337729136149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,32,0.010626667075686984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,2560,0.010778666370444827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,2560,0.019161777363883126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,768,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,2048,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,768,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,256,0.00227111101978355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,1024,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,256,0.03679644399219089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,512,0.025729777084456548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,32,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,12288,0.05724622143639458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,256,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,3584,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,6144,0.05978577666812473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,4096,0.026573333475324843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,32,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,2560,0.08672622177335952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,2048,0.13655466503567165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,1024,0.012114666402339935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,256,0.009045333498054082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,4096,0.017257778180970084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,16384,0.12583555115593806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,128,0.030840890275107488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,12288,0.030971556901931763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,10240,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,1536,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,3072,2.382577684190538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,1024,0.03278755479388767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,1024,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,16384,0.2385662131839328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,12288,0.03695288962788052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,5120,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,512,0.06641244226031832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,4096,10.530175103081598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,5120,0.029595557186338637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,7168,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,5120,0.06042489078309801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,512,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,7168,0.13212266233232287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,65536,2.985333336724175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,3072,0.3675333393944635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,1024,0.017165333032608032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,3584,0.039846221605936684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,51200,0.17112444506751168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,128,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,512,0.18568800555335152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,12288,0.17168800036112467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,5120,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,1024,0.015049777097172208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,5120,0.02472355630662706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,128,0.002453333387772242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,512,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,51200,0.7656746440463595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,1024,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,7168,0.05585600270165337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,1536,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,6144,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,10240,0.1723173326916165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,1536,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,64,0.004882666799757215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,65536,0.08310133218765259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,768,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,2048,0.03886577818128798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,16384,0.10674844185511272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,51200,0.4484497706095378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,6144,0.02459200057718489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,768,0.07225066423416138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,3072,0.060510224766201444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,768,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,6144,0.5387271245320638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,16384,0.11169066694047715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,16384,0.14827733569675022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,2048,0.07969511217541166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,1536,0.005817777580685086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,512,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,12288,0.03083733386463589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,2048,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,5120,0.04446666770511203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,7168,0.06673689021004571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,51200,0.04026133484310574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,256,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,3584,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,3584,0.014684443672498068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,3072,0.015057777365048727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,10240,0.17328888840145537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,32,0.014169777433077494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,7168,0.031192888816197712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,1024,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,8192,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,512,0.018199111024538677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,3584,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,12288,1.985330581665039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,6144,0.024262222978803847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,1024,0.08929600318272908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,8192,1.6891821755303278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,16384,0.03349155518743727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,64,0.007828444242477417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,3072,0.028774221738179524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,2048,0.028823110792371962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,7168,0.004536889079544279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,6144,0.039946665366490684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,7168,0.3056097825368245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,128,0.23845420943366158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,6144,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,2048,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,16384,0.05979910824033949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,6144,0.009475555684831407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,3584,0.049680888652801514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,7168,0.013105777402718862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,256,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,1536,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,16384,1.2673217985365126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,256,0.021253334151373968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,32,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,512,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,3072,0.02164711058139801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,4096,0.022701332966486614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,256,0.001981333312061098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,7168,0.053800000084771045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,10240,0.06018400192260742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,6144,0.024852444728215534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,16384,0.21322666274176705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,1536,0.0480915539794498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,256,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,512,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,1536,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,16384,0.02233155568440755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,6144,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,5120,0.03462666604253981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,12288,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,5120,0.062275555398729115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,1536,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,128,0.0018977777411540349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,65536,16.661449856228298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,4096,0.03198400139808655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,512,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,2048,0.00813333359029558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,16384,0.09707111120223999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,2560,0.2799297703637017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,7168,0.04371822211477491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,6144,0.045379555887646146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,1536,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,1536,0.019754666421148512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,128,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,16384,0.14453332953982884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,6144,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,1536,0.020358221398459542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,5120,0.2456888887617323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,2560,0.12637955612606472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,128,0.02178222272131178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,32,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,4096,0.039065778255462646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,2048,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,12288,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,8192,0.8267475763956705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,3584,0.028672890530692205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,1024,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,16384,0.0662248863114251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,3584,0.0849999984105428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,3584,0.047263112333085805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,256,0.016034666034910414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,3584,0.2920871045854357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,6144,0.018623999423450895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,256,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,51200,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,256,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,32,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,3584,0.023899555206298828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,10240,0.016780444317393832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,768,0.22940354877048066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,768,0.012753778033786349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,1024,0.09011644124984741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,4096,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,7168,0.010828444527255164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,1536,0.015544889701737298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,10240,0.16850044992234972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,32,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,64,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,128,0.005088888936572605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,1024,0.28654665417141384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,65536,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,12288,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,1536,0.014635556273990206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,5120,0.036700444089041814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,2560,0.03250755535231696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,65536,0.0783404442999098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,1536,0.025847999585999384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,6144,0.11264888445536296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,12288,0.04231466518508064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,12288,0.021190222766664293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,768,0.007751999629868402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,1536,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,8192,0.07343022028605144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,2048,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,8192,0.10263999965455796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,1024,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,51200,0.2945342328813341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,8192,0.024703111913469102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,1024,0.006395555618736479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,51200,0.4287555482652452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,1536,0.02402399977048238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,10240,0.03912533322970072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,64,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,512,0.006708444406588872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,51200,0.12866132789187962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,256,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,16384,0.04411910971005758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,8192,4.734246148003472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,3072,1.3004195955064561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,65536,17.794331868489582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,16384,0.3086186779869927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,3072,0.015226667126019796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,512,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,128,0.010653333531485664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,4096,0.045448889334996544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,2560,0.017242666747834947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,128,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,1024,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,3584,0.04038222299681769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,2560,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,3584,0.025480000509156123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,2048,0.01125955581665039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,32,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,3584,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,65536,2.1726417541503906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,5120,0.9122213787502713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,4096,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,10240,0.012187555432319641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,6144,0.5091013378567165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,10240,0.04002933369742499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,2560,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,12288,0.06646044386757745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,7168,0.0210533340771993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,1024,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,3584,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,256,0.0074515557951397365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,51200,0.24710223409864637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,6144,0.267798211839464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,1024,0.019279999865425956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,3072,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,10240,0.10722311337788899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,768,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,12288,0.027276444766256545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,128,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,16384,1.244566175672743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,12288,0.03626666797531976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,7168,0.5724097887674967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,3072,0.03953155544069078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,51200,0.059559113449520536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,2560,0.014753777119848462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,51200,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,32,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,6144,0.013655111193656921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,3584,0.04570400052600437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,8192,0.2526346577538384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,2048,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,2048,0.005382222019963794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,5120,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,128,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,128,0.015571556157535978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,4096,0.01312355531586541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,8192,0.11496799521976048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,1536,0.010266666611035665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,768,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,6144,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,64,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,3584,0.027278222971492346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,2560,0.015879111157523263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,3072,0.034727109803093806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,16384,0.0677644411722819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,64,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,12288,0.04056977894571092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,768,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,16384,0.04814666509628296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,7168,0.10584621959262425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,2048,0.012688889271683164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,10240,0.02608355548646715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,256,0.016442666451136272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,64,0.00675644435816341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,1536,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,2048,0.025864000121752422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,6144,0.032510221004486084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,4096,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,8192,0.04229955540763008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,65536,4.419090694851346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,8192,2.9838951958550344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,16384,0.10574755403730605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,1536,0.005695111221737332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,8192,0.025915554828113977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,12288,0.029820445511076186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,51200,4.261297861735026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,512,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,8192,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,51200,0.06171822547912598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,10240,3.3084328969319663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,768,0.01791288952032725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,256,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,10240,0.05441155698564318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,6144,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,12288,0.5706799825032552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,64,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,12288,0.04049955474005805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,5120,0.011985777980751462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,5120,0.007877333296669854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,6144,0.01928444372283088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,51200,0.10379733641942342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,8192,0.00906311141120063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,1536,0.1346613301171197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,65536,0.07441422012117174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,3072,0.021523555119832356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,8192,0.026339555780092876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,7168,0.1101448867056105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,512,0.007690666450394525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,51200,0.5579671329922146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,65536,0.3531155586242676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,7168,0.011204444699817233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,4096,0.060481780105166964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,512,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,6144,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,65536,0.5490951008266872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,128,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,2048,0.007118222614129384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,51200,0.5902249018351237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,512,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,8192,0.0073315559162033936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,32,0.020918221937285528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,51200,0.05725955300860935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,3584,0.6279662450154623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,3584,0.04641155401865641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,5120,1.1083520253499348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,256,0.11115199989742702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,65536,0.1436328887939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,1536,0.014908444550302295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,3072,0.015692444311247934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,512,0.013080889152155982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,65536,4.616959889729817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,10240,0.3386133246951633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,2560,0.08966666460037231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,4096,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,65536,0.26728532049391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,3072,0.007496889266702864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,32,0.003567999849716822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,64,0.001790222194459703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,2560,0.033225778076383806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,10240,0.03427377674314711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,256,0.0024782222592168385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,7168,0.017631111873520743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,65536,0.38065600395202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,3072,0.011640888949235281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,32,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,32,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,1024,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,7168,0.010616888602574667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,16384,0.13233866956498888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,1536,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,256,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,2048,0.07789688640170626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,256,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,2560,0.020696888367335003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,10240,0.012714666624863943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,3584,0.6304862234327528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,4096,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,10240,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,10240,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,512,0.012807110945383707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,2048,0.15659911102718777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,32,0.027221333649423387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,256,0.018926221463415358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,4096,0.037011557155185275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,256,0.01254666679435306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,51200,0.24550223350524902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,2048,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,65536,0.07467555337482028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,2048,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,7168,0.05559200048446655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,2048,0.01994044416480594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,7168,0.9406986236572266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,2560,0.024142222272025213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,4096,0.030751999881532457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,768,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,512,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,10240,0.3937404420640733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,2048,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,51200,0.4398053222232395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,51200,0.05098311106363932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,3072,0.00811111099190182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,256,0.014735110931926303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,2560,0.2877724435594347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,1024,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,64,0.0052897776994440295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,65536,0.1790693336062961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,1024,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,512,0.014871999621391296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,16384,0.20429066816965738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,10240,0.11141422059800889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,3072,0.0317795541551378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,7168,0.010088889135254754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,12288,0.009402666654851701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,256,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,2560,0.06443911128573947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,7168,0.05805155303743151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,3072,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,8192,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,64,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,12288,0.069069332546658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,16384,0.26716799206203884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,65536,0.08201333549287584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,32,0.0019360000474585427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,10240,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,4096,0.008461332983440822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,256,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,8192,0.11000088850657146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,2048,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,5120,0.10994488663143581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,12288,0.08649688959121704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,3072,0.013379555609491138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,1024,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,8192,0.2595626778072781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,7168,0.09580355882644653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,5120,0.017858665850427415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,512,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,16384,0.039093332158194646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,256,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,4096,0.10677778058581883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,128,0.021016889148288306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,10240,0.04266755448447334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,32,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,1024,0.017257778180970084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,12288,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,3584,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,6144,0.018152889278199937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,5120,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,512,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,1536,0.019121777680185106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,1024,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,4096,0.04047022263209025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,16384,0.04471022221777174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,7168,0.017148445049921673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,10240,0.015020444989204407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,256,0.15554489029778376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,256,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,128,0.12128622002071804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,2048,0.014208000567224292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,1024,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,2560,0.007575111256705389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,3072,0.006776888751321369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,3072,0.0787386695543925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,2048,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,12288,0.013742222554153867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,12288,0.06643378072314791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,51200,0.12993866867489284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,12288,0.03744533326890733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,65536,0.40923555692036945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,3072,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,512,0.012725333372751871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,7168,0.015765332513385348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,1024,0.012449777788586087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,8192,0.023783110909991797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,128,0.014532445205582513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,7168,0.032474666833877563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,256,0.01201333353916804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,1024,0.026775999201668635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,2048,0.009974222216341231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,5120,0.012703110774358114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,4096,0.11925600634680854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,16384,0.1128702163696289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,10240,0.0697644419140286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,6144,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,3072,0.021935999393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,5120,0.0380586650636461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,8192,0.07595822546217176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,8192,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,1024,0.025418667329682246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,2560,0.11091111103693645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,6144,1.5452097786797419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,10240,0.028070221344629925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,1024,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,4096,0.051510221428341336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,4096,0.025415110919210646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,10240,0.03871466716130575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,51200,0.10711110962761773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,65536,0.11320000224643284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,3584,0.024193776978386775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,65536,0.3659075631035699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,1536,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,512,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,16384,0.24258134100172254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,3072,0.032060444355010986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,768,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,16384,0.36100443204243976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,1024,0.004502222355869082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,12288,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,1024,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,256,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,2560,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,51200,6.053867763943142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,128,0.003855111284388436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,4096,0.35170933935377335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,10240,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,256,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,6144,0.10870666636361016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,8192,0.05584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,2048,0.00453955555955569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,256,0.028112000889248315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,768,0.0177839994430542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,16384,0.016377778516875375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,256,0.020226667324701946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,3072,0.06872177786297269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,768,0.04506133331192864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,2560,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,8192,0.013971555564138623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,768,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,6144,0.029380444023344252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,51200,1.5978897942437067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,8192,0.04730222291416592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,12288,0.03513333201408386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,65536,0.2781386640336779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,65536,3.1754497951931424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,2048,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,5120,0.025289777252409194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,4096,0.013906665974193148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,512,0.013295110729005603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,8192,0.011557333171367645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,8192,0.057578669653998486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,1024,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,2048,0.06668888860278659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,6144,0.013678222066826291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,1024,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,512,0.06784533129798041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,4096,0.04619822237226698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,512,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,65536,0.0860826638009813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,512,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,1536,0.02903911140229967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,3584,0.10763021972444321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,3072,0.014599111345079211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,1536,0.03824977742301093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,512,0.010061333576838175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,10240,0.0762595534324646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,12288,0.3570844332377116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,4096,0.20279288291931152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,256,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,2048,0.03852355480194092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,65536,0.06191022528542412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,512,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,3584,0.06631555822160509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,8192,0.4640524652269151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,2560,0.013541333377361298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,4096,1.1591200298733182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,7168,0.03115733464558919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,256,0.0064693333374129395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,5120,0.008598222500748104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,7168,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,51200,0.17029333114624023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,6144,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,768,0.2509546544816759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,256,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,1024,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,5120,0.7536480161878797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,1024,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,128,0.009717333647939894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,5120,0.03436622354719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,65536,0.0681555536058214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,5120,0.022967111733224656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,1024,0.02502311103873783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,5120,0.015644444359673392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,128,0.014831110835075378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,7168,0.06189599964353773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,6144,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,12288,0.4139395554860433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,3584,0.03756444321738349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,7168,0.881422201792399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,5120,0.11475466357337104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,51200,0.2000968853632609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,7168,0.44350398911370176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,7168,0.029574222034878198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,5120,0.06341955396864149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,1024,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,5120,0.06336177719963922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,51200,0.10620177454418606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,1536,0.018236445056067575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,16384,0.590628465016683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,6144,0.013612444202105204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,7168,0.02085866696304745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,16384,0.09416177537706162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,16384,0.26626043849521214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,4096,0.01498133275243971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,512,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,51200,0.17855377991994223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,1536,0.004084444294373195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,256,0.04976177877850003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,2560,0.033222221665912204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,2560,0.05609066618813408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,256,0.018632888793945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,12288,0.17501866817474365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,10240,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,512,0.002833777831660377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,16384,0.15234577655792236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,10240,0.031122667921913996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,3072,0.09253155522876316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,5120,0.04923911227120293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,1536,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,256,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,5120,0.44136444727579754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,10240,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,768,0.04168711105982462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,3584,0.12031643920474583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,8192,0.14553244908650717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,768,0.008104888929261101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,32,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,6144,0.009617778162161509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,51200,0.049921777513292104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,7168,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,4096,0.030140442980660334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,3584,0.04445688923199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,2560,0.00720355576939053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,3072,0.007910221815109253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,6144,0.022207111120224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,512,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,10240,0.01831466621822781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,51200,0.05117688907517327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,512,0.06434577703475952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,768,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,1536,0.055983112918006055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,768,0.10283022456698948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,768,0.04000799854596456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,4096,0.03963111175431146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,7168,0.2267084386613634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,16384,8.035487704806858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,1024,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,512,0.020928889513015747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,5120,0.018753778603341844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,1024,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,16384,0.022642667094866436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,16384,0.17415644062889946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,2048,0.015842666228612263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,16384,0.10740711291631062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,16384,0.16170488463507757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,768,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,512,0.01629866659641266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,5120,0.020280889338917203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,32,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,2048,0.03457422057787577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,64,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,3584,0.08527200089560615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,6144,0.013535111314720578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,10240,0.045626666810777455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,8192,0.025006221400366888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,3584,0.007689778175618913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,5120,0.009033777647548253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,2560,0.07366755273607042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,51200,0.048885332213507764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,256,0.00222044438123703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,64,0.017575111654069688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,1024,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,256,0.016840888394249808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,51200,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,10240,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,16384,0.5177609125773112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,256,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,32,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,51200,0.06320622232225206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,64,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,2560,0.004883555488453971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,7168,4.327080196804471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,8192,0.014537778165605334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,1024,0.01500266624821557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,10240,0.03466755482885573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,51200,0.19442310598161486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,65536,0.26499199867248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,128,0.021199999584092036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,51200,0.1635217799080743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,5120,0.026723555392689172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,3072,0.1027253336376614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,3072,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,256,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,128,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,3072,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,1536,0.021407110823525324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,512,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,3072,0.017869333426157635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,512,0.07783910963270399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,1024,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,128,0.0053226666318045715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,16384,0.03457155492570665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,10240,0.02056711084312863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,128,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,5120,0.04788177874353197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,16384,0.1301342248916626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,1024,0.027806222438812256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,1536,0.009574222067991892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,51200,0.11601866616143121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,2048,0.006895111252864202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,4096,0.007845333053006066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,10240,0.05442133214738634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,2560,0.03889866669972738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,768,0.019695111446910434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,51200,1.8600347306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,128,0.012334222594896952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,10240,0.03234311276011997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,2048,0.06548177533679538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,7168,0.011298666397730509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,768,0.003485333174467087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,16384,0.04634488953484429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,65536,0.1806115574306912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,3584,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,6144,0.014008000493049622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,64,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,3584,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,64,0.06981422503789265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,8192,0.02732444471783108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,256,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,12288,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,5120,0.02938400043381585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,1536,0.23948089281717935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,51200,0.8590355449252658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,2048,0.03291733397377862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,3072,0.08322755495707194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,5120,0.5555013550652398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,1536,0.0389066669676039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,1024,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,51200,0.022729777627521094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,16384,19.71526421440972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,1024,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,768,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,10240,0.027816888358857896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,16384,0.11035644345813328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,2048,0.014936889211336771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,2048,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,3072,0.05362666646639506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,51200,0.05896800094180637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,768,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,256,0.006681777950790193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,65536,0.06570755773120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,2048,0.03102933367093404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,2560,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,1536,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,7168,0.009418666362762451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,32,0.03779733180999756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,128,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,4096,0.06890222099092272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,16384,0.13976266649034289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,51200,0.10966133409076267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,51200,0.038198222716649376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,8192,0.010840000377760993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,3072,0.2675466537475586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,51200,0.062462223900689014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,2560,0.0418888893392351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,256,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,32,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,5120,0.06526222493913439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,512,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,7168,0.13249066140916613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,8192,0.011034666664070554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,64,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,2048,0.07224622037675646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,128,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,3584,0.011371555427710215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,6144,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,5120,0.03493244449297587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,8192,0.0240320000383589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,256,0.002404444540540377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,2048,0.019339554839664035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,16384,0.7876275380452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,8192,0.022628444764349196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,128,0.008489777644475302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,65536,0.021151999632517498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,256,0.020119110743204754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,3584,0.03831555445988973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,512,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,10240,0.16437955697377524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,8192,0.11185422208574082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,64,0.0025048889219760895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,1024,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,6144,0.03461422191725837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,768,0.025263999899228413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,5120,0.3744853337605794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,1536,0.2507573233710395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,65536,0.20635110802120635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,8192,0.015079999963442484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,128,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,512,0.012124444047609964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,1536,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,256,0.010000000397364298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,1024,0.007084444165229797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,8192,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,2560,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,3584,0.03308622373474969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,128,0.005393777870469623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,7168,0.04531822270817227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,768,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,768,0.003031111011902491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,2560,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,51200,0.04243377844492594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,8192,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,8192,0.04474488894144694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,1024,0.007044444481531779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,12288,0.019479110836982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,2560,0.010072888599501716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,5120,0.03482666611671448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,7168,0.12441244390275742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,1536,0.012519111235936483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,1536,0.008032889001899296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,8192,2.053959952460395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,768,0.022058667408095464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,1024,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,6144,0.02680533296532101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,12288,1.0905413097805448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,128,0.020270221763186984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,1024,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,2560,0.07105422019958496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,3584,0.008788444101810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,51200,0.05306577682495117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,3072,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,256,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,5120,0.12237955464257134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,65536,0.5610942310757107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,65536,9.01405504014757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,16384,0.02232088810867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,128,0.005102222164471944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,12288,0.5983813073900011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,12288,0.01830755505296919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,16384,0.09620622131559585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,128,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,51200,0.5098880132039388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,512,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,512,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,8192,0.025011556016074285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,4096,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,51200,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,768,0.04037244452370538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,7168,0.07047377692328559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,12288,0.01942133406798045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,12288,0.2905093299018012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,128,0.006190222170617845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,2560,0.04616444309552511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,64,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,51200,0.14573689301808676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,51200,0.10452977816263835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,8192,3.295891655815972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,2048,0.02336888843112522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,768,0.02500533395343357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,768,0.02977422210905287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,256,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,2048,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,5120,0.010520888699425591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,5120,0.025234666135576036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,7168,0.2564977804819743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,3072,0.09240266349580552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,10240,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,10240,0.16678222020467123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,10240,0.055359999338785805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,4096,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,2048,0.036172442966037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,768,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,128,0.015832000308566622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,256,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,51200,0.07796177599165174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,256,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,65536,0.44213154580858016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,65536,0.07095111078686185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,2048,0.00489688871635331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,7168,0.010974222587214576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,10240,0.031151112582948472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,3072,0.023642667465739783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,16384,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,2560,0.00904799997806549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,2048,0.014488000008794995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,1536,0.011125333607196808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,7168,0.008742222355471717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,65536,0.21652000480227998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,256,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,8192,0.06338489055633545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,2560,0.0990897748205397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,12288,0.02976977825164795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,768,0.012404444317022959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,6144,0.09954933325449626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,128,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,10240,0.043748444981045194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,10240,0.05292000042067634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,64,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,1024,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,8192,0.030008001459969416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,1536,0.020759999752044678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,6144,0.19499377409617105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,51200,5.949735005696614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,128,0.022311111291249592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,3072,0.44515111711290145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,4096,0.06988444593217638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,2560,0.014615999327765571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,5120,0.009389333426952362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,1024,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,512,0.008801778157552084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,12288,0.008147555920812819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,4096,0.04503555430306328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,64,0.0037208890749348533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,128,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,2048,0.014092443717850579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,16384,0.5425413449605306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,51200,0.6416684256659614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,12288,0.049069331751929395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,1536,0.006538666784763336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,3584,0.013023110727469126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,1024,0.028759111960728962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,3584,0.04013333386845059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,8192,0.4389759964413113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,51200,0.3427662319607205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,7168,0.05186222328080071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,256,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,8192,0.014027555783589682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,16384,0.02381155557102627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,512,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,128,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,16384,0.03264710969395108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,7168,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,512,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,768,0.006046222315894232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,512,0.011782222323947482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,51200,0.05166755451096428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,7168,0.01666933298110962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,51200,0.2880000008477105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,1024,0.020527111159430612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,3072,0.03644088904062907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,4096,0.03316266669167413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,7168,0.07041866911782159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,5120,0.07303288910124037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,65536,0.06142933501137627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,1024,0.025571554899215698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,3584,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,3072,0.03760177890459696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,7168,0.023295111126369897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,16384,0.08981421920988296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,65536,0.21576800611284044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,51200,3.8972295125325522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,2048,0.115192002720303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,10240,2.8466478983561196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,2560,0.08645688825183445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,256,0.01132266637351778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,5120,0.05377066797680325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,65536,0.008353778057628209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,7168,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,65536,0.05922222137451172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,8192,0.02015555567211575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,128,0.007672888537247975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,51200,1.9169510735405817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,32,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,4096,0.01462399959564209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,2560,0.013518222504191928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,51200,0.2775164445241292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,768,0.04851288927925957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,5120,0.021818665994538203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,64,0.020276443825827706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,768,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,512,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,512,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,1024,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,3072,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,768,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,512,0.011362666885058085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,3072,0.1543617778354221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,3072,0.09448444181018406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,5120,0.03373333480623033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,10240,0.09130933549669053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,3072,0.013550221920013428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,3072,0.012599111431174807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,16384,0.0158906661801868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,3584,0.07578311363855998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,2560,0.007814221911960179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,256,0.00244355553554164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,4096,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,1536,0.05616088708241781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,16384,0.012352888782819113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,2048,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,1024,0.005647111270162795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,256,0.0064186664919058485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,2560,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,7168,0.02757511039574941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,51200,0.059133330980936684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,65536,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,3072,0.5484302308824327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,1024,0.07252799802356295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,51200,0.05255200134383308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,1024,0.01924711134698656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,6144,0.18713688850402832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,64,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,7168,0.044886221488316856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,12288,0.22367821799384224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,3584,0.015106666419241162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,1024,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,7168,0.08145155509312947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,51200,0.05635733074612088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,128,0.02016888890001509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,2560,0.014059555199411182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,4096,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,8192,0.03521777855025397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,128,0.00198399999903308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,3584,0.055833776791890465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,64,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,12288,0.3052204449971517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,4096,0.04790133237838745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,4096,0.01757333344883389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,512,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,8192,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,1024,0.022410665949185688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,2048,0.004600888739029567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,2048,0.052694221337636314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,65536,0.41667289204067653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,3584,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,65536,0.0686195558971829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,1536,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,1536,0.00942133367061615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,65536,0.06125689215130276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,51200,0.05458044343524509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,3584,0.018696000178654987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,2560,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,512,0.0052879999081293745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,32,0.002208000048995018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,256,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,65536,0.6214950879414877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,5120,0.03823644585079617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,5120,3.027425765991211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,32,0.03762222329775492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,512,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,5120,0.029496000872717962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,16384,0.017793777916166518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,10240,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,16384,0.12796266873677573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,6144,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,16384,1.0057546827528212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,5120,0.09575555721918742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,1024,0.0791520012749566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,65536,0.055827558040618896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,3584,0.011984888878133563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,12288,0.01464355488618215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,10240,0.01739644507567088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,12288,0.05024888780381945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,64,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,10240,1.231962627834744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,3072,0.01607733302646213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,7168,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,2560,0.006698666761318843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,64,0.004048888882001241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,2560,0.028982222080230713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,3072,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,3072,0.015990222493807476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,3584,0.010456889039940305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,4096,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,7168,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,768,0.01773422294192844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,16384,0.07053777906629774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,4096,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,2560,0.06431733237372504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,1536,0.0035537779331207275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,65536,0.210826661851671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,12288,0.04805422160360548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,128,0.011425777441925473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,5120,0.10228088829252456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,3584,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,512,0.005657777604129579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,51200,0.038057777616712786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,8192,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,32,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,2560,0.030187557140986126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,65536,2.1074896918402777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,256,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,12288,0.04009777638647292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,3584,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,1536,0.031215111414591473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,2048,0.017264000243610807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,10240,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,3584,0.01407199932469262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,4096,0.023256889647907678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,2048,0.019918221566412184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,1536,0.018995554910765756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,1536,0.025360888904995386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,1024,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,10240,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,64,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,7168,0.04794399936993917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,4096,0.018724444839689467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,5120,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,768,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,32,0.009130666653315226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,8192,0.028753777345021565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,3072,0.0240000006225374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,768,0.003830222205983268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,256,0.02242311173015171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,768,0.008405333591832055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,3584,0.15370933214823404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,65536,0.6462817721896702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,5120,0.022058667408095464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,3072,0.2706071005927192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,512,0.015268445014953613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,64,0.011682666838169098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,6144,0.024375110864639282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,128,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,5120,0.017176000608338248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,7168,0.03932799895604452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,2048,0.011579555769761404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,1536,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,10240,0.019508444600635104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,32,0.01686844395266639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,8192,8.037530687120226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,8192,0.014890667464998035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,5120,0.06109244293636746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,16384,2.4985858069525824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,5120,0.016744888491100736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,3072,0.023380445109473333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,64,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,65536,14.12432861328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,10240,0.20983733071221244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,1024,0.02029600077205234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,1024,0.007678222325113084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,4096,0.058324442969428174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,768,0.016892444756295946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,64,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,51200,0.07451022333568998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,2560,0.017875555488798354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,1536,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,128,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,8192,0.029191109869215224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,7168,0.014501333236694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,512,0.0038746665749284956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,64,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,12288,0.030870222383075293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,512,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,4096,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,6144,0.013798221945762634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,1536,0.0035439998739295532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,256,0.006173333360089197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,768,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,8192,0.04328711165322197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,65536,0.08886044555240208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,51200,0.27528354856703013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,7168,0.02126577827665541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,7168,0.031424889961878456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,51200,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,5120,0.021547555923461914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,768,0.007784888976150089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,3584,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,32,0.0032115555885765287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,5120,0.024029332730505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,16384,4.163063049316406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,8192,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,3072,0.012448000411192576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,16384,0.02083911167250739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,4096,0.01235466698805491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,3072,0.019713777634832595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,5120,0.00452533322903845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,12288,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,1024,0.03903466794225905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,65536,1.0210391150580513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,7168,0.5588897599114312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,512,0.036547554863823786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,128,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,256,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,12288,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,3584,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,1024,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,7168,0.04082577758365207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,8192,0.015815999772813585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,5120,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,51200,0.22955645455254448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,128,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,65536,0.06844000021616618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,2048,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,768,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,3584,0.033486220571729876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,1536,3.5407386355929904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,5120,0.020781333247820537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,6144,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,768,0.027776888675159876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,8192,0.02349600030316247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,8192,0.19927467240227592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,64,0.02163999941613939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,2560,0.07853333155314128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,768,0.05731111102634006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,128,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,12288,0.16116710503896078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,1536,0.018072000808186002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,51200,0.0540284448199802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,2560,0.02241777711444431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,128,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,1024,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,51200,0.07873688803778754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,128,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,128,0.014868444866604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,10240,0.010653333531485664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,128,0.0021928888228204516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,10240,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,5120,0.021020443903075323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,2560,0.19302311208513048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,2560,0.023405333360036213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,256,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,10240,0.02110933264096578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,12288,0.013924444715181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,32,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,10240,0.026076444321208533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,6144,0.04845866560935974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,128,0.014455111490355598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,10240,0.014354666074117025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,7168,0.025788444611761305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,16384,0.4524933497111003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,7168,0.03433777888615926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,32,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,16384,0.08905866410997178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,3584,0.04029777646064758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,128,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,4096,0.016420443852742512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,8192,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,768,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,2560,0.0074933336840735535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,16384,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,8192,0.025965332984924316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,3072,0.0075733330514695905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,2048,0.039635555611716375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,2048,0.49535110261705184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,4096,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,4096,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,5120,0.030363556411531236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,512,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,256,0.031143112315071955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,1536,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,128,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,512,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,768,0.03323200013902452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,3072,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,3072,0.03229511115286086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,256,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,16384,0.1343582206302219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,32,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,16384,0.0782160030470954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,512,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,128,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,65536,0.09977422157923381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,51200,0.0380257765452067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,10240,0.05541333225038317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,5120,0.05901955233679878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,16384,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,10240,0.1643431054221259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,10240,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,65536,0.06433155801561144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,8192,0.014477334088749357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,3072,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,4096,0.06589777602089776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,768,0.03491111265288459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,6144,0.020375111036830477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,2048,0.011447110937701331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,7168,0.07729778024885389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,768,0.015386665860811869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,16384,0.13819555441538492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,128,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,2560,0.6003697713216146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,6144,0.01126399967405531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,7168,0.04016977879736159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,512,0.014841778410805596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,16384,0.0968853301472134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,256,0.022152000003390845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,2560,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,512,0.002607999990383784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,256,0.002263111165828175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,6144,0.7103466457790799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,16384,1.7213315963745117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,2560,0.2070773310131497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,65536,0.1424328883488973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,65536,0.16127555900149876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,1536,0.21871733665466309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,2048,0.07715821928448148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,3072,0.03050666716363695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,512,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,3072,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,768,0.030898667044109766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,7168,0.05141333407825894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,64,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,1536,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,16384,0.022224000758594934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,65536,0.14078577359517416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,65536,0.1979688935809665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,2048,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,1024,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,32,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,8192,0.02179644505182902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,4096,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,65536,0.08666221963034736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,10240,0.3781235482957628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,1536,2.7865342034233938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,10240,0.02311822275320689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,512,0.023333332604832117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,65536,0.10860444439782037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,1536,0.026094221406512793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,8192,0.15899644957648382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,51200,1.26839107937283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,768,0.06318044662475586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,16384,0.0462408893638187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,2560,0.03725688987308078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,4096,0.008159110943476358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,5120,0.02576444380813175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,8192,0.18694844510820177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,256,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,5120,0.04205333193143209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,5120,0.017415111263593037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,2560,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,1536,0.01887999971707662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,6144,0.09036978085835774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,512,0.051741331815719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,16384,0.7823537720574273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,1024,0.06534044610129462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,5120,0.01608000033431583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,2048,0.007511110769377814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,3072,0.1328631109661526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,5120,0.016809778081046212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,768,0.14374132951100668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,512,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,16384,0.26829777823554146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,6144,0.029521776570214167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,3584,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,768,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,256,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,512,0.005688888745175467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,512,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,2048,0.008035555481910706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,1536,0.8723822169833713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,65536,0.38731911447313094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,12288,0.262572447458903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,7168,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,256,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,256,0.015136000182893542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,64,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,12288,0.02997777859369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,1024,0.02270577847957611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,51200,0.07627911037868924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,2560,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,3072,0.02344888945420583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,6144,0.013387555877367655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,256,0.0034871109657817413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,3584,0.007800888684060838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,65536,0.8180319997999402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,65536,0.029010666741265193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,5120,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,64,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,65536,0.11370755566491021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,8192,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,8192,0.05881777736875746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,64,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,256,0.0021173333128293357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,12288,0.025942222939597234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,64,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,2048,0.004515555583768421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,128,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,16384,0.2318293253580729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,3072,0.0255422227912479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,10240,0.030978666411505804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,16384,0.046471112304263644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,2048,0.004308444344335132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,2048,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,2560,0.030419554975297716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,8192,0.025651555922296312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,768,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,16384,0.01647733317481147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,6144,0.06168533033794827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,5120,0.012463111016485425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,4096,0.02077866593996684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,2048,0.010758221977286868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,5120,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,32,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,51200,0.23496445020039877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,1024,0.9168649249606662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,8192,0.04528888728883532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,4096,0.010695999695195092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,3584,0.6222186618381077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,1536,0.024821332759327356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,8192,0.022359111242824133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,16384,0.009269333548016017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,65536,0.07637066973580255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,256,0.012358222570684222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,4096,0.031947556469175555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,51200,0.0769795576731364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,512,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,4096,0.009646221995353699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,6144,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,128,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,256,0.061590221193101674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,1024,0.02267733381854163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,65536,4.575439876980252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,65536,0.2428800000084771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,1024,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,256,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,3584,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,2048,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,1024,0.017895110779338412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,512,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,10240,0.17676266034444174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,4096,0.01755644381046295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,12288,0.018897778458065458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,16384,0.7808266745673286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,1536,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,3072,0.008061333662933773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,3584,0.009679111341635386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,16384,0.05311110946867201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,7168,0.021759111020300124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,51200,0.06180622180302938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,6144,0.04319999946488274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,256,0.01718488832314809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,1536,0.07391644186443753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,256,0.01699911057949066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,32,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,65536,0.04795555604828728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,2048,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,512,0.016901334126790363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,64,0.01238666640387641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,768,0.003165333428316646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,7168,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,256,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,3584,0.07044355736838447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,512,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,1536,0.05155555407206217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,6144,0.05965244770050049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,512,0.022954665952258643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,768,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,16384,0.02355466617478265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,4096,0.014426666829321118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,3072,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,128,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,3072,0.03917510973082648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,3584,0.007699555820888943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,512,0.014640889234013028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,4096,0.038853334056006536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,65536,0.15815644794040257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,768,0.005715555614895291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,5120,0.018015111486117046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,2048,0.029871109459135268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,10240,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,65536,0.2868408891889784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,768,0.021556443638271753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,3072,0.02628533376587762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,10240,0.087226668993632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,256,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,1024,0.007747555772463481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,65536,0.09254222446017796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,51200,0.5926239755418565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,256,0.015592000550693937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,7168,0.029501332177056208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,51200,2.5792196061876083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,1024,0.03674844569630093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,51200,0.062464886241488986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,5120,0.3769822120666504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,7168,0.05714044306013319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,65536,4.324724409315321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,768,0.41735466321309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,16384,0.08759733041127522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,51200,0.49611732694837785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,256,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,64,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,12288,0.04874666531880697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,1024,0.012110222544935016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,32,0.005375999957323074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,1536,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,32,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,6144,0.5214622285630968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,3072,0.037498666180504694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,2048,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,5120,0.013399110900031196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,2048,0.03889155719015334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,128,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,8192,0.08362578021155463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,32,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,16384,0.11704177326626247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,1024,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,2048,0.01240711079703437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,4096,0.016142222616407607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,12288,0.03177422285079956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,7168,0.03418755531311035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,5120,0.02463200026088291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,65536,0.7748213344150119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,2560,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,2560,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,6144,0.04394400119781494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,10240,0.012716444830099741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,3584,0.004654222064548069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,128,0.013761777844693927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,256,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,3072,0.01889333294497596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,10240,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,1024,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,7168,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,16384,0.03388444582621256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,1536,0.009137777818573846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,3584,0.04617422156863742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,8192,0.14961777793036565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,12288,0.08800444338056777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,51200,1.3044142193264432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,6144,0.09413244326909383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,51200,2.8424390157063804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,12288,0.030048890246285334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,768,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,1024,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,128,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,3072,0.057672891351911754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,6144,0.11463467280069987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,12288,0.045125332143571645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,2048,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,128,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,3072,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,6144,0.09442489014731513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,256,0.004722666823201709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,3584,0.008439111212889353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,16384,0.01999288962946998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,2048,0.010310221877362993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,3584,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,256,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,512,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,8192,0.03983199927541945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,1024,0.029959112405776978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,7168,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,6144,0.04971111151907179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,5120,0.4139137797885471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,3584,0.016823111308945548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,16384,0.06549866994222005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,2560,0.008050666915045844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,4096,0.12178666061825222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,2048,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,12288,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,5120,0.18205600314670137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,10240,0.025053333905008104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,1536,0.015447111593352424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,2560,0.1505848831600613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,16384,0.05106044477886624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,51200,0.5304880142211914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,10240,0.06215911441379123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,256,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,1536,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,32,0.016417778200573392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,12288,0.06778844197591145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,2048,0.00655733338660664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,16384,0.06366044282913208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,65536,0.060876442326439746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,6144,0.016720889343155753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,3584,0.019612444771660697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,5120,0.027632888820436265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,3072,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,16384,0.18077954981062147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,256,0.006214222146405115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,512,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,512,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,12288,0.07417688767115276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,12288,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,16384,0.04226933254135979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,2048,0.029470221863852605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,2560,0.020852444900406733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,128,0.061556445227728955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,64,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,128,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,1024,0.2729297743903266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,256,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,64,0.02230577833122677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,3584,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,4096,0.02089599933889177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,3584,0.31463556819491917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,6144,0.03178222311867608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,3072,0.07723910941017999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,1024,0.0037831109431054858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,3584,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,4096,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,8192,0.13820444213019475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,256,0.015295111470752291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,2048,0.24927822748819986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,10240,0.10122133625878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,2560,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,512,0.012490666574902005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,2560,0.03418133325046963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,10240,0.03923111160596212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,128,0.0021084443562560612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,512,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,16384,0.06993333498636882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,6144,0.06001600291993883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,768,0.038988444540235735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,128,0.005764444255166584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,5120,0.31917866071065265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,5120,0.035275555319256253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,5120,0.011571555501884885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,6144,0.1322480042775472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,3072,0.02232088810867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,768,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,16384,0.05539822247293261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,128,0.010957332948843638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,16384,0.05639288822809855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,8192,0.03525244527392917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,768,0.00776444458299213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,3072,0.045253333118226796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,5120,0.10277599758572048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,6144,0.030844443374209937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,10240,0.0855120023091634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,12288,0.026434666580624048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,5120,0.19290133317311606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,12288,0.05979733334647285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,65536,1.2274320390489366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,4096,0.04215822286076016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,2048,0.0058346668051348785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,1024,0.025732444392310247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,65536,5.07808091905382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,8192,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,128,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,7168,0.02180266711446974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,4096,0.062163558271196157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,3584,0.03806489043765598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,1536,0.01890222231547038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,6144,0.021691555778185528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,128,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,2560,0.08703200022379558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,3584,0.024297777149412367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,65536,0.057518223921457924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,768,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,512,0.011090666883521609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,51200,0.9364444944593641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,5120,0.036706666151682533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,16384,0.022299556268586054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,3584,0.008041777544551426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,256,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,768,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,4096,0.05393777622116936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,16384,0.05971199936336941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,16384,0.052341332038243614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,4096,0.024460444847742718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,64,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,2048,0.0177484436167611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,32,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,3584,0.02530844509601593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,256,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,2560,0.0476453337404463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,512,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,2048,0.023659555448426142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,128,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,3584,0.021384000778198242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,256,0.0033973333322339584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,12288,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,1024,0.07488177882300483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,7168,0.02146311104297638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,512,0.01142133358452055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,7168,0.181604438357883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,12288,0.19242400593227812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,768,0.007204444044166141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,2560,0.02743822170628442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,8192,0.023609777291615803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,1536,0.008186666501892937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,5120,0.19415644804636636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,65536,0.03436444534195794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,10240,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,5120,0.1837093300289578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,4096,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,2048,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,5120,0.021528888079855178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,4096,0.022638221581776936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,512,0.0024782222592168385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,5120,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,256,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,7168,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,3584,0.07135111093521118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,51200,2.0347396002875433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,2560,0.030227555168999568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,10240,0.04819644490877787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,7168,0.007583111524581909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,51200,0.05588622225655449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,256,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,256,0.030646221505271062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,4096,0.04056888818740845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,128,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,10240,0.1050213310453627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,3072,1.0305378172132704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,32,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,3072,0.13210933738284641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,7168,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,128,0.023359111613697473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,32,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,16384,0.07759733332528009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,6144,0.05436444282531738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,3072,0.008847999903890822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,1024,0.0038817777401871155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,65536,0.11414933204650879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,51200,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,256,0.013272888958454132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,1024,0.022649778260125056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,7168,0.08500978019502427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,2048,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,4096,0.02909066610866123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,51200,0.8476524353027344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,4096,0.213119109471639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,51200,0.07050044669045342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,768,0.013547555440002017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,12288,0.7311635547214084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,65536,0.31247109837002224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,5120,0.00904622260067198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,10240,0.09821689128875732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,512,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,128,0.015453333655993143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,64,0.030583110120561387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,51200,0.45232534408569336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,2048,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,65536,0.10981688896814983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,12288,0.7304248809814453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,7168,0.08289511336220635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,768,0.01408622165520986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,16384,3.159616894192166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,3584,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,7168,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,51200,0.7892755402459039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,256,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,2048,0.007328000333574083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,1536,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,256,0.11899733543395996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,768,0.043226665920681424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,32,0.0018097776919603348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,2560,0.15128533045450845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,2560,0.052708443668153554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,8192,0.1271253294414944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,16384,0.021439111895031396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,256,0.05592888593673706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,51200,0.41858845286899143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,10240,0.018516444497638278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,1024,0.014925332532988654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,8192,0.1539706654018826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,3584,0.203311112191942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,4096,0.008160889148712158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,4096,0.012640888492266336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,10240,0.035424888134002686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,7168,0.036566221051745944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,1024,0.006208000083764394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,2560,0.12590577867296007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,64,0.040822221173180476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,8192,0.02677333354949951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,10240,0.02516088883082072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,128,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,32,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,256,0.023873777853118047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,16384,0.238975101047092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,65536,0.658083544837104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,16384,0.020257777637905545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,512,0.023344000180562336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,512,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,7168,0.07287644677691989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,16384,0.05383822321891785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,128,0.04092177748680115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,1536,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,16384,0.021055999729368422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,3584,0.007892444729804993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,512,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,10240,0.3953999943203396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,128,0.012392889294359418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,2560,0.030470222234725952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,1024,0.1272488832473755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,1536,0.006530666516886816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,256,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,768,0.005642666584915585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,12288,0.050127999650107495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,12288,0.013479111095269522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,2048,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,128,0.013863999810483722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,768,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,5120,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,512,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,5120,0.048510223627090454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,6144,0.009739555418491364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,1024,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,12288,0.10637333658006455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,2560,0.03453777896033393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,32,0.007087111473083496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,3584,0.011951111257076263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,4096,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,12288,0.03637422124544779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,32,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,12288,0.03264977865748935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,768,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,3584,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,5120,0.02091022166940901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,3072,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,3584,0.03705599904060364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,7168,0.011437333292431302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,3584,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,3072,0.03299999899334378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,3072,0.11858133474985759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,65536,1.4442568884955511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,5120,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,3072,0.046966221597459584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,2560,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,2560,0.03628888726234436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,1024,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,32,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,3072,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,51200,0.052687999274995595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,1536,0.020449777444203694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,2560,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,2560,0.7337537871466743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,16384,0.29023199611239964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,3072,0.008647110727098253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,8192,0.02369155486424764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,5120,0.0690515571170383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,8192,0.01591911084122128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,8192,0.021935999393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,65536,0.25965955522325307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,256,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,5120,0.01961866683430142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,4096,0.014059555199411182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,64,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,2560,0.008619555168681675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,768,0.0031395554542541504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,8192,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,51200,0.07825422286987305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,4096,0.029870222012201946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,12288,0.1401475535498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,128,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,4096,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,1536,0.022867555419603985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,65536,0.019732443822754752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,8192,0.02739288906256358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,65536,0.06975466675228544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,256,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,2560,0.025238222546047632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,12288,0.016858667135238647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,5120,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,256,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,1024,0.018599111172888014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,3072,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,2560,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,12288,0.013422222601042854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,10240,0.024678222007221643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,1024,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,1024,0.02534488836924235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,10240,0.03322044346067641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,7168,0.05996444490220812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,2560,0.013033777475357056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,3072,0.0836026668548584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,10240,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,3072,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,2560,0.01793599956565433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,1024,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,2048,0.015170666906568738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,128,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,1536,0.17657689253489176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,3072,0.025273778372340735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,512,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,7168,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,2048,0.052912887599733144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,128,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,3072,0.007009777757856581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,12288,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,3072,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,32,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,1536,0.017266665895779926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,128,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,8192,0.14365243911743164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,1024,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,7168,0.019137778215938143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,10240,0.03394844465785556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,512,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,768,0.007542221910423703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,768,0.01773955590195126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,768,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,8192,0.015798222687509324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,256,0.006346666564544042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,10240,0.7396195729573568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,2048,0.021985777550273474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,4096,0.010758221977286868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,512,0.020649777518378366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,8192,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,2560,0.0846231116188897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,3584,0.04181866513358223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,12288,0.031544887357287936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,3584,0.012517333030700684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,2560,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,4096,0.05018755462434557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,512,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,2048,0.0987866653336419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,6144,0.018989332848125033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,65536,27.020499335394963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,12288,0.06590221987830268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,8192,0.16507644123501247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,1536,0.003568888952334722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,51200,0.20262844032711455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,7168,0.013192888763215808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,12288,0.04055733482042948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,64,0.008368888662921058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,512,0.013567999833159976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,4096,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,12288,0.030343112018373277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,768,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,65536,0.8192062377929688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,7168,0.041484445333480835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,10240,0.019184889064894784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,6144,0.031109333038330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,16384,0.060532444053226046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,65536,0.5847822295294868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,256,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,12288,0.22637777858310273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,128,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,10240,0.08551733361350165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,1536,0.027242667145199243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,3584,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,65536,26.196395874023438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,3072,0.02036711076895396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,51200,0.1590284506479899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,6144,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,512,0.010320888625250923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,6144,0.04302488764127096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,64,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,8192,0.029824889368481104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,16384,0.21711733606126574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,1024,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,2560,0.1511955526140001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,512,0.040960888067881264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,4096,0.09487644169065688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,1024,0.036347554789649114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,1536,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,1536,0.00535733335547977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,3072,0.015595555305480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,1024,0.02589688864019182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,65536,0.24347111913892958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,1024,0.08271644512812297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,3072,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,7168,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,7168,0.010024888647927178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,12288,0.011781333221329583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,2560,0.02287111183007558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,12288,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,10240,0.06217955880694919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,12288,1.3026239607069228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,8192,0.010846222440401712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,51200,0.053378668096330434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,3584,0.22701777352227104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,2560,0.016069332758585613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,12288,0.08184444242053561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,51200,0.051015999582078725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,7168,0.08141511016421847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,6144,0.16893422603607178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,256,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,3584,0.012859555582205454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,7168,0.05462400118509928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,3584,0.0452497767077552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,65536,0.10534666644202338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,1024,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,256,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,4096,0.025711110896534387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,4096,0.05236533284187317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,10240,0.015066666735543145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,3584,0.06679377953211467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,7168,0.11489333046807183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,5120,10.285585191514757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,8192,0.018904000520706177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,2560,0.00702044450574451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,32,0.1725697782304552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,768,0.05009955498907301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,32,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,10240,0.018950222267044913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,3584,0.037951111793518066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,512,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,3584,0.03011111087269253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,1536,0.07647111018498738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,512,0.01923199991385142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,6144,0.014532445205582513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,4096,0.022069333328141108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,1536,0.004473777694834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,12288,0.08063822322421603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,1536,0.01902044481701321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,6144,0.009760888914267222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,2048,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,512,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,10240,0.030753778086768255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,1024,0.025369778275489807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,1024,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,16384,0.03914311197068956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,3584,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,12288,0.013332444760534497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,10240,0.0361697773138682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,512,0.009299555586444007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,2560,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,6144,0.04342400034268697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,8192,0.025120000044504803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,51200,0.3168959882524279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,10240,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,128,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,1536,0.022700443863868713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,8192,0.020568889048364427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,6144,0.08354133367538452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,6144,0.023708444502618577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,768,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,2560,0.010194666683673859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,3072,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,6144,0.040595554643207125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,7168,0.09781866603427464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,1536,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,256,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,10240,0.03780533207787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,12288,0.05423288875155979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,10240,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,3072,0.14514311154683432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,256,0.014504000544548035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,1024,0.015832000308566622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,6144,0.04458933406405979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,768,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,12288,0.16279378202226427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,2048,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,8192,0.0302782224284278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,2560,0.022666666242811415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,3584,0.009558222360081142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,51200,0.10408177640702988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,3584,0.01886755559179518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,1024,0.04334311021698845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,7168,0.02052088909678989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,2048,0.03013333347108629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,6144,0.010114666488435533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,65536,0.1508622169494629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,1536,0.026956443985303242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,2560,0.39929689301384824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,7168,1.3241679933336046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,768,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,768,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,5120,0.023983110984166462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,2560,0.01039644413524204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,2048,0.0112631113992797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,16384,0.0416515568892161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,2048,0.014276444911956787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,5120,0.014569777581426831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,51200,0.03146133489078946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,1024,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,4096,0.026915555198987324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,32,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,8192,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,1024,0.03157777918709649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,12288,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,51200,0.4193395508660211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,1024,0.08322755495707194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,256,0.005976000179847081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,8192,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,8192,0.09450844261381362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,12288,3.6395361158582897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,51200,1.4083902570936415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,6144,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,10240,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,4096,0.11699644724527995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,3584,0.07059555583530001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,65536,0.05979910824033949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,2560,0.014498665928840637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,3072,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,12288,0.06533422072728474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,1024,0.006339555399285422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,32,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,256,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,512,0.006776888751321369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,32,0.08238933483759563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,512,0.011433777709801992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,8192,0.08868533372879028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,7168,0.03103733393881056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,65536,0.08060266574223836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,768,0.005631999837027655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,2048,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,1024,0.017228444417317707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,1536,0.014194667339324951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,256,0.015192889504962497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,7168,0.3461306624942356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,5120,0.03124089042345683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,7168,0.01369511087735494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,3072,0.01815733313560486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,5120,0.13924266232384577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,4096,0.12944977813296848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,51200,0.056477334764268666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,7168,0.12807822227478027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,65536,0.9581901762220594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,128,0.010552888943089379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,2048,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,6144,0.047091556919945605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,4096,0.3141386773851183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,16384,0.5446524620056152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,65536,0.20901066727108428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,2560,1.9923937055799696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,6144,0.07064711385303073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,256,0.13298932711283365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,1024,0.040688001447253756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,1024,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,2560,0.027956446011861164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,3584,0.030119998587502375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,64,0.0021111110432280437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,2048,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,2560,0.147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,2560,0.01220266686545478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,1024,0.009936889012654623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,5120,0.01891644464598762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,768,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,65536,0.5583937962849935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,3072,0.07027466429604425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,7168,2.0741733974880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,768,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,51200,0.07401155763202243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,16384,0.054577777783075966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,256,0.046168890264299184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,768,0.05485511157247755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,7168,0.13008888562520346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,2560,0.025034666061401367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,128,0.017258667283587985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,1024,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,768,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,768,0.008519111408127679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,4096,0.07994488875071208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,2560,0.6556160185072157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,2048,0.01142133358452055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,1024,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,51200,1.5793244043986003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,10240,0.035550223456488714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,4096,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,3072,0.03688977824317084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,1536,0.028322666883468628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,65536,8.974432203504774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,10240,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,12288,0.07471555471420288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,16384,0.16968177424536812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,12288,0.08382399876912434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,768,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,32,0.021711111068725586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,3584,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,512,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,3584,0.0073902226156658595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,1024,0.020303999384244282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,128,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,1536,0.01993777851263682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,51200,0.6514568858676487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,12288,0.060360888640085854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,65536,0.3364577823215061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,3584,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,5120,0.024754666619830664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,8192,0.029139555162853662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,128,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,1024,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,128,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,6144,0.07487199703852336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,10240,0.12490933471255833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,5120,0.08290933238135444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,65536,0.06232533189985487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,10240,0.03289777702755398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,51200,1.1561066309611003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,768,0.017616000440385606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,16384,0.030239999294281006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,1536,0.0336097776889801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,2560,0.16091644763946533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,32,0.03670755690998501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,256,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,12288,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,65536,1.3330781724717882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,8192,0.02642400066057841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,5120,0.16436533133188883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,128,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,32,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,8192,0.02548355526394314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,12288,0.05729599793752035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,64,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,12288,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,768,0.020569778150982328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,51200,0.4870737923516168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,5120,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,4096,0.05684355232450697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,6144,0.04753955536418491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,768,0.0037164443896876443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,3072,0.1552142169740465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,51200,0.2288720077938504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,64,0.043326222234302096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,1024,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,2560,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,51200,0.049752000305387706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,8192,0.15722755591074625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,16384,0.04982133375273811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,2560,0.015384000208642749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,16384,0.025241777300834656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,51200,1.520859506395128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,65536,0.7495626873440213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,12288,0.39121066199408633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,3072,0.015589333242840238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,1024,0.02089511189195845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,1536,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,64,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,4096,0.010853333605660332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,3072,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,6144,0.042066666815016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,1024,0.016805332567956712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,4096,0.012260444462299347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,6144,0.05923200315899319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,1024,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,2048,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,6144,0.01309866706530253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,5120,0.009382222261693742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,2560,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,512,0.017873777283562552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,2048,0.004330666528807746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,2560,0.0206675562593672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,65536,0.12946400377485487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,10240,0.03966933488845825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,10240,0.03640177845954895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,2560,0.012375999655988483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,6144,0.023932443724738225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,65536,0.07647466659545898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,8192,0.06018933322694567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,1024,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,5120,0.05563466416464912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,8192,0.13841956191592747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,3584,0.04877777894337972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,65536,0.1699999968210856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,512,0.006837333242098491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,4096,0.051029334465662636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,32,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,8192,0.14821333355373806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,2048,0.023043556345833674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,7168,0.009875555833180746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,51200,0.09521244631873237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,256,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,65536,1.4022764629787867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,768,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,2560,0.04054133428467645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,7168,0.028479999966091577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,10240,0.7044115596347384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,2048,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,128,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,16384,0.1764648887846205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,12288,0.020225778222084045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,5120,1.1869502597384982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,65536,0.07862222194671631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,768,0.01601777805222405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,128,0.007902222375075022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,51200,0.1054426696565416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,2048,0.41973421308729386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,7168,0.017143110434214275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,3072,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,8192,0.025616889198621113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,5120,0.021893334057595994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,4096,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,2560,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,128,0.013898667362001208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,10240,0.02289955483542548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,2560,0.05719733238220215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,10240,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,256,0.005316444569163852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,10240,0.012517333030700684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,128,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,32,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,16384,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,3072,0.02734755476315816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,65536,0.04263733492957222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,51200,0.09909866915808783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,51200,0.19626843929290771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,3072,0.006870222174459034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,65536,6.353068457709418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,2560,0.01089244418674045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,32,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,64,0.0028293333533737394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,768,0.018509333332379658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,64,0.06421244144439697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,32,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,16384,0.06271466943952772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,65536,0.07947733004887898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,51200,0.12086844444274902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,65536,1.2389412985907662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,16384,0.0815182195769416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,512,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,5120,1.0266648398505318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,2560,0.02397511071628994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,2560,0.01586666703224182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,128,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,8192,0.01477688882086012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,3584,0.1018844445546468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,3584,0.03980266716745164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,4096,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,8192,0.051274667183558144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,512,0.020015999674797058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,2048,0.02446222139729394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,3072,0.030407110850016277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,51200,0.0676399999194675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,6144,0.03163822160826789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,2048,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,65536,0.060461335712009005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,1024,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,51200,0.11944711208343506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,51200,0.7303395801120334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,512,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,7168,0.059770663579305015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,12288,0.0765626695421007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,10240,0.42257600360446507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,8192,0.41491466098361546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,65536,0.06691022051705255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,128,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,51200,0.3671120007832845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,256,0.033599111768934466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,3584,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,5120,0.012873777912722694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,5120,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,2048,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,65536,0.14805244074927434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,256,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,3584,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,65536,0.06570666366153292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,8192,0.07701689004898071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,3584,0.028899557060665552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,6144,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,3072,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,64,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,1024,0.1166951126522488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,51200,0.15756089157528347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,1024,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,64,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,32,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,51200,1.375393761528863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,1024,0.016195555528004963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,256,0.026899556318918865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,12288,0.04077333211898804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,6144,0.1089973317252265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,5120,0.010503110786279043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,7168,0.014261333478821648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,1536,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,5120,0.0447306666109297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,2560,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,16384,0.016362667083740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,5120,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,7168,0.018401778406567044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,2560,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,7168,0.015370666980743408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,8192,0.030321776866912842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,3072,0.02167111138502757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,32,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,1024,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,1536,0.00962488849957784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,5120,0.014711111783981323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,3584,0.11016888750924005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,10240,0.023747555083698694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,7168,0.16019554932912192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,2560,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,6144,0.1997413370344374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,16384,0.026337777574857075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,1024,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,4096,0.5165466732449001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,768,0.0034773333205117118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,1024,0.0038577777643998465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,12288,0.014236445228258768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,4096,0.145541336801317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,1536,0.22157600190904406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,12288,0.07292355431450738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,12288,0.020289777053727042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,512,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,4096,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,12288,0.23931556277804902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,4096,0.01108266661564509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,16384,0.021015110943052504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,3072,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,1536,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,3584,0.008612444831265343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,1536,0.17251288890838623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,10240,0.028103109863069322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,6144,0.03758488761054145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,6144,1.1395538118150499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,1024,0.016571554872724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,1024,0.02728355593151516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,512,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,1536,0.01096088853147295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,256,0.005631999837027655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,768,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,512,0.018216000662909616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,65536,0.08903377585940891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,6144,0.0521360006597307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,51200,0.06700177987416585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,32,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,16384,0.038078222009870745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,2560,0.012125333150227865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,2048,0.022101332743962605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,2048,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,6144,0.011559999651379056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,32,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,64,0.08835022317038642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,16384,0.04061333338419596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,3072,0.023659555448426142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,2048,0.017295110556814406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,5120,0.020963556236690942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,1536,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,32,0.001959111127588484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,51200,0.14118756188286677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,512,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,8192,0.039100444979137845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,3584,0.00981155534585317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,2048,0.0273973329199685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,3584,0.06073333157433403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,2560,0.029318223396937054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,256,0.028135998381508723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,7168,0.02219911085234748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,16384,0.022252443763944838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,51200,0.1859582265218099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,6144,0.03422755665249295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,1536,0.07418488793902926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,2560,0.06236977709664238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,512,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,6144,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,512,0.008210666477680206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,2048,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,51200,0.12917066944970026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,64,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,65536,1.9969083997938368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,128,0.02158311174975501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,32,0.027552889453040227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,512,0.017095999585257638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,2048,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,768,0.02009600069787767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,3584,0.8069760004679362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,768,0.00996088898844189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,512,0.12221422460344102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,64,0.002566222308410539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,2048,0.009400000174840292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,256,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,256,0.04929688904020521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,768,0.012493333054913415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,16384,0.03427822060055203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,65536,0.07410311036639743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,51200,0.11661688486735027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,10240,0.03625422385003831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,12288,0.06115466356277466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,10240,0.01851466629240248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,2048,0.06489955716662936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,16384,0.022660444180170696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,6144,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,51200,0.04735733403099907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,6144,0.04176622298028734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,2048,0.023040889037979975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,2560,0.004925333377387789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,3072,0.07014577918582492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,1536,0.016920000314712524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,51200,0.505400021870931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,1024,0.0038373333712418876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,65536,0.06862844361199273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,16384,0.047263112333085805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,8192,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,6144,0.017285333739386666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,2048,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,4096,0.01796888808409373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,65536,4.124134063720703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,6144,0.03840622305870056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,6144,0.01722755531469981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,65536,0.3230080074734158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,1024,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,6144,0.01939288940694597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,10240,0.4390675491756863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,3072,0.026527111728986103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,3072,0.025197333759731714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,51200,0.11546756161583795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,3072,0.02964799933963352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,6144,0.07101777527067396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,64,0.0022844444546434614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,2048,0.01889155473974016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,2560,0.022653333014912073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,3584,0.02885066800647312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,6144,0.010510221951537661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,8192,0.024735111329290602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,128,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,7168,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,3072,0.4599733352661133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,16384,0.14873422516716853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,768,0.0704524450831943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,1024,0.004337777694066365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,1536,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,2560,0.0103511114915212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,128,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,5120,0.019838222199016146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,1536,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,1024,0.006514666808976068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,5120,0.1864782174428304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,51200,0.06637955374187894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,3584,0.0813902219136556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,12288,0.17967822816636828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,16384,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,5120,0.014504000544548035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,65536,1.1140356063842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,8192,0.46861690945095485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,1536,0.007705777883529663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,3584,0.016568000117937725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,51200,0.10740533139970566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,128,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,5120,0.004660444541109933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,51200,0.19205333126915825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,6144,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,512,0.006728000111050076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,6144,0.20064266522725424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,4096,0.004593777573770947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,3072,0.01402311192618476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,768,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,256,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,768,0.006004444426960415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,1024,0.008845333423879411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,51200,0.03795999950832791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,51200,0.4207911226484511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,768,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,5120,0.12510489092932806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,32,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,5120,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,2560,0.013944000005722046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,5120,0.11213244332207574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,51200,11.714558919270834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,512,0.05263288815816244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,512,0.011485333244005838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,12288,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,3584,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,5120,0.03163289030392965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,16384,0.021239999267790053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,256,0.01498222185505761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,51200,0.049288001325395375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,32,0.0217884447839525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,3584,0.03151999910672506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,256,0.02476355599032508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,64,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,5120,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,4096,0.10602400037977432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,12288,0.31099645296732586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,5120,0.03590666585498386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,4096,0.015857777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,65536,0.08392444584104751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,4096,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,6144,0.10827822155422634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,1536,0.03581777877277798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,1536,0.0692124432987637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,256,0.022055110997623865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,5120,0.0422435568438636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,7168,0.06332977612813313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,51200,0.6120542420281304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,2560,0.012483555409643384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,6144,0.07220800055397882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,1536,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,8192,0.025222222010294598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,256,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,51200,0.05158222383923001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,128,0.015453333655993143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,65536,2.2041609022352433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,3584,0.213682664765252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,768,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,8192,0.4029235574934218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,512,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,1024,0.09031111001968384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,12288,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,1024,0.020900444851981267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,2560,0.008353778057628209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,51200,0.06329955657323201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,3584,0.012320000264379712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,2048,0.01812000075976054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,16384,0.039148443275027804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,32,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,51200,0.052685333622826465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,16384,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,2048,0.02350133326318529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,1536,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,1024,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,2048,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,1536,0.028639998700883653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,16384,0.01716888944307963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,2560,0.10527111424340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,3584,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,6144,0.03211466802491082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,2048,0.10297155380249023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,64,0.018562666243977018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,16384,0.01853511068556044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,32,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,51200,0.05052888724539015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,3072,0.09371199872758652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,12288,0.016514667206340365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,8192,0.09063466389973958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,2048,0.039242668284310236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,4096,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,128,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,2560,0.013902222116788229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,10240,0.09047200282414754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,5120,0.33151377571953666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,5120,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,4096,0.04890666736496819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,8192,0.049957331683900624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,8192,0.05663377708858914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,6144,0.046500444412231445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,6144,0.005196444276306364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,12288,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,512,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,2048,0.06928266419304742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,1536,0.01923822197649214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,51200,0.05243022243181864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,10240,0.027121777335802715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,6144,0.116321775648329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,65536,0.10878488752577041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,51200,0.1713848908742269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,128,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,2048,0.026886221435334947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,2560,0.028871999846564397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,64,0.01677600045998891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,10240,0.1981955501768324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,512,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,768,0.04405511087841458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,2048,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,2048,0.025073778298166063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,8192,0.20626489321390787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,7168,0.03216444452603658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,4096,0.050541334682040744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,65536,0.03700266612900628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,12288,0.5608320236206055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,16384,0.14349422189924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,6144,0.08524088727103339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,512,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,3072,0.020075556304719713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,2048,0.006864000111818314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,7168,0.011828444070286222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,1024,0.012120889292822944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,51200,0.0961022244559394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,7168,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,2560,0.022608000371191237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,32,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,51200,0.06612444586224027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,65536,0.06649688879648845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,768,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,1536,0.010183999935785929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,7168,0.012164444559150271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,64,0.0432168874475691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,12288,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,128,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,64,0.002152888932161861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,3072,0.30587466557820636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,2048,0.010144000252087912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,768,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,768,0.010317333042621613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,3072,0.017408889200952318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,256,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,512,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,8192,0.077711111969418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,64,0.0037413334680928122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,128,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,6144,0.04474400149451362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,128,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,2560,0.03401688734690348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,4096,0.13224889172448054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,10240,0.17777155505286324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,32,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,2560,0.13503911760118273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,5120,0.011579555769761404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,128,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,1536,0.4986106554667155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,4096,0.02307555576165517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,512,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,5120,0.01756800048881107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,10240,12.859464009602865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,2560,0.021686222818162706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,512,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,3072,0.48997868431939023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,12288,0.0188711103465822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,1536,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,2560,0.7214986483256022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,7168,0.02998577886157566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,3584,0.025412445267041523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,65536,0.44968711005316836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,3584,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,4096,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,256,0.009135111338562435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,1536,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,768,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,1536,0.004634666774008009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,5120,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,512,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,2048,0.019527110788557265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,512,0.00868888861603207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,4096,0.08575999736785889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,3584,0.008028444316652086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,2048,0.050809777445263334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,1024,0.017528888252046373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,1024,0.005315555466545953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,51200,0.05229511194758945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,512,0.08492177724838257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,512,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,3584,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,16384,0.1360666619406806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,3072,0.12831910451253256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,65536,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,256,0.03302399979697333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,5120,0.024728000164031982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,65536,0.1253004471460978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,12288,0.03913866811328464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,5120,0.05679821968078613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,12288,0.018741332822375827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,128,0.0019439999014139175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,3072,0.013779555757840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,4096,0.07901777823766072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,1536,0.0314995547135671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,2048,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,6144,0.01605422298113505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,1024,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,65536,0.06324355469809638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,4096,0.012122666670216454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,6144,0.07348089085684882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,64,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,16384,0.06423733631769817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,65536,0.2167226605945163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,6144,0.020745777421527438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,3584,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,768,0.0343537794219123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,32,0.011397333608733283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,51200,0.21806132793426514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,32,0.11504088507758246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,3584,0.1880497799979316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,1024,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,256,0.004980444494220945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,8192,0.04320533408059014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,3584,0.09710577461454604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,3072,0.10096089045206706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,4096,0.09531733062532212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,51200,0.22221867243448892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,4096,0.03577155537075467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,6144,0.29332976871066624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,65536,0.10036800305048625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,6144,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,4096,0.014919110470347934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,128,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,1024,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,51200,1.5028844409518773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,128,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,16384,0.028159111738204956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,4096,0.11415644486745198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,128,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,2560,0.012612444659074148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,256,0.015544000599119397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,12288,0.26377866003248424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,5120,0.01382844481203291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,2048,0.02995733420054118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,128,0.003399111123548614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,12288,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,10240,0.05223200056287977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,256,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,2048,0.519902229309082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,2048,0.011007111105653973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,1024,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,1024,0.01460444430510203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,8192,0.22717600398593477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,6144,0.04436799883842468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,6144,0.10172088940938313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,12288,0.01391999920209249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,65536,0.11024266481399536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,8192,0.08020355304082234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,3584,0.007711999946170383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,2560,0.02405333353413476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,12288,0.21822578377193877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,4096,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,5120,0.020938666330443487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,4096,0.04731733269161648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,128,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,51200,0.1725306643380059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,1024,0.04453688859939575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,51200,0.05150755577617221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,128,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,10240,0.012452444268597497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,32,0.00608800020482805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,768,0.014850666125615438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,16384,2.941875457763672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,2560,0.0550746652815077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,1536,0.03763466742303636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,6144,0.045145776536729604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,7168,0.024656888511445787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,128,0.019558222757445443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,64,0.0020808887978394828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,1024,0.042520887321896024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,1024,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,6144,0.025077333052953083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,2048,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,3584,0.00638755535085996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,12288,0.6625119845072428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,8192,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,65536,0.6784426901075574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,2048,0.02465244465404087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,5120,0.02258577777279748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,7168,0.08559644222259521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,16384,0.0295466681321462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,5120,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,51200,0.05615466833114624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,4096,0.21579821904500326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,3072,0.07039999961853027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,1024,0.006080889039569431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,8192,0.1398435566160414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,7168,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,256,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,51200,0.09618933333290948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,16384,0.05446666810247633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,3072,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,2048,0.018166222506099276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,51200,0.21931821770138213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,1536,0.009760888914267222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,128,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,16384,0.03392622205946181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,1536,0.0462773342927297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,2048,0.17335377799140084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,16384,0.018333334061834548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,32,0.13731911447313097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,10240,0.13510578208499485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,65536,0.22477600309583876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,2048,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,3072,0.03496977686882019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,7168,0.012109333442317115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,768,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,768,0.015248000621795654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,768,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,128,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,3072,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,5120,0.30238133006625706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,16384,0.11928178204430474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,16384,0.07320800092485216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,16384,0.019722667005327012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,3072,0.0092284447617001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,2560,0.5089324315388998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,2560,0.029493331909179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,7168,0.054824001259273954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,65536,0.06184266673194038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,65536,16.845430162217884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,5120,0.16957333352830675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,4096,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,10240,0.025571554899215698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,2560,0.007894222107198503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,2048,0.008556444611814287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,5120,0.06092266903983223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,10240,0.03914311197068956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,8192,0.01923466722170512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,512,0.03982222080230713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,2048,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,65536,0.07927022377649943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,4096,0.1206666628519694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,256,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,4096,0.008047999607192146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,65536,0.20507200558980307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,65536,0.09183733330832587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,256,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,3584,0.008902221918106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,3584,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,3584,0.026191999514897663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,3072,0.014498665928840637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,5120,0.01664266652531094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,2048,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,64,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,51200,0.8118746545579699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,32,0.007592000067234039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,4096,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,5120,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,8192,0.22178044584062365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,10240,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,256,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,512,0.07931377490361531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,256,0.0855368905597263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,8192,0.2618782255384657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,7168,0.042135109504063926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,7168,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,256,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,3072,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,768,0.023648000425762598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,7168,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,1024,0.005981333139869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,512,0.016466667254765827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,12288,1.4455786810980902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,12288,0.01756266587310367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,2560,0.016880000631014507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,3072,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,7168,0.010313777459992303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,12288,0.3086408774058024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,51200,0.06014221906661987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,256,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,7168,0.0789546635415819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,51200,0.05480533175998264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,65536,0.35883378982543945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,65536,0.15419467290242514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,3072,0.0332897769080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,8192,0.011880888707107969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,6144,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,8192,0.08421155479219224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,128,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,3072,0.026708443959554035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,6144,0.054245332876841225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,2048,0.5360408888922797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,128,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,7168,0.01644444465637207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,2560,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,2560,0.11443377865685357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,12288,0.11205511622958714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,1536,0.045143110884560474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,1536,0.007800000409285228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,1536,0.004705777598751916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,1536,0.004327110946178436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,512,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,5120,0.033434665865368314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,1024,0.11403821574317084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,3072,0.08700978093677098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,10240,0.05430755681461758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,2048,0.09820622205734253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,512,0.03647022114859687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,8192,0.03901866740650601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,5120,0.020643555455737643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,512,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,2048,0.02953333324856228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,6144,0.01797422269980113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,12288,0.016238222519556682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,4096,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,7168,0.022051556242836848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,128,0.01957244508796268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,16384,0.031397332747777305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,768,0.007101332975758447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,768,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,2560,0.49681780073377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,256,0.018637332651350234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,5120,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,768,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,5120,0.058499554793039955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,4096,0.04752177662319607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,2048,0.047525333033667676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,768,0.0028826666788922418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,3072,0.0329093337059021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,2048,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,128,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,4096,0.03401511245303684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,6144,0.02163911031352149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,4096,0.011310222248236338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,65536,0.18659999635484484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,128,0.015950222810109455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,64,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,4096,0.008204444415039485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,8192,0.0639537771542867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,12288,0.04384533233112759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,1536,0.04858488837877909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,7168,0.01982133256064521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,2560,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,1536,0.07622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,64,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,2048,0.008040889269775814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,2560,0.012159110771285163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,12288,0.5032986534966363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,1536,0.005370666583379109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,7168,0.021378666162490845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,64,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,1536,0.004504888835880491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,2048,0.00600355573826366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,5120,0.09170666668150161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,1536,0.005716444303592046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,4096,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,16384,0.01570222278436025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,1024,0.03310488992267185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,65536,0.47096978293524844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,128,0.029126223590638902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,32,0.0063786668082078295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,512,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,512,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,12288,0.06113600068622165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,512,0.022170666191313002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,1536,0.01829155617290073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,16384,0.04586044285032484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,1536,0.015793777174419828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,3584,0.010595555934641095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,4096,0.008259555531872643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,6144,0.21658133135901558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,1536,0.08393777741326226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,16384,0.37482311990525985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,7168,0.21677866246965197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,3072,0.07804355356428358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,32,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,65536,0.022516444325447083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,2560,0.2918915483686659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,256,0.013120888835854001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,3584,0.01484266585773892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,12288,0.02565244502491421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,1536,0.037951111793518066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,512,0.012659555508030785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,512,0.01645866698688931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,16384,0.05883555279837715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,2560,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,3072,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,65536,0.19208621978759766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,2560,0.020596444606781006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,1536,0.05078666739993625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,6144,0.040115555127461754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,4096,0.024996444582939148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,16384,0.027279110418425664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,10240,0.02605955633852217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,5120,0.16993510723114014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,3584,0.04157422317398919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,2048,0.012784000072214337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,2048,0.01753777762254079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,768,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,3072,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,7168,0.027673777606752183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,16384,0.016772444049517315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,6144,0.17600622442033556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,3072,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,65536,0.09669066800011529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,7168,0.04636889033847385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,128,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,8192,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,32,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,51200,0.23666487799750435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,128,0.014456000592973499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,65536,0.22574843300713432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,2048,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,3072,0.007247111035717859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,6144,0.01201333353916804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,12288,0.018175111876593698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,512,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,65536,0.7092017597622342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,4096,0.0072266666425599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,32,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,10240,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,6144,0.17572000291612414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,1536,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,7168,0.03470222155253092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,2048,0.041478223270840116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,12288,0.03919733232922024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,8192,0.03267111049758063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,4096,0.04889600144492256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,10240,0.07972088787290785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,10240,0.16767466068267822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,128,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,6144,0.07552711168924968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,3072,0.007686222592989604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,512,0.005352889084153705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,4096,0.009508444203270806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,8192,0.13303644127315944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,16384,0.1371955606672499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,12288,0.02474844455718994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,32,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,128,0.003248888792263137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,768,0.019952888290087383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,128,0.01180355581972334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,3584,0.03173333406448364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,3584,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,5120,0.025286222497622173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,2560,0.048215111096700035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,5120,0.023406222462654114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,12288,0.03382755650414361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,12288,0.07017955515119764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,3072,0.022392888863881428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,65536,0.668524424235026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,4096,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,1536,0.015477332803938123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,2048,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,3584,0.0079297779334916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,2048,0.0410462220509847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,12288,0.026649778087933857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,10240,0.03311289019054837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,32,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,3584,0.018736888964970905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,768,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,3072,0.01533511115445031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,65536,0.09730044338438247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,16384,0.2891768879360623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,256,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,5120,0.00977955593003167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,1024,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,6144,0.11425156063503689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,512,0.03049955434269375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,7168,0.046966221597459584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,768,0.10182488626903957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,5120,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,3072,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,256,0.011788444386588203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,16384,0.2635137769911024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,51200,1.9127218458387587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,65536,0.0643084446589152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,256,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,2048,0.010629333555698395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,7168,0.0447271102004581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,8192,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,12288,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,8192,0.07140444384680854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,7168,0.010393777655230628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,256,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,2560,0.02595022155178918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,768,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,7168,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,768,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,256,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,512,0.02090399960676829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,7168,0.030914667579862807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,768,0.030065778228971694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,512,0.0070266665683852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,3072,0.021761778328153823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,65536,0.0862391127480401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,3072,0.040327111879984535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,2560,0.006892444358931647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,1024,0.002990222225586573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,8192,0.13850666417015925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,16384,0.03558311197492812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,1536,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,5120,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,4096,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,65536,0.06466399961047702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,4096,0.8482809066772461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,16384,0.043291555510626904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,2048,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,32,0.026935999592145283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,5120,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,5120,0.029826667573716905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,768,0.018566222654448617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,5120,0.0288702216413286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,8192,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,12288,0.19947732819451225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,7168,0.13581866688198513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,3072,0.03390489021937052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,8192,0.11126755343543158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,65536,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,12288,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,1024,0.06679644187291463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,1536,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,1024,0.02050755586889055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,51200,0.43454665607876247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,1536,0.03418133325046963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,65536,0.8172977765401205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,16384,0.05208977725770739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,6144,0.035064889325035944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,10240,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,3072,0.015259555644459195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,12288,0.025119110941886902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,16384,2.397905773586697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,6144,0.03433422247568766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,256,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,512,0.0032248888164758682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,8192,0.021874666213989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,16384,0.17183378007676867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,1536,0.010055111514197456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,4096,0.019630221856964957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,65536,0.5197200245327419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,5120,0.6337422264946831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,10240,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,3584,0.15179999669392905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,1536,0.01482311056719886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,128,0.0063164445261160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,256,0.0022035555707083806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,4096,0.05399644374847412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,51200,3.867847866482205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,12288,0.05708444118499756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,3072,0.015836444165971544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,10240,0.04600444436073303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,256,0.03556088937653436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,256,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,16384,0.1890568865670098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,8192,0.004847110973464118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,16384,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,4096,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,10240,1.8796382480197484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,64,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,10240,0.07342666387557983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,1536,0.009759999811649323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,16384,0.038219554556740656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,4096,0.5910319752163357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,1024,0.05807022253672282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,10240,0.03077510992685954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,768,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,512,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,128,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,4096,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,512,0.0222746663623386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,128,0.024107555548350017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,4096,0.6986231274074979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,1536,0.026014222039116755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,6144,0.07758577664693196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,3584,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,768,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,3072,0.18788088692559135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,2560,0.06485155555937025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,8192,0.021930666433440313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,51200,0.04802044563823276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,10240,0.18505777253044975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,768,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,32,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,16384,0.036068442795011736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,64,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,51200,0.46114222208658856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,3584,0.035001777940326266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,2048,0.006729777902364731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,512,0.048835555712381996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,8192,0.45486312442355686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,65536,0.08676355414920384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,256,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,2560,0.024667556087176006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,51200,0.04610133171081543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,2560,0.07429689168930054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,512,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,3584,0.044785777727762856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,6144,0.021939555803934734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,65536,0.25401777691311306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,2048,0.006096000058783426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,8192,1.1724657482571073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,3072,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,10240,0.015093333191341825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,8192,0.10340444246927898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,51200,0.6878515349494086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,256,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,3072,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,2048,0.06468800041410658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,16384,0.020440888073709276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,51200,0.290882666905721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,64,0.004987555659479565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,256,0.005526222288608551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,12288,0.009867555565304225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,768,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,2560,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,3072,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,10240,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,65536,14.25567626953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,16384,0.22434488932291666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,6144,2.7311325073242188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,2048,0.08885688914193048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,65536,0.06490133206049602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,64,0.018600889378123812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,256,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,8192,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,2048,0.01828533411026001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,12288,0.02397777802414364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,12288,0.04026222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,4096,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,4096,0.025118221839269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,5120,0.059362669785817467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,3072,0.025437333517604407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,512,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,16384,0.0307279990779029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,1536,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,5120,0.22401422924465605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,5120,0.04760533240106371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,4096,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,768,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,768,0.0291244453854031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,2560,0.026748443643252056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,512,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,512,0.019796444310082328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,1024,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,51200,0.13688711325327554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,768,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,3584,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,10240,0.03467288944456313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,5120,0.19298488563961455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,768,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,1536,0.028752889898088243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,3072,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,64,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,12288,0.03456622362136841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,512,0.02052888936466641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,512,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,10240,0.015672889020707872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,10240,0.013361777696344586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,6144,0.009783999787436591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,2560,0.03394666645261977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,10240,0.071706665886773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,128,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,12288,0.11235200034247504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,256,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,10240,0.06951110892825656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,1024,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,64,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,128,0.0752355522579617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,2048,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,4096,0.03564088874393039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,512,0.05916621949937609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,7168,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,3072,0.042259554068247475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,51200,0.04720177915361193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,512,0.00925066653225157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,512,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,10240,0.849730650583903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,65536,0.6546071370442709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,4096,0.23415467474195692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,128,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,4096,0.09180800120035808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,4096,0.02973777718014187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,2560,0.010962666736708747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,3584,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,51200,0.40252799457973903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,1536,0.201255997021993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,16384,0.2639253404405382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,1024,0.0038648889296584656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,7168,0.012452444268597497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,128,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,1024,0.015954666667514376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,1536,0.023137778043746948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,512,0.011175110936164856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,1024,0.020217777954207528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,1024,0.006810666786299811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,1024,0.014927110738224454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,6144,0.14900622102949354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,7168,0.03144444359673394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,7168,0.03377244538731045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,51200,0.17893955442640516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,64,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,3584,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,768,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,768,0.003220444545149803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,65536,0.5614568922254775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,3584,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,512,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,1024,0.1872417794333564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,6144,0.013674666484196981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,3584,0.013191999660597907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,16384,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,12288,39.484168158637154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,10240,0.08115288946363661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,768,0.004547555413511065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,16384,0.1175653272204929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,3584,0.013235555754767524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,3584,0.015041778484980265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,3072,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,768,0.07954221963882446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,6144,0.3975528875986735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,51200,0.10917155610190497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,12288,0.037782222032547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,8192,0.010196444061067369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,512,0.0964426663186815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,51200,0.42644532521565753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,12288,0.01906844476858775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,256,0.008018666671382057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,7168,0.01776444415251414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,51200,0.038667556312349104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,1536,0.08322933647367689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,7168,0.021031111478805542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,51200,1.991409725613064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,7168,0.016075554821226332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,5120,0.06377689043680827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,5120,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,1536,0.0561351113849216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,3072,0.007335111498832703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,1024,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,7168,0.02004888819323646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,128,0.11235288778940837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,8192,0.09087022145589192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,3072,0.02385422256257799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,12288,0.09614844454659356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,3584,0.15574310885535345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,16384,0.03324088785383437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,1024,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,2560,0.004865777575307422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,5120,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,51200,0.036943111154768206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,16384,0.061103999614715576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,65536,4.5018157958984375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,2048,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,512,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,5120,0.08549066384633382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,512,0.008071111308203803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,8192,0.023392000132136877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,51200,0.10651733477910359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,51200,0.07386310895284016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,7168,0.24293777677747938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,3584,0.047241777181625366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,7168,0.26287465625339085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,2560,0.08797333637873332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,12288,0.16017156177096897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,2048,0.03579555617438422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,7168,0.13408354918162027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,12288,0.08452888992097642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,7168,0.10950755410724217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,3584,0.05850311120351156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,12288,0.18879289097256133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,128,0.05836444430881076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,16384,0.07481244537565443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,4096,0.023741333021057978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,12288,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,5120,0.019638222124841478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,768,0.017902221944597032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,768,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,512,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,768,0.018575110369258456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,65536,0.9932195875379773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,4096,0.03618399964438544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,512,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,65536,0.12191643979814316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,256,0.10982933309343125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,51200,0.8565662172105578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,5120,0.011306666665607028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,768,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,768,0.043281777037514575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,2560,0.047525333033667676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,16384,0.17593333456251356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,256,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,65536,0.7831368976169162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,7168,0.02327644493844774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,1536,0.020269332660569083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,2048,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,256,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,16384,0.06313421991136339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,12288,0.10575911071565415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,3584,0.007804444266690149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,128,0.005819555371999741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,512,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,64,0.00666222224632899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,64,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,51200,0.13824000623491076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,65536,0.10133333338631524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,65536,0.0557617810037401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,7168,0.011711111499203576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,64,0.006680000159475539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,51200,0.5414835611979166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,5120,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,32,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,512,0.03197955422931247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,3584,0.19618844985961914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,32,0.008470222353935242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,3584,0.031158222092522517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,10240,2.7133102416992188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,3584,0.08708710802925958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,256,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,2560,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,16384,0.11916532781389023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,8192,0.1369475523630778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,10240,0.6096551153394911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,8192,0.05412800113360087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,128,0.016919111212094624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,8192,0.03683822353680929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,128,0.011033777561452655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,2560,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,1536,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,7168,0.19954133033752441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,5120,0.02438933319515652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,51200,0.06165511078304715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,768,0.026856889327367146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,128,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,51200,0.13992799652947321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,256,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,5120,0.0064195555945237475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,768,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,512,0.04714577727847629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,12288,1.2476622263590496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,512,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,16384,0.021789333886570398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,16384,0.22896978590223524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,2048,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,4096,0.07277866866853502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,64,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,16384,0.03053599927160475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,3072,0.03823466764556037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,768,0.05601955784691704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,2560,0.03176088796721564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,4096,0.031938665442996554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,5120,0.04526577724350823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,768,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,10240,0.21450311607784697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,32,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,65536,0.008076444268226624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,51200,0.01789866718981001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,128,0.016131555040677387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,6144,0.051487111383014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,5120,0.01737955543729994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,16384,0.030420445733600195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,16384,1.1460978190104167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,10240,0.06738666693369548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,64,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,51200,0.018039110634062026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,1536,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,2048,0.007584888901975419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,7168,0.0114720000161065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,1536,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,4096,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,1536,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,16384,0.01272266689274046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,8192,0.20895200305514863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,1536,0.019894222418467205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,128,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,2560,0.04077777928776211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,8192,0.01260444439119763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,2048,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,4096,0.015248000621795654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,6144,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,4096,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,6144,0.3398115634918213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,12288,0.044974221123589404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,7168,0.017480888300471835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,2560,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,12288,0.037011557155185275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,10240,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,2560,0.02708622150950962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,2560,0.21708888477749297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,1536,0.025418667329682246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,128,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,3584,0.1345920032925076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,3072,0.00573955559068256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,3072,0.029211554262373183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,12288,0.026978666583697002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,16384,0.03297422329584757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,6144,0.0446631113688151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,3584,0.01955111159218682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,16384,0.05555555555555555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,1024,0.01166044423977534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,3584,0.013663999736309052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,8192,0.020986666282018025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,64,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,2560,0.05190844337145487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,1536,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,12288,0.1422391070259942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,51200,0.23447378476460776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,5120,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,7168,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,10240,0.019877332780096266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,51200,0.07654577493667603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,65536,0.3998240100012885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,4096,0.03459199931886461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,1024,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,768,0.2804257869720459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,3072,0.0850177804629008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,32,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,8192,0.15767289532555476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,128,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,16384,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,4096,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,2560,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,256,0.013075555364290873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,6144,0.054915553993648954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,65536,0.04916888806555006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,7168,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,512,0.013297778036859302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,3584,0.05031822125116984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,3072,0.022724444667498272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,8192,0.02159022291501363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,2560,0.031147556172476873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,12288,0.020831111404630873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,5120,0.00902844468752543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,8192,0.07603555255466037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,32,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,1536,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,5120,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,16384,0.02027555637889438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,51200,0.24301687876383463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,1024,0.05156888895564609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,1536,0.003912888881233004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,5120,0.022988445229000513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,3072,0.012626666989591388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,2048,0.010669333239396414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,51200,0.3631617758009169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,7168,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,12288,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,1536,0.08716533581415813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,4096,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,64,0.005312888986534542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,8192,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,3584,0.009472889204819998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,32,0.04497333367665609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,2560,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,1536,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,5120,0.016685333516862657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,12288,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,2048,0.024718221690919664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,768,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,12288,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,10240,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,5120,0.1778764459821913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,10240,0.019330667124854196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,512,0.006856888946559694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,256,0.04455555478731791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,8192,0.16225332683987087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,51200,0.31249867545233834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,2048,0.06314222017923991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,12288,0.015274667077594332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,768,0.018283555905024212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,4096,0.022224000758594934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,10240,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,1024,0.0038373333712418876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,65536,1.098656866285536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,256,0.02734311090575324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,2048,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,5120,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,2560,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,64,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,5120,0.044475555419921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,128,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,12288,0.10018577840593125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,4096,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,12288,0.021194666624069214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,512,0.0510062211089664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,8192,0.03945599993069967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,32,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,5120,0.009299555586444007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,2560,0.008569777839713627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,1024,0.013825777504179211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,65536,0.9695030848185221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,1024,0.03851022322972616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,5120,0.044907556639777295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,256,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,4096,0.024089778463045757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,128,0.12056977219051784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,2048,0.06280355321036445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,8192,3.238527086046007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,4096,0.18551466200086805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,512,0.01495911180973053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,3072,0.022964444425370958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,4096,0.03046400017208523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,128,0.12433421611785889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,6144,0.07573333051469591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,16384,0.059416890144348145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,128,0.006206222292449739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,4096,0.03934222128656175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,12288,0.09076444307963054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,1536,0.01707466608948178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,2048,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,65536,0.24278044700622559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,3584,0.039107554488711886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,128,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,1024,0.06168533033794827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,2560,0.017198221551047433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,8192,0.03276888860596551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,7168,0.15660533640119764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,3072,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,2560,0.08083466688791911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,4096,0.04225066635343763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,3072,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,3072,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,5120,0.0203813330994712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,3072,0.007707555260923173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,6144,0.031178666485680476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,4096,0.026937777797381084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,768,0.020607999629444547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,10240,0.6306808789571127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,10240,0.04595288965437147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,4096,0.009484444227483537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,8192,0.03130577670203315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,5120,0.009138666921191746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,7168,0.09238311317231919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,3584,8.817217508951822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,10240,0.03388355506791009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,4096,0.07149244679345025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,1536,0.44492446051703555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,128,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,768,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,5120,0.026428444517983332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,6144,0.04977955420811971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,1024,0.02779200010829502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,768,0.004192000048028098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,256,0.002983111060327954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,65536,0.32353332307603627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,8192,0.3780488967895508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,3584,0.0325546662012736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,32,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,7168,0.2985706594255236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,256,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,1024,0.003519999898142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,768,0.016582222448454965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,256,0.07373778025309245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,64,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,3072,0.018068444397714403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,128,0.01347555551264021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,2560,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,2048,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,65536,0.1642168892754449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,2048,0.05708533525466919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,128,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,12288,0.1090008881356981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,768,0.018160889546076458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,65536,0.08266844352086385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,12288,0.03167288833194309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,7168,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,12288,1.2008248435126412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,32,0.15412178304460314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,2048,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,4096,0.08015111419889662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,7168,0.029847111966874864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,64,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,3584,0.03523111012246873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,2048,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,2560,0.10536266697777642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,8192,0.01107022249036365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,4096,0.03384177883466085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,2560,0.007162666983074612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,2560,0.028873778051800195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,7168,0.023407111565272015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,2048,0.006395555618736479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,3072,0.013079110946920184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,16384,0.27356089486016166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,65536,0.057709336280822754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,2560,0.014576888746685453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,128,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,7168,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,5120,0.019641778535313077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,1024,0.019906666543748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,7168,0.12667200300428602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,3584,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,65536,0.09798133373260498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,51200,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,5120,0.012520000338554382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,5120,0.053871999184290566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,768,0.00849422232972251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,5120,0.02460177739461263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,2048,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,7168,0.07308799690670438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,2048,0.05644977755016751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,512,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,16384,0.2402391168806288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,10240,0.031502223677105375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,5120,0.08792799711227417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,256,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,4096,0.00508266646001074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,2048,0.4098186757829454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,65536,0.5563839806450738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,16384,0.7063537703620063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,6144,0.03369955552948846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,10240,0.2183955510457357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,128,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,10240,0.16615022553337946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,51200,0.07076978021197848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,5120,0.18317066298590767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,2048,0.014273777604103088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,16384,0.03942222065395779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,1536,0.027588443623648748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,5120,0.0832773314581977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,12288,0.023723555935753718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,128,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,6144,0.136790222591824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,6144,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,2048,0.6216950946384007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,768,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,768,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,10240,0.35350667105780703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,51200,0.0711946686108907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,2560,0.48705599043104386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,5120,0.019985778464211356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,1024,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,3584,0.021518222159809534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,12288,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,5120,0.031423999203575984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,5120,0.12257511085934109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,6144,0.019150222341219585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,3072,0.045067555374569364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,2048,0.12132355901930068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,12288,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,8192,0.14911911222669813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,512,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,51200,0.04877600073814392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,1024,0.015230221880806817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,4096,0.11867200003729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,768,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,65536,0.17962133884429932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,51200,0.3117226759592692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,4096,0.12428800264994304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,7168,0.013975999421543546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,51200,0.23183022605048287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,2048,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,65536,0.05595466825697157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,1536,0.07607111003663805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,51200,0.10325600041283502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,12288,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,8192,0.011944000091817645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,6144,0.011208888557222156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,5120,0.07879466480678983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,768,0.006325333482689328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,1024,0.016530666086408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,65536,0.054921779367658824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,768,0.021589333812395733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,7168,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,1536,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,65536,0.09143733315997654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,3584,0.585444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,6144,0.06294133265813191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,51200,0.1114657719930013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,2048,0.02223199937078688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,51200,0.10849955346849229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,128,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,10240,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,3072,0.013316444224781461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,10240,0.04527555571662056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,256,0.015543111496501498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,10240,0.013453333742088742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,2560,0.0796639985508389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,3584,0.04114933477507697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,7168,0.1228133307562934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,7168,0.04001333316167196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,8192,0.10692622264226277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,5120,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,3584,0.35268621974521214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,3072,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,256,0.016554666890038382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,768,0.021712000171343487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,128,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,12288,0.032567110326555036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,768,0.03252266512976752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,10240,0.013521778086821238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,3584,0.0461351109875573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,3584,0.007928888830873702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,2560,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,2048,0.03423733181423611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,2048,0.08425422509511311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,4096,0.018732445107565988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,6144,0.03576177689764235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,6144,0.02330488959948222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,2048,0.012395555774370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,1536,0.004416888786686791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,5120,0.09380000167422825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,6144,0.01650311052799225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,768,0.013760888742076026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,12288,0.05380710959434509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,2048,0.07175822390450372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,6144,0.011209777659840055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,768,0.022612444228596155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,256,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,512,0.10322933064566718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,10240,0.07188800308439466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,256,0.008244444098737504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,64,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,65536,0.07240000036027697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,2048,0.09638488954967922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,128,0.4177404509650336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,7168,0.030384000804689195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,10240,0.21931200557284883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,128,0.0063164445261160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,64,0.008586666650242275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,5120,0.00871377769443724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,128,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,12288,0.06411911381615533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,8192,0.02481066683928172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,5120,0.03207733233769735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,2560,0.1221546729405721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,3584,0.1006951133410136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,1536,0.0700311130947537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,1024,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,7168,0.044665777020984225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,2048,0.010816888676749336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,3072,0.016144000821643405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,4096,0.01884533299340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,65536,0.0800666676627265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,2560,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,1024,0.009292444421185387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,64,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,65536,0.1258995532989502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,3584,0.1800044510099623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,3072,0.343215995364719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,1024,0.011980444192886353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,5120,0.11005155245463054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,5120,0.04545688960287306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,8192,0.044176889790429004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,4096,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,128,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,64,0.016166221764352586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,128,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,7168,0.019931554794311523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,10240,0.08543377452426487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,8192,0.25204534000820583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,65536,0.05063555637995402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,12288,0.037440889411502414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,10240,0.0192284451590644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,2560,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,256,0.005869333528810077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,16384,0.02309333284695943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,51200,0.32261421945359975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,2560,0.01646222174167633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,2560,0.08520711130566067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,8192,0.10869155327479045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,32,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,4096,0.16157244311438665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,2048,0.047669334544075855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,51200,0.0630542238553365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,8192,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,768,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,2048,0.01109688894616233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,768,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,16384,0.5899510913425022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,10240,0.16243022018008763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,32,0.004507555729813046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,7168,0.1782968839009603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,5120,0.008600888980759515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,51200,0.04955733484692044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,32,0.03520977828237746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,6144,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,12288,0.1336017714606391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,12288,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,2560,0.021224000387721594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,3584,0.02398666739463806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,8192,0.07215555508931477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,7168,0.14798488881852892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,1536,0.03672622309790717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,12288,0.22363021638658312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,7168,0.1398622194925944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,65536,0.06110666857825386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,512,0.04722133278846741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,3072,0.016979555288950603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,3584,0.11594311396280925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,3072,0.06629244486490886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,1024,0.029304888513353135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,2048,0.05505511164665222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,10240,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,65536,0.2269359959496392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,512,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,16384,0.01568266749382019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,128,0.019557333654827543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,1536,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,4096,0.027126222848892212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,51200,3.1371822357177734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,16384,0.0879288911819458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,51200,0.1411448849572076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,10240,0.10788444677988689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,12288,0.023131555981106226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,32,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,16384,0.03501244386037191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,16384,0.028091556496090356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,16384,0.018423999349276226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,64,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,768,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,7168,0.02847466535038418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,2048,0.021588444709777832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,10240,0.03792444533771939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,10240,0.03905155592494541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,10240,0.05310577816433377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,256,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,768,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,12288,0.02862666712866889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,8192,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,2048,0.004362666772471534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,12288,0.7723191049363879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,128,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,768,0.011335111326641507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,4096,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,6144,0.165793776512146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,1536,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,51200,3.0644124348958335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,1536,0.008231110870838165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,256,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,5120,0.030050665140151978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,6144,0.1274053388171726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,2048,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,32,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,2048,0.010452444354693094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,768,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,3584,0.07528711027569242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,1536,0.004311111238267687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,10240,1.4456293318006728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,4096,0.3777137862311469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,10240,0.018603555030292936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,768,0.01752000053723653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,1024,0.01422755585776435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,256,0.030960000223583643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,3072,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,3072,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,5120,0.012220444778601328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,1536,0.013674666484196981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,10240,0.023507555325826008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,12288,0.07585599687364367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,768,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,65536,0.08179377847247653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,8192,0.09635644488864475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,65536,0.15629066361321345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,4096,0.019564444820086162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,1024,0.018930666976504855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,10240,0.021743110484547083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,256,0.018536888890796237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,7168,0.5465448697408041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,5120,0.11441511578030056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,4096,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,6144,0.09968444373872544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,8192,2.0403058793809676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,12288,0.9988258149888781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,2560,0.046487109528647534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,8192,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,8192,0.0251528885629442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,8192,0.028060442871517603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,512,0.013422222601042854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,2560,0.026921777261628047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,3072,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,768,0.0025537777692079544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,2560,0.006457777900828256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,768,0.006049777898523543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,128,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,3584,0.10453599691390991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,512,0.01443288889196184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,8192,0.016902221573723685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,65536,4.917778862847222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,1024,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,7168,0.03089155422316657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,768,0.01408088869518704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,65536,0.06435377730263604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,16384,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,10240,0.12472444110446507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,4096,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,65536,19.28677707248264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,16384,0.0478115545378791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,8192,0.7032133208380805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,7168,0.0640159977806939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,128,0.0850328869289822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,768,0.018595554762416415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,5120,0.016946666770511203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,12288,0.4306951098971897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,8192,0.15922666920555964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,3072,0.14343021975623235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,7168,0.017794667018784415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,3584,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,32,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,8192,1.4245778189765081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,1536,0.011608888705571493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,3584,0.027152000202072993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,6144,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,2560,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,768,0.006088888893524806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,128,0.008369777765538957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,2048,0.025387555360794067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,3584,0.023415111833148535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,1536,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,1024,0.012899555265903473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,3584,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,512,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,2560,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,3584,0.0206675562593672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,12288,0.017864000466134813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,12288,0.06952533457014295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,64,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,16384,0.2836959891849094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,16384,0.1032079988055759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,2560,0.23033422893948027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,2048,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,64,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,10240,0.3319386641184489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,768,0.027621333797772724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,51200,3.885064019097222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,3584,0.0838693314128452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,64,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,10240,0.1429848935869005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,5120,0.014293332894643148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,2560,0.004899555610285865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,10240,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,3584,0.013148444394270578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,128,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,2048,0.04770222306251526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,5120,0.012096000214417776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,51200,0.15674489074283177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,512,0.04073777794837952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,2560,0.007092444433106317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,12288,0.014750222365061441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,128,0.0716444452603658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,4096,0.14489955372280544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,51200,0.07066488928265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,256,0.006896889044178857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,10240,0.06296800242529975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,128,0.01682577696111467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,768,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,128,0.003478222423129611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,32,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,51200,1.386856926812066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,512,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,64,0.0017786667578750187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,2048,0.00739644467830658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,12288,0.032582223415374756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,256,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,3584,0.41552087995741105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,6144,0.017616889543003507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,128,0.0247013337082333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,1024,0.019906666543748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,64,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,1536,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,3584,0.17320888572269014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,3584,0.07298755645751953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,1536,0.01274400038851632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,768,0.026143999563323125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,3584,0.016150222884284127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,65536,0.2652782334221734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,4096,0.004579555657174853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,6144,0.04868355393409729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,768,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,65536,0.1979431046379937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,1536,0.003724444243643019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,10240,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,2048,0.02382044494152069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,3072,0.7595440016852485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,65536,0.15720533000098333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,7168,1.5330382453070746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,10240,0.08263822396596272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,2048,0.01939377850956387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,51200,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,16384,3.908686319986979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,16384,0.4644808769226074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,51200,0.027709333433045283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,3584,0.01871200071440803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,2048,0.01331733332739936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,2048,0.027962668074501887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,16384,0.016530666086408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,12288,0.11106399695078532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,4096,0.024724443753560383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,512,0.0041857775714662345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,768,0.005406222409672207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,1024,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,768,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,51200,0.0991342200173272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,512,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,3072,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,5120,0.017224889662530687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,51200,0.1520355542500814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,512,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,16384,0.13937866687774658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,6144,0.02331466641690996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,51200,1.5616391499837239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,2048,0.11303733454810248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,3584,0.1491288873884413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,7168,0.01462311049302419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,64,0.015282667345470853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,5120,0.03259733319282532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,512,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,5120,0.02467022173934513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,2560,0.02955199943648444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,51200,0.6072630882263184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,768,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,6144,0.02681688964366913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,512,0.022431110342343647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,32,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,2048,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,51200,4.21034410264757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,128,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,64,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,1024,0.01555022266176012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,6144,0.07615111271540324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,4096,0.011355555719799466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,3584,0.029973334736294214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,3584,0.015063110325071545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,8192,0.24229865603976777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,16384,0.02873599860403273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,3072,0.03729866610633002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,64,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,3584,0.007737778127193451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,256,0.16671821806165907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,3072,0.027414222558339436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,3584,0.06106311082839966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,768,0.012117333710193634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,8192,0.09010310967763265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,8192,0.06542844242519803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,1024,0.16351377964019775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,256,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,256,0.0112631113992797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,1536,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,512,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,5120,0.05166044500139025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,6144,0.0063368889192740125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,128,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,128,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,10240,0.7639830907185873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,256,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,5120,0.09586311048931545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,3072,0.05070133341683281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,3584,0.008360000120268928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,3072,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,12288,0.27888533804151744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,3584,0.009110222260157267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,65536,0.1463706625832452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,768,0.0042151109211974675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,7168,0.2556106779310438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,51200,1.4768115149603949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,6144,0.9847013685438367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,2560,0.06597066587871976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,768,0.003444444388151169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,10240,0.16464800304836696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,2560,0.0086986662613021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,3584,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,16384,0.03648622168434991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,6144,0.03614933292071024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,64,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,32,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,6144,0.01178311142656538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,256,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,3584,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,12288,0.01720266706413693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,256,0.01351377781894472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,16384,2.2924062940809464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,2560,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,2560,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,65536,0.16129067209031847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,12288,0.14501244492001003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,1024,0.006400888992680444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,512,0.0025884444928831523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,256,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,4096,0.030456887351142034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,65536,0.41354044278462726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,65536,0.37733154826694065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,1536,0.045597334702809654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,8192,0.037208000818888344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,16384,0.022387555903858606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,2560,0.02530133393075731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,12288,0.23635376824273002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,512,0.036715557177861534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,2048,0.009920000202125972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,768,0.0061048890153567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,1536,0.14607111612955728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,2560,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,3584,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,3584,0.014745778507656522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,2048,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,3072,0.022485334012243483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,2048,0.014537778165605334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,2048,0.0789546635415819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,5120,6.649306827121311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,6144,0.019488889310095046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,1024,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,10240,0.09274133046468098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,512,0.0026871110830042097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,768,0.00774133370982276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,10240,0.058232890235053174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,7168,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,64,0.04597244328922696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,51200,0.3549359904395209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,3584,0.007687110867765214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,3584,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,4096,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,4096,0.06288711229960124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,4096,0.010833777487277985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,7168,0.03624533282385932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,512,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,6144,0.06822311215930514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,1024,0.04890666736496819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,16384,0.020945777495702107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,128,0.020592000749376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,1024,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,8192,2.7962773640950522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,16384,0.20833155843946669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,1536,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,128,0.015451555450757345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,4096,0.007695111135641734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,7168,0.0357893341117435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,512,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,6144,0.23469421598646376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,64,0.08267822530534533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,128,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,4096,0.012717333104875354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,256,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,51200,0.08232533269458346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,1024,0.013772444592581855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,65536,0.2551608880360921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,128,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,51200,0.02132355504565769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,2560,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,768,0.01791555517249637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,12288,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,16384,0.2425626648796929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,12288,0.07127733363045587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,1536,0.004585777719815572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,3072,0.02385866641998291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,10240,0.06836799780527751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,16384,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,12288,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,5120,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,51200,3.2632702721489797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,512,0.047671112749311656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,8192,0.02296977738539378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,6144,0.6899946530659994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,32,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,8192,0.014319111903508505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,768,0.03375466664632162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,1536,0.16912710666656494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,256,0.010293333066834344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,5120,0.023756444454193115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,5120,0.034500443273120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,256,0.017834666702482436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,16384,0.03347288899951511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,2560,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,512,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,12288,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,10240,0.08391555812623765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,12288,0.305440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,3584,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,12288,0.029901332325405542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,3072,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,2560,0.022687999738587275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,512,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,1024,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,10240,0.20309155517154268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,3072,0.04889066682921516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,8192,0.11628177430894639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,51200,0.07035733593834771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,768,0.06514044602711995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,51200,0.2139457729127672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,512,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,51200,0.08826044532987808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,1024,0.0029191111938820947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,1024,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,1024,0.05281066563394335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,7168,0.5487066904703776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,3584,0.016917333006858826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,16384,0.09460800223880345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,10240,0.1485146681467692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,128,0.013047110703256396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,2560,0.034840001000298396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,2048,0.015782222151756287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,8192,0.011798222031858234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,65536,0.0925546685854594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,1536,0.0236471113231447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,1536,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,768,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,512,0.0029164445069101122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,3584,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,65536,0.015777778294351365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,2048,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,8192,0.09666044182247585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,2048,0.024546666277779475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,4096,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,512,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,512,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,65536,0.08290400107701619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,65536,0.10673955414030288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,7168,0.03922577699025472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,16384,0.7535297605726453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,3584,0.03734933336575826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,768,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,7168,0.055736886130438916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,10240,0.017838221457269456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,768,0.00612088872326745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,768,0.011913778053389655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,768,0.018529777725537617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,768,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,768,0.00309244439833694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,3072,0.6655422316657172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,768,0.027020444472630818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,6144,0.016001777516471017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,65536,19.38485378689236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,6144,0.7666862275865344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,6144,0.028145776854621038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,64,0.002537777854336633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,5120,0.2758764425913493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,8192,0.07037688626183404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,2560,0.019326221611764696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,8192,0.11775822109646267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,64,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,1024,1.1462311214870877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,1536,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,3584,0.0632293356789483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,16384,0.08788977728949653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,2560,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,512,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,2048,0.7035706837972006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,6144,0.046225776274998985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,1536,0.02273599969016181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,1536,0.012118221984969245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,1024,0.06303999821345012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,8192,0.06782488690482245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,3584,0.028135110934575398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,256,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,1024,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,512,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,768,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,32,0.0053680001033677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,16384,0.015802666544914246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,65536,0.19830932882097033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,6144,0.060681780179341636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,2560,0.012563555604881711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,3584,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,256,0.033077332708570696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,5120,0.008453333543406593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,6144,0.03904977771970961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,51200,0.1296959982977973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,2560,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,16384,12288,0.04946755700641208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,12288,0.13423911730448404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,512,0.022219555245505437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,512,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,2560,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,128,0.016144000821643405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,65536,0.06324978007210626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,256,0.005647999958859549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,8192,0.3115333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,2048,0.04163466559516059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,768,0.020473778247833252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,768,0.01959377858373854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,16384,0.14324177636040583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,3584,0.049161778555976025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,2560,0.011491555306646558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,7168,0.23323199484083387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,12288,0.05487999982304043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,51200,55.81580268012153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,16384,0.5579991340637207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,5120,0.03332355618476868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,5120,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,65536,0.14898044533199734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,10240,0.10118310981326634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,2560,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,6144,0.6669155756632487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,2560,0.060039109653896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,5120,0.049441777997546725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,4096,0.2004879977968004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,2560,0.05301244391335381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,5120,0.03071111109521654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,1024,0.004726222405831019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,2048,0.07544800308015612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,512,0.04591644472546048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,256,0.02010577751530541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,1024,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,512,0.004572444491916233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,256,0.031934221585591636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,51200,0.3747848934597439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,256,0.007365333537260692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,2048,0.06070578098297119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,768,0.015172445111804538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,8192,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,768,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,1024,0.01278133359220293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,128,0.03702044486999512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,64,0.011600889265537262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,1536,0.10070844491322835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,4096,0.022035555707083807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,1536,0.015504888362354703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,2560,0.021920889616012573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,51200,2.4253173404269748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,12288,0.019320888651741874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,7168,0.03535733289188809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,8192,0.03596444593535529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,128,0.010061333576838175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,16384,0.10390488968955146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,6144,0.021526222427686054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,10240,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,512,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,768,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,6144,0.010405333505736457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,2048,0.010755555497275459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,2560,0.016232000456915963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,12288,0.017864889568752713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,2560,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,512,0.037639111280441284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,10240,0.18108622233072916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,16384,1.0312897364298503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,7168,0.020642666353119742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,16384,0.06970044639375475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,64,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,10240,0.020961778031455144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,256,0.07138222455978394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,2048,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,2560,0.05316088928116692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,5120,0.03548977772394816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,8192,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,10240,0.025683555338117812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,768,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,3072,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,65536,0.1942737764782376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,512,0.003208000005947219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,2560,0.045088890526029796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,1536,0.0047564444442590075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,64,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,12288,2.186643600463867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,3584,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,2048,0.04582488867971632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,6144,0.2336195574866401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,10240,0.028766221470303003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,512,0.025059555967648823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,256,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,1024,0.04572088850869072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,8192,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,8192,0.13591910733117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,6144,0.3839724328782823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,256,0.020505777663654752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,2048,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,16384,0.029851555824279785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,2560,0.012233778006500669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,7168,0.014727999766667684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,3072,0.028175112273957994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,2560,0.03304622239536709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,6144,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,7168,0.18440177705552843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,768,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,1536,0.4716862042744954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,3072,0.026704000102149114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,1024,0.029112888707054987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,4096,0.0617120001051161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,1536,0.07423022058275011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,512,0.06558755371305677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,2560,0.05450488792525398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,32,0.0047360000511010485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,7168,0.020266667008399963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,1024,0.010328888893127441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,512,0.05666666560702854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,51200,0.1296373340818617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,128,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,64,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,3072,0.007146666447321574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,1536,0.029448890023761328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,7168,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,1024,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,3584,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,2560,0.03312799996799893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,65536,0.05322844452328152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,2560,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,1536,0.009688888986905416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,2560,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,10240,0.08580622408125137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,5120,0.03802311089303758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,16384,0.09089155329598321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,2048,0.016756445169448853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,8192,0.02349244389269087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,7168,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,10240,0.3528275489807129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,128,0.003907555507289039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,1536,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,768,0.03275200062327915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,1536,0.033260444800059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,3072,0.03866666555404663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,2560,0.029911110798517864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,3072,0.020937777227825586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,4096,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,16384,0.025324443976084392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,3072,0.013039111263222165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,1024,0.02819733487235175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,1536,0.19181956185234916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,8192,0.04941777719391716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,4096,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,768,0.008944888909657797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,512,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,5120,0.1553973356882731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,2048,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,16384,0.048447112242380776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,5120,0.02796622117360433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,8192,0.011448889142937131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,3072,0.02102222210831112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,7168,0.06610400146908231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,10240,0.0880897773636712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,1536,0.0169840008020401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,12288,0.02773155603143904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,4096,0.0407591097884708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,2048,0.0314631097846561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,2560,0.008715555899673039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,65536,0.22431021266513398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,2560,0.034625778595606484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,2048,0.04819200105137295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,5120,0.3014720016055637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,5120,0.049069331751929395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,16384,0.030941334035661485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,10240,0.45643022325303817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,51200,0.1266497770945231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,64,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,10240,0.031141334109836157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,64,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,1024,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,3584,0.053767111566331655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,5120,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,1536,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,3584,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,51200,0.0969911085234748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,7168,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,1536,0.12967022260030112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,51200,0.2723555564880371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,128,0.015255111787054272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,768,0.05762400229771932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,10240,0.028591109646691218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,2048,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,2048,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,256,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,16384,0.07797511418660481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,512,0.06546844376458062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,2048,0.028284443749321833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,16384,0.08555644750595093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,256,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,12288,2.0697405073377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,768,0.02142488956451416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,3072,0.2334364520178901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,768,0.043828444348441235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,2048,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,2048,0.0428071121374766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,3072,0.4183519946204291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,6144,0.027081777652104694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,16384,0.03241066800223456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,64,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,1536,0.01869244376818339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,7168,0.0750453339682685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,2560,0.04545422063933479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,3072,0.011208000282446543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,65536,0.3189057774013943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,10240,0.12610577212439641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,128,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,1024,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,16384,0.3520195484161377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,2048,0.024030221833123103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,6144,0.009755555954244401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,1536,0.011110222174061669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,128,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,10240,0.0931733316845364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,51200,0.5623431205749512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,128,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,3584,0.015040000279744467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,16384,0.41865865389506024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,3584,0.18414399358961317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,256,0.0038764443662431505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,8192,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,3584,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,5120,0.011424889167149862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,1536,0.011946666571829053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,10240,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,2560,0.019550222489568923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,768,0.02882844540807936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,6144,0.03808444407251146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,2560,0.018819555640220642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,16384,0.03934222128656175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,16384,0.026889777845806543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,10240,0.01716088917520311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,2048,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,10240,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,3072,0.007430222299363878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,2560,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,8192,0.140791998969184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,51200,0.20123467180464003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,32,0.0017768889665603638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,7168,0.027087999714745417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,1536,0.012142221960756512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,10240,0.0075039996041191956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,4096,0.02032711108525594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,6144,0.021306667062971327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,768,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,1536,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,3584,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,65536,1.1401991314358182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,1024,0.02609777781698439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,16384,0.026679111851586237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,1536,0.03650666607750787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,12288,0.043600890371534556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,3584,0.09860889116923015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,32,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,512,0.015871110889646742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,3584,0.028590222199757893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,128,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,3072,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,6144,0.012897777888509961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,1536,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,1024,0.012084444363911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,12288,0.02538577715555827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,8192,0.48170312245686847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,51200,0.05526133378346761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,8192,0.012895111408498553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,4096,0.041053334871927895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,51200,0.6899902025858561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,12288,0.020246222615242004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,12288,0.011023999916182624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,1024,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,512,0.015285332997639975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,2048,0.015085332923465304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,32,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,16384,0.016494222813182406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,12288,0.05427111188570658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,256,0.02530844509601593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,3584,0.007703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,2048,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,65536,0.06772444645563762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,7168,0.1252408954832289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,1536,0.016772444049517315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,768,0.2270533243815104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,2048,0.04144622219933404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,5120,0.19518400563134086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,3584,0.053744001521004565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,12288,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,256,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,128,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,7168,0.0432204438580407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,6144,0.025518221987618342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,51200,0.06408622529771593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,4096,0.04458666510052151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,6144,0.018183999591403537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,4096,0.02552355494764116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,4096,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,12288,0.6445688671535915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,256,0.0058577776783042485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,1536,0.020909332566791113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,3584,0.020522667302025687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,32,0.007831110722488826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,5120,0.02049511174360911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,5120,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,3072,0.03124799993303087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,3584,0.022575111852751836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,5120,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,12288,0.014426666829321118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,3072,0.009475555684831407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,32,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,5120,1.9169805314805772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,12288,0.16552000575595432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,8192,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,2560,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,64,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,256,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,256,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,5120,0.011799111134476133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,1536,0.031125333574083116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,8192,0.0174346665541331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,128,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,8192,0.016784888174798753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,1536,0.006010666489601135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,3584,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,51200,0.06357510884602864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,65536,0.07181333170996772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,65536,1.5049787097507055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,1024,0.03138577938079834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,5120,0.10667733351389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,4096,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,2048,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,128,0.010446222292052375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,3072,0.01055466632048289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,2560,0.027836445305082533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,512,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,16384,0.0317466656366984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,1536,0.013757333159446716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,6144,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,3072,0.05045688814587063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,3584,0.046385778321160205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,8192,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,2560,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,1024,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,5120,0.02642222245534261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,3584,0.015159999330838522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,512,0.01187111106183794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,16384,0.04283200038803948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,4096,0.08248444398244222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,768,0.0032186667538351486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,4096,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,1536,0.042824887567096286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,768,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,5120,0.03755555550257365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,7168,0.012718222207493253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,128,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,65536,3.3673352135552297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,2048,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,5120,0.025828444295459326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,3584,0.013612444202105204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,6144,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,64,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,1536,0.012540444731712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,3584,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,3072,0.04982577761014303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,6144,0.14960622787475586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,7168,0.0442808899614546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,128,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,10240,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,12288,0.013789333403110504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,10240,0.04961511161592272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,1024,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,2560,0.015590222345458137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,1536,0.019680889116393197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,2048,0.00920800036854214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,5120,0.012050666742854647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,65536,0.8459617826673719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,768,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,51200,0.022487110561794702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,65536,0.13678844769795737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,5120,0.1930746634801229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,7168,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,768,0.006663110935025745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,51200,0.061103112167782254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,4096,0.008606221940782335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,5120,0.43343732092115617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,51200,0.13666489389207628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,12288,0.05390222205056084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,7168,0.12335644827948676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,12288,0.7546382480197482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,65536,12.240211486816406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,16384,0.02509333358870612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,3584,0.00792266676823298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,65536,0.01917155583699544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,2560,0.23466934098137748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,2048,0.00740533322095871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,2560,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,65536,0.030066665675905015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,6144,0.14758222632937962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,51200,0.12247555785708958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,16384,0.022322666313913133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,64,0.020796444680955675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,51200,0.03994489047262404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,1024,0.006015111174848344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,8192,0.0921599997414483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,12288,0.048581331968307495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,3584,0.009394666386975182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,65536,0.12980622715420193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,256,0.007676444119877285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,65536,0.30768267313639325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,12288,0.02571733295917511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,3072,0.014373333917723762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,51200,0.43010844124688047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,12288,0.0651457773314582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,3072,0.03319911162058512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,256,0.021029333273569744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,16384,0.017102221647898357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,12288,0.02551733288500044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,32,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,7168,0.0228426671690411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,3584,0.01367111090156767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,8192,0.012607110871209038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,1536,0.053505778312683105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,3072,0.02604977786540985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,7168,0.024548444483015273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,4096,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,4096,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,2560,0.0787573324309455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,5120,0.03696533375316196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,1024,0.01755644381046295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,65536,0.06798666715621948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,3072,0.21760088867611357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,2560,0.004856889032655292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,1536,0.14442488882276747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,2560,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,51200,0.059346669250064425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,1024,0.003175111073586676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,12288,0.2825555536482069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,7168,0.02649066680007511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,1536,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,12288,0.48000086678398984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,1536,0.011757333245542316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,256,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,256,0.011176000038782755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,4096,0.02124533388349745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,512,0.11716622776455349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,6144,0.02000799940692054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,2560,0.4086586634318034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,1024,0.009524444739023844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,3072,0.10740444395277236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,3072,0.28158667352464467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,256,0.025126222107145522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,1024,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,1536,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,4096,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,5120,0.0143857780430052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,3584,0.21333866649203828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,12288,0.3048551082611084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,32,0.027446221974160936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,51200,0.42224799262152773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,51200,0.21693866782718232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,10240,0.26705778969658744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,64,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,51200,0.0882800022761027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,51200,0.06013955672581991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,2048,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,3584,0.035473777188195124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,1024,0.03440444336997138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,4096,0.0746426648563809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,2048,0.009663110805882348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,8192,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,6144,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,1024,0.019288889235920377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,2048,0.016406221522225272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,256,0.019713777634832595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,3072,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,512,0.0036551112102137674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,3072,0.020510221521059673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,256,0.002206222257680363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,1024,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,6144,0.019261333677503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,2560,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,12288,0.061273780133989125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,1024,0.037054220835367836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,3584,0.033999999364217125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,65536,0.25201333893669975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,256,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,2048,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,4096,0.14225155777401394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,8192,0.17907110850016275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,1024,0.006752000086837345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,12288,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,16384,0.11442044046190049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,512,0.010946667028797997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,32,0.00240355564488305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,5120,1.5038151211208766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,64,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,3584,0.01793066660563151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,6144,0.018280888597170513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,5120,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,10240,0.8875102467007108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,3584,0.030633777379989624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,5120,0.03763377666473389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,32,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,3072,0.041821334097120494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,6144,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,64,0.01424533294306861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,51200,0.05747377872467041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,3584,0.10828266541163127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,3072,0.0547653337319692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,768,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,7168,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,256,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,1536,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,128,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,256,0.01679733395576477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,51200,0.09648178021113078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,51200,1.2676622602674696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,7168,0.01685066686736213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,3072,0.016679111454221938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,5120,0.04688622223006355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,51200,1.152406268649631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,5120,0.25112711058722603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,32,0.003607999947335985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,3072,0.008164444731341468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,51200,0.059608002503712974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,2048,0.11456444528367783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,1536,0.014183999763594734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,768,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,5120,0.00906488878859414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,1024,0.003354666754603386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,7168,0.07058222426308526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,5120,0.08597066667344834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,1536,0.006138666636413998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,16384,0.021359999974568684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,3072,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,256,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,32,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,32,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,1536,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,768,0.19643378257751465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,1536,0.028259555498758953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,512,0.0052968888647026485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,32,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,6144,0.018394667241308425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,512,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,768,0.10222488641738892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,3584,0.010083555347389644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,64,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,5120,0.10416889190673828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,8192,0.16100977526770696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,10240,0.2031395567788018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,8192,0.23155999183654785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,2560,0.08992355399661595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,3072,0.0075502221783002215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,1536,0.025648888614442613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,51200,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,4096,0.010399111443095738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,51200,0.18516711393992105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,1536,0.009132444030708736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,3072,0.012608889076444836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,16384,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,4096,0.016880000631014507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,256,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,128,0.017855111095640395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,12288,0.026706667410002813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,7168,0.009940444595283931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,16384,0.7275582419501411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,2048,0.0069955554273393415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,3584,0.04191733400026957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,128,0.010762666662534079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,3584,0.01665599975321028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,1536,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,51200,0.10520533058378433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,7168,0.07944977945751615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,3584,0.007059555914666917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,3072,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,3072,0.10907910929785834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,5120,0.5829626719156901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,65536,0.08697155449125503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,5120,0.15061422189076742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,8192,0.20063289006551108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,8192,0.06542666753133138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,2560,0.1027066641383701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,2048,0.012411555482281579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,1536,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,768,0.0031111112071408164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,3072,0.007576000359323289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,4096,0.05581955777274238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,6144,0.015065777632925244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,10240,0.02828177809715271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,1536,0.031876444816589355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,5120,0.060192002190483936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,256,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,512,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,3584,0.5022835731506348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,64,0.001791999985774358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,2048,0.37509245342678493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,768,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,2048,0.09561866521835327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,8192,0.2405102252960205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,768,0.04132533404562209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,64,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,2048,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,51200,0.1522293355729845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,2560,0.006562666760550604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,3584,0.04973599976963467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,32,0.0053191110491752625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,256,0.0029315555261241067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,6144,0.028712001111772325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,768,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,2048,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,1536,0.007320000065697565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,1024,0.015419556034935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,7168,0.027512888113657635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,64,0.021611554755104914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,2560,0.012472000386979846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,8192,0.09718755880991618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,51200,0.1697831153869629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,3584,0.021550221575631037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,4096,0.011759999725553723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,2560,0.009335111412737105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,16384,0.25693954361809623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,2560,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,3584,0.07131733496983846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,16384,0.1428204509947035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,8192,0.02257866660753886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,1536,0.02404622236887614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,3072,0.04246844516860115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,16384,0.020949333906173706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,10240,0.032427555984920926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,4096,0.02696177860101064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,2560,0.04767910970581902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,256,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,7168,0.026524444421132404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,2560,0.19615022341410318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,32,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,768,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,512,0.0025920000755124623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,2560,0.015556444724400839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,1024,0.008015111088752747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,512,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,8192,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,128,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,3072,0.06936711072921753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,6144,0.025829333398077223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,256,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,16384,0.05307644605636597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,512,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,1024,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,1536,0.019462222854296368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,51200,0.13652443885803223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,51200,9.441980150010851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,6144,0.01924888955222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,128,0.0255395554833942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,65536,0.0690915584564209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,256,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,256,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,6144,0.055488000313440956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,12288,0.07034666670693292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,256,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,4096,0.029446221060223047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,768,0.0181013329161538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,7168,0.08203111092249553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,1536,0.010479999913109673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,2048,0.05512711074617174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,256,0.054995556672414146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,16384,0.5321102142333984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,512,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,4096,0.01399466726515028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,51200,0.05691288577185737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,768,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,7168,0.017481777403089736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,12288,0.05198755529191759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,3072,0.020190222395790946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,65536,0.12652977307637533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,512,0.0409840014245775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,1536,0.017837334010336135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,1536,0.05074577861362033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,10240,0.06489600075615777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,6144,0.14984888500637478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,128,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,2048,0.009345778160625035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,12288,0.14037688573201498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,1536,0.014013333453072442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,1024,0.019738665885395475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,65536,0.07037778033150567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,512,0.11531378163231744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,1536,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,2560,0.00462755560874939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,16384,0.025975111458036635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,51200,3.8968255784776478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,1024,0.0517884459760454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,3584,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,8192,0.06855199734369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,10240,0.030193779203626845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,256,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,128,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,1536,0.1066746645503574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,128,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,5120,0.022534221410751343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,3584,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,12288,0.02535733415020837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,65536,0.4817182223002116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,8192,0.010838222172525195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,3584,0.02437066700723436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,65536,0.0759377810690138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,4096,0.041096889310412936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,8192,0.04106577899720933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,5120,0.03798666596412659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,1536,0.004628444297446145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,2048,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,7168,0.022652443912294176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,7168,0.02808977829085456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,2048,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,3584,0.23136178652445474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,5120,0.040814220905303955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,2048,0.021700445148679946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,256,0.016494222813182406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,7168,0.04703199863433838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,8192,0.011589333415031433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,1024,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,1536,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,3584,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,5120,0.0712053312195672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,32,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,12288,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,12288,0.06401777929729886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,51200,0.4051511022779677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,512,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,3072,0.0729457802242703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,64,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,8192,0.029623111089070637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,2560,0.034161776304244995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,2560,0.012700444294346703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,5120,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,2048,0.051338666015201144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,4096,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,2560,0.04141866829660203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,256,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,2560,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,2048,0.028532445430755615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,12288,0.06358222166697185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,3584,0.1281777752770318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,16384,0.01648533344268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,128,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,256,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,512,0.013639999760521783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,65536,0.06450577576955159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,65536,0.14847644170125326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,1536,0.015650666422314115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,3072,0.0281315545241038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,768,0.015650666422314115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,256,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,8192,0.08134133285946317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,7168,0.027307555079460144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,4096,0.012606222596433429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,768,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,2560,0.01756088932355245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,5120,0.06321066617965698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,65536,0.10112177663379246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,8192,0.17537600464291045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,10240,0.28621066941155326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,2048,0.008540444076061249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,128,0.3433413240644667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,5120,0.011712888876597086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,2560,0.015354666444990369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,768,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,7168,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,32,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,12288,0.032884445455339216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,12288,0.01937244501378801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,4096,0.02091377807988061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,3584,0.14162311289045545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,512,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,65536,0.15233777629004583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,2560,0.010492444038391113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,16384,0.4852355321248372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,16384,0.021871111459202234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,512,0.014690667390823364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,128,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,8192,0.024874667326609295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,51200,0.07855288849936591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,1536,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,512,0.09850933154424031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,16384,0.03897422220971849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,2048,0.028993778758578833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,768,0.09239200088712905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,256,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,4096,0.052071998516718544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,3584,0.06688621971342298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,768,0.02171733313136631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,1536,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,2048,0.011027555498811932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,3072,0.011519999967681037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,6144,0.035233779086007014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,7168,0.03108533223470052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,64,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,51200,0.04926310976346334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,6144,0.052236444420284696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,65536,1.0396124521891277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,12288,0.13680888546837702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,3072,0.012515555653307172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,16384,0.04335822330580818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,2560,0.060956445005204946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,65536,1.0366524590386283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,65536,0.18217689461178252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,256,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,16384,0.09593688779407078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,512,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,3584,0.17323111163245308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,2560,0.05786222219467163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,51200,0.8989751603868273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,10240,0.16091022226545545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,512,0.03732088870472378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,6144,0.04218488931655884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,3072,0.05412888858053419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,10240,0.045068442821502686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,32,0.006172444257471297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,3584,0.03564622335963779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,65536,1.3636088901095922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,1536,0.1296879980299208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,1024,0.006032889087994893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,3584,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,3584,0.010995555255148145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,3584,0.16035289234585232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,2560,0.07711111174689399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,1536,0.009967111051082611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,64,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,1536,0.008223999705579545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,6144,0.2833991050720215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,128,0.015154666370815702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,512,0.013009777499569787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,6144,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,128,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,16384,0.03596355517705282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,256,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,3072,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,8192,0.010136889086829292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,12288,0.09925599892934163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,2048,0.030444443225860596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,1024,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,128,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,4096,0.07213333580229017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,128,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,16384,0.23402667045593262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,65536,0.12248533301883274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,3072,0.01187999960449007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,3072,0.016490666402710807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,3584,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,2048,0.015038222074508667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,128,0.005061333378156026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,5120,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,65536,0.15065333578321669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,51200,1.027700424194336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,512,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,4096,0.02348888913790385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,4096,0.021839110387696162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,8192,0.010805333654085795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,4096,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,8192,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,4096,0.24899911880493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,5120,0.4324355655246311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,65536,0.9708408779568143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,5120,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,1024,0.005998222364319696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,32,0.01646844380431705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,16384,0.04670133193333944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,8192,0.3229084544711643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,6144,0.012063110868136087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,8192,0.013850666582584381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,8192,0.0938017765680949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,1024,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,128,0.1320399973127577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,256,0.0038568890757030914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,128,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,256,0.015789333317014907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,3072,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,1024,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,65536,0.06583378050062391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,3584,0.029142220815022785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,5120,0.07784355349010892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,65536,0.0749066670735677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,32,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,16384,0.023794665932655334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,5120,0.01887999971707662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,12288,0.016328000360065036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,6144,0.009376000199053023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,10240,0.02749600013097127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,16384,0.0391902228196462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,12288,0.2526720099978977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,4096,0.04108355442682902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,10240,0.5589697625901964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,10240,0.24229865603976777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,4096,0.04012444284227159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,16384,0.028765334023369685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,12288,0.02056533263789283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,65536,0.13495911492241755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,12288,0.01643200053109063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,8192,0.09007022115919326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,5120,0.09007822142706977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,3072,0.01903022163444095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,256,0.014155555102560254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,51200,0.05446755554940966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,2560,0.0158897770775689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,65536,0.1979964441723294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,6144,0.10201422373453777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,64,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,16384,0.07169689072502984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,16384,0.023037332627508376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,5120,0.01943288909064399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,512,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,5120,0.03926400012440152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,51200,0.17761511272854277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,51200,0.18546399805280897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,4096,0.04850844542185465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,51200,0.027414222558339436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,65536,0.1338177786933051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,7168,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,3584,0.02251911163330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,2048,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,32,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,16384,0.1789760059780545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,8192,0.03971022367477417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,16384,0.06618133518430921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,64,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,32,0.013565333353148567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,8192,0.3602462344699436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,5120,0.04660800099372864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,64,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,16384,0.13820978005727133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,256,0.008256889051861232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,10240,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,65536,0.821632014380561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,256,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,64,0.005653333332803514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,512,0.020639111598332722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,3072,0.015666666958067153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,65536,0.1266675525241428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,512,0.03350311186578538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,6144,0.03842133283615112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,3072,0.03271199928389655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,768,0.035990221632851496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,65536,0.08643555641174316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,8192,0.1440382268693712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,3072,0.10417155424753825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,64,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,2048,0.03696177734269036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,5120,0.012346666720178394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,16384,0.02346844474474589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,8192,0.04412000046836006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,10240,0.03630400035116408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,256,0.05039466751946343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,6144,0.06212266948488024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,16384,0.050797333319981895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,8192,0.03052088949415419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,8192,0.11832443873087566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,64,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,32,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,3072,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,12288,0.01365600029627482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,1536,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,10240,0.02790044413672553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,7168,0.036624001132117376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,3584,0.5032329029507107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,12288,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,10240,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,32,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,8192,0.4664986928304036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,2048,0.031903998719321355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,3584,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,7168,0.07357333103815715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,16384,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,5120,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,3072,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,32,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,1536,0.013219555219014486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,3072,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,10240,0.047352890173594155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,2560,0.02027555637889438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,3584,0.013768889009952545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,128,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,12288,0.2214097711775038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,8192,0.058424002594418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,8192,0.3702755504184299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,3072,0.06744088729222615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,4096,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,7168,0.01223111069864697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,768,0.006659555352396435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,2560,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,256,0.014325333966149224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,65536,0.984466658698188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,2560,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,3072,0.021831111775504217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,6144,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,2048,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,8192,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,3072,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,512,0.015471999843915304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,1024,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,3584,0.013304000099500021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,1024,0.01864177816443973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,2048,0.043466667334238686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,8192,0.09689866834216648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,1024,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,51200,0.07149600320392184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,1024,0.027787556250890095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,512,0.0031822222388452957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,6144,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,512,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,768,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,1024,0.003408000080121888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,3072,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,512,0.04804089003139072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,1536,0.03274222215016683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,4096,0.023045332895384893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,2560,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,3584,0.09705511066648696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,10240,0.012054222325483957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,1024,0.003588444242874781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,1024,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,2048,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,65536,0.4070746633741591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,768,0.21031644609239367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,4096,0.014169777433077494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,2048,0.018020444446139865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,4096,0.023119111855824787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,768,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,768,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,65536,0.1693120002746582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,5120,0.016018667154841952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,64,0.012139555480745105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,3072,0.016907556189431083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,512,0.004506666627195147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,4096,0.053130666414896645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,65536,0.16645156012641057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,5120,0.018722666634453666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,2048,0.006700444552633498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,3584,0.05376533336109585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,32,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,4096,0.015319999721315173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,2560,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,65536,0.057032889790005155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,10240,2.0980745951334634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,10240,0.02351911034848955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,4096,0.024491555160946313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,256,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,256,0.006325333482689328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,1536,0.006439111298984951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,16384,0.059281779660118945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,5120,0.01240533341964086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,7168,0.062459554937150746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,3072,0.0853911108440823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,1536,0.053730666637420654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,65536,3.179938634236654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,8192,0.015172445111804538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,5120,0.015615999698638916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,3584,0.016043555405404832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,5120,0.025260445144441392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,3072,0.02897955642806159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,51200,0.14177155494689941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,5120,0.05128533310360379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,3584,0.24698045518663195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,12288,0.20569688744015166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,65536,0.07247910896937053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,768,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,7168,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,5120,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,256,0.005695999910434087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,6144,0.4328426784939236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,32,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,2560,0.014906666345066495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,10240,0.5221501986185709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,16384,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,8192,1.8465707567003038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,1536,0.1228791077931722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,768,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,10240,0.035079111655553184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,65536,0.0777404440773858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,5120,0.07512178023656209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,2048,0.01352533366945055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,32,0.03572622272703383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,7168,0.03865689039230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,5120,0.04066488809055752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,256,0.05902044640647041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,3072,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,12288,0.1460675530963474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,10240,0.03274044394493103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,768,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,32,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,6144,0.008551999926567078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,512,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,12288,0.034938666555616595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,10240,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,2560,0.014537778165605334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,6144,0.04741244514783224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,1024,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,6144,0.47484710481431747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,768,0.04224266608556112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,64,0.011423110961914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,65536,0.6161280208163792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,3072,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,51200,0.08487733205159505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,256,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,2560,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,5120,0.01258933295806249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,2048,0.008806222014957005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,10240,0.029115554359224107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,2560,0.047988444566726685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,1536,0.02905422117975023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,7168,0.036927110619015165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,7168,0.040989332728915744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,51200,0.04820177621311612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,10240,0.08621244298087226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,16384,0.05081511206097073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,1024,0.007707555260923173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,10240,0.01941333380010393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,10240,0.05755199988683065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,4096,1.2795644336276584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,4096,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,6144,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,51200,0.07147022088368733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,10240,0.027465777264700994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,3584,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,6144,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,8192,0.015579554769727917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,3072,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,10240,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,16384,0.018935110833909776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,2560,0.038310223155551486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,8192,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,65536,0.3887288835313585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,3584,0.08294133345286052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,512,0.015972443752818637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,1536,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,128,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,1536,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,6144,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,2048,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,12288,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,65536,0.0766071081161499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,1536,0.06691822078492907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,3584,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,51200,0.4885573387145996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,128,0.04456444581349691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,256,0.015331554743978711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,65536,0.04938488867547777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,64,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,2560,0.05773422453138563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,512,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,2048,0.006686222222116258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,768,0.03723466727468703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,8192,0.4298746585845947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,12288,0.016879111528396606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,5120,5.513942294650608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,3584,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,65536,0.13411466280619302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,256,0.01443822185198466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,3072,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,12288,0.040532443258497454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,8192,0.013340444200568728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,2560,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,3584,0.015048000547620984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,7168,0.026169778572188482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,2048,0.06598755386140612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,2560,0.012839111189047495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,5120,0.0211173329088423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,8192,1.1583351559109158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,1536,0.00599644457300504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,1024,0.01924711134698656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,12288,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,4096,0.050148444043265454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,1536,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,5120,0.11916266547309028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,7168,0.040941334433025785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,3072,0.012895999683274163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,256,0.0032835555159383347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,65536,0.44085245662265354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,65536,0.05272000034650167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,5120,0.034295111894607544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,2048,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,16384,0.2958302232954237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,3072,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,768,0.02796799937884013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,2560,0.033017777734332614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,12288,0.056735999054378934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,16384,0.06843822532229953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,6144,0.017356443736288283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,3072,0.013393777940008374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,16384,0.022537777821222942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,2560,0.4066035482618544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,512,0.04719822274314033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,512,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,6144,0.019693333241674636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,1536,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,6144,0.01954844428433312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,10240,0.09108977847629124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,16384,2.4068116082085504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,8192,0.020417778028382193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,6144,0.01992177797688378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,4096,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,3072,0.06048444244596693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,3072,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,12288,0.043895999590555825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,6144,0.03687644335958693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,65536,0.06001777781380547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,10240,0.016420443852742512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,32,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,2048,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,10240,0.03698933455679152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,2560,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,1024,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,1536,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,6144,0.035988443427615695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,16384,0.0670693318049113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,16384,0.18086932765112984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,16384,0.010330666270520952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,12288,0.012809777425395118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,1536,0.016880000631014507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,8192,0.09904088576634724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,51200,0.15236800246768528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,8192,0.9353244569566516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,16384,0.02270222206910451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,16384,0.01089599976936976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,16384,0.29680267969767254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,1024,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,2560,0.026937777797381084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,51200,0.07517600059509277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,3072,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,6144,1.4305173026190863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,4096,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,128,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,2560,0.03583466675546434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,5120,0.04088000125355191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,1024,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,5120,0.00982844415638182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,128,0.017056889004177518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,12288,0.046012444628609546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,4096,0.04849866694874234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,3584,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,10240,0.029098666376537744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,1536,0.029829333225886028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,2048,0.017198221551047433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,512,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,2048,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,1024,0.03722755445374383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,5120,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,12288,0.5450871255662706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,3584,0.02648711038960351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,16384,0.2517368793487549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,3072,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,4096,0.04351644383536445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,1024,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,3072,0.008485332959228093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,768,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,51200,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,65536,0.049469331900278725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,10240,0.03934577769703335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,3584,0.01795644395881229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,5120,0.11697155899471706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,1024,0.01796266602145301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,1536,0.0037919998996787598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,6144,0.010503110786279043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,12288,0.06534133354822795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,5120,0.18103733327653673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,10240,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,256,0.0064604443808396654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,6144,0.014126222994592456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,4096,0.2549946573045519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,128,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,5120,0.02161066730817159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,256,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,6144,0.04535288943184746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,128,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,2560,0.040362666050593056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,2560,0.31816622945997447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,512,0.10638311174180771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,3584,0.01054222219520145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,1024,0.06042844719356961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,256,0.015094222293959724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,51200,0.2865848806169298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,8192,0.01738933391041226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,10240,0.02996088729964362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,256,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,8192,0.05031555559900072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,10240,0.026407111022207472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,6144,0.016116445263226826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,65536,0.5851279894510905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,512,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,32,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,8192,0.0510168903403812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,3072,0.00904088881280687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,32,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,768,0.10424000024795532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,3072,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,10240,0.013977777626779346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,256,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,10240,0.0125946667459276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,7168,0.014347554908858405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,128,0.03065955638885498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,65536,0.20085155963897705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,5120,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,2048,0.012563555604881711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,5120,0.12091733349694146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,6144,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,1536,0.027850667635599773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,2048,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,1024,0.021856889128684998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,4096,0.027284445034133062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,1024,0.019983111156357657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,65536,0.06518489122390747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,4096,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,2560,0.03692355420854356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,1536,0.006044444524579578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,2048,0.051544000705083214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,8192,0.03422044383154975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,2560,0.05141777793566386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,2048,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,16384,0.045810666349199086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,2560,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,128,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,6144,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,7168,0.05038133263587952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,2560,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,51200,0.046795556942621865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,3072,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,2048,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,512,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,768,0.020090666082170274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,51200,0.09986044300927056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,12288,0.4693137804667155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,768,0.06446666849984063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,7168,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,10240,0.10197866625256008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,3584,0.01145244472556644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,5120,0.020079111059506733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,2048,0.006319111006127463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,7168,0.012690666649076672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,64,0.004620444443490771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,3584,0.010990222295125326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,1536,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,32,0.002096000024014049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,3584,0.02831377916865879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,1024,0.011931555966536203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,2048,0.02241777711444431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,10240,0.08988622162077163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,12288,0.047615110874176025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,12288,0.025384889708624944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,10240,0.031017776992585924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,128,0.0780586666531033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,8192,0.022175111704402502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,256,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,2048,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,256,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,7168,0.028785778416527644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,128,0.014160888062583076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,5120,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,4096,0.016215110818545025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,768,0.007354666789372762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,768,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,65536,0.11899644798702663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,4096,0.04177422324816386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,768,0.03211999932924906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,768,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,1024,0.016902221573723685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,10240,0.016328000360065036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,4096,0.006700444552633498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,32,0.04126044445567661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,1024,0.028373334142896865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,512,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,3584,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,2560,0.019904888338512845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,768,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,2048,0.03611822260750665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,51200,4.338712056477864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,4096,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,1024,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,3584,1.3236657248602974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,65536,0.22044977876875135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,2560,0.025059555967648823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,10240,0.07560888926188152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,51200,0.06155733267466227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,128,0.1670222282409668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,2560,0.024689777029885188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,12288,0.21244888835483125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,12288,0.023660444551044043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,768,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,65536,0.028065777487225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,16384,0.02683555583159129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,128,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,2048,0.010207110808955299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,128,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,64,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,4096,0.02067466742462582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,2048,0.027273777458402846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,2048,0.039204445150163435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,16384,0.16752889421251085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,2560,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,12288,0.4253528912862142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,7168,0.09893333249621922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,1024,0.1379973358578152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,2048,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,16384,0.0857422219382392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,10240,0.011711111499203576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,5120,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,10240,0.0739920006857978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,512,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,8192,0.00741244438621733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,3584,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,16384,0.029339555237028334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,512,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,65536,0.26394134097629124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,7168,0.05120266808403862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,51200,0.12285421954260932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,65536,0.2993306583828396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,8192,0.9297599792480469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,6144,0.013624889155228933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,2048,0.004661333229806689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,1024,0.06372800138261583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,64,0.027208000421524048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,8192,0.024897777371936377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,1536,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,8192,1.3671947055392797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,2048,0.7509102291531033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,5120,0.04275288846757677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,16384,0.04246222310596042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,16384,0.08776533603668213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,65536,0.47921154234144425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,768,0.06992710961235894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,10240,0.013295110729005603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,7168,0.011795555551846823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,4096,0.04951911171277364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,1536,0.2803991105821398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,16384,0.019654222660594516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,2560,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,64,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,512,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,16384,0.022984888818528917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,3584,0.01958311100800832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,256,0.015797333584891427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,8192,0.07177511188719007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,8192,0.056306666798061795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,768,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,1024,0.16683110925886366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,12288,0.7373244497511121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,51200,0.09830311271879409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,128,0.012249777714411417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,3584,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,16384,0.021728888154029846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,8192,0.014353776971499125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,2048,0.029022223419613306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,6144,0.05304622319009569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,768,0.04452177882194519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,12288,0.05088622371355692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,2560,0.019202666150199044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,5120,0.02265600032276577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,7168,5.709815131293403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,65536,0.12233866585625543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,32,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,3072,0.011075555450386472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,32,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,7168,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,4096,0.008949333594905006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,12288,0.2700124316745334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,6144,0.013843555417325763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,65536,0.08401866753896077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,8192,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,2048,0.041025777657826744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,512,0.0243493335114585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,16384,0.022594667143291895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,512,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,256,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,32,0.08436799711651272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,51200,0.09832000070148045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,51200,0.8883244196573893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,768,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,2048,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,128,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,2048,0.010679999987284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,2560,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,3072,0.09451822439829509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,2560,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,51200,0.015139554937680563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,16384,0.12325510713789199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,4096,0.07257511218388875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,1536,0.0199955552816391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,3584,0.03260355525546604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,768,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,32,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,3072,0.025965332984924316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,4096,0.02068088948726654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,5120,0.10476177930831909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,8192,0.08267822530534533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,51200,0.05722577704323662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,4096,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,768,0.003882666842805015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,5120,0.031750222047170006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,768,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,6144,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,1536,0.02935289012061225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,2560,0.03564888901180691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,3072,0.01347733371787601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,3584,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,6144,0.01074577785200543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,1536,0.12230043941073948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,64,0.04104355639881558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,1536,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,5120,0.6880293422275119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,3584,0.029319110843870375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,6144,0.08492800262239243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,768,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,64,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,512,0.009106666677527958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,512,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,32,0.0017991111510329777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,65536,0.15676177872551814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,256,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,16384,0.02015822297996945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,32,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,256,0.036343110932244196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,3584,0.06890400250752766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,64,0.0041982221106688184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,12288,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,128,0.024832889437675476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,4096,0.06051199966006809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,2048,0.022672888305452134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,6144,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,12288,1.0193653106689453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,256,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,512,0.0038204445607132385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,4096,0.025063110722435847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,5120,0.011707555916574268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,3072,0.0200311111079322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,8192,0.007863110966152614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,12288,2.9892853630913625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,51200,0.10443111260732015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,32,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,3072,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,4096,0.031833777825037636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,65536,0.7810399797227648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,6144,0.040876444843080305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,5120,0.02516355613867442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,1536,0.06957599851820204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,12288,0.24264444245232475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,65536,0.0569653312365214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,65536,0.42006667455037433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,2560,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,12288,0.11064800288942124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,3072,0.011136000355084738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,51200,0.152164445983039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,51200,0.049822221199671425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,3072,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,16384,0.021913778450753953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,256,0.00240000006225374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,1536,0.05998133288489448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,10240,0.23730754852294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,5120,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,65536,0.048696001370747886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,2048,0.00463733325401942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,12288,0.03222044308980306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,7168,0.2759946717156304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,768,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,3584,0.015568888849682279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,64,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,1024,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,1536,0.04125333494610257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,3584,0.018914666440751817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,2560,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,3072,0.01826133330663045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,128,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,16384,0.05833244323730469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,256,0.010138666464222802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,768,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,12288,0.20278310775756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,16384,0.15850310855441624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,65536,0.2590968873765734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,10240,0.11248355441623265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,51200,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,512,0.014558222558763293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,768,0.011679111255539788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,2048,0.007483555210961237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,256,0.015346666177113851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,51200,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,1024,0.14584355884128147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,5120,0.01992177797688378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,8192,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,16384,0.07589511076609294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,51200,0.04721866713629829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,1536,0.06924711333380805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,512,0.01629955569903056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,51200,0.46599912643432617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,64,0.009713778065310584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,51200,1.0139386918809679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,6144,0.14403555128309461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,2048,0.004171555654870139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,10240,0.5265697903103298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,8192,0.09320000145170425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,256,0.019239111079110038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,12288,0.016653334101041157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,51200,0.04831555485725403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,16384,0.032272001107533775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,1536,0.03639199998643663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,128,0.014819555812411837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,256,0.006723555425802867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,12288,0.05998133288489448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,5120,0.03455555438995361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,1024,0.012355555262830524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,8192,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,12288,0.024030221833123103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,2560,0.04405155446794298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,10240,0.011663999822404651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,16384,0.12695733706156412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,512,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,1024,0.030046221282747056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,64,0.16047022077772352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,1536,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,6144,0.036729776197009616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,8192,0.05339466532071432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,3584,0.016030222177505493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,2560,0.4073048962487115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,5120,0.023037332627508376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,2048,0.022091555926534865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,51200,0.08316444688373141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,512,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,16384,0.26388798819647896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,51200,1.3226648966471355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,51200,0.019650666250122916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,51200,0.42869867218865293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,1536,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,2048,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,128,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,10240,0.05594578054216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,2560,0.34882044792175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,768,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,1024,0.20644711123572457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,128,0.014557333456145393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,3584,0.018607111440764535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,10240,0.00463377767139011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,128,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,65536,0.0864844454659356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,8192,0.043213334348466664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,5120,0.9238746431138781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,16384,0.021194666624069214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,51200,0.55440886815389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,10240,0.1639288928773668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,128,0.011824000212881299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,2560,0.17137244012620714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,5120,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,8192,0.022138666775491502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,64,0.05497422152095371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,1024,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,8192,0.021879111727078755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,256,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,51200,0.061463110976748996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,3072,0.026314665873845417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,2048,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,8192,0.022284444835450914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,65536,0.04996622271007962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,512,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,768,0.031151112582948472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,2048,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,768,0.003318222239613533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,12288,0.018862222631772358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,512,0.024600888291994732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,32,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,2560,0.052926222483317055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,512,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,4096,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,5120,0.028384000062942505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,1024,0.012321777641773224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,2048,0.00423200014564726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,512,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,65536,0.14707556035783556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,51200,0.1482435597313775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,768,0.129530668258667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,1536,0.019111999207072787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,7168,0.045574221346113414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,128,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,5120,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,7168,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,16384,0.033972445461485125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,768,0.010434666441546546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,1024,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,3072,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,128,0.01722400055991279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,65536,0.05178755521774292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,3072,0.10939733187357585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,65536,0.01142133358452055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,512,0.0038720000949170855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,12288,0.1492497788535224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,1536,0.03449599941571554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,512,0.018523555662896898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,1024,0.0057911111248864075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,8192,0.02292799949645996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,51200,0.05531466669506497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,1536,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,1536,0.02937955657641093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,2048,0.023738667368888855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,1024,0.005688888745175467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,1024,0.0324942237801022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,512,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,10240,0.014113777213626437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,51200,0.00962488849957784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,51200,0.4987084600660536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,16384,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,4096,0.01904533306757609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,5120,0.053655998574362866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,16384,0.41739288965861004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,7168,0.053399999936421715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,2048,0.023039110832744177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,512,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,6144,0.1842648850546943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,51200,0.1013093325826857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,128,0.0298977792263031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,512,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,3072,0.012293332980738746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,7168,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,65536,0.5906097624037001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,1536,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,3584,0.014501333236694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,65536,0.06363644202550252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,512,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,256,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,128,0.011237333218256632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,128,0.02985422147644891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,3584,0.02404622236887614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,128,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,4096,0.0038968887594011095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,768,0.06172977553473579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,7168,0.024149333437283833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,65536,1.679483519660102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,128,0.030027555094824895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,1024,0.02204622162712945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,768,0.02758933272626665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,768,0.016415110892719693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,12288,0.08124533626768324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,12288,0.016380444169044495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,1536,0.011897777517636618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,7168,0.04999822378158569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,5120,0.02073955535888672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,6144,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,1536,0.019254222512245178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,768,0.014304000470373364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,8192,0.02981511089536879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,768,0.021589333812395733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,512,0.00740799970097012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,3584,0.1948853333791097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,51200,0.5296559863620335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,256,0.006798222247097228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,3584,0.05370755659209358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,6144,2.375621371799045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,65536,1.10716003841824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,8192,0.06389777527915107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,8192,0.02676888969209459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,16384,0.09107644690407647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,6144,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,51200,0.04821244544453091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,3072,0.026330666409598455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,16384,0.012494222157531314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,12288,0.04412000046836006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,512,0.31073421902126735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,1536,0.003567111161020067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,5120,0.012094222009181976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,2560,0.06310488780339558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,32,0.017795556121402316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,1024,0.026718222432666357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,12288,0.2822195423973931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,12288,0.3955253230200873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,3584,0.019554666346973844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,768,0.004554666578769684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,1536,0.10451022121641372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,2048,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,32,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,32,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,2560,0.01716888944307963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,3072,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,3072,0.13463021649254694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,5120,0.00907199995385276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,768,0.014500444134076437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,51200,0.10852799812952678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,256,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,1024,0.03765955567359924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,2560,0.020598222812016804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,6144,0.009476444787449306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,4096,0.027680888772010803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,12288,0.02749155627356635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,256,0.02151022189193302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,1536,0.015815999772813585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,32,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,4096,0.03199555476506551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,1536,0.03830577929814657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,65536,1.7287670771280925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,8192,0.06930133369233873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,2048,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,3072,0.03707733419206407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,1024,0.07414399915271334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,4096,0.08574044704437256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,65536,0.23153956731160483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,51200,0.0642017788357205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,2560,0.02544533378548092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,1536,0.032256000571780734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,768,0.008368888662921058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,2048,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,1536,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,4096,0.012522666818565793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,6144,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,2560,0.009000000026490953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,6144,0.31387199295891655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,8192,0.02776355544726054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,16384,0.05713333023918999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,3072,0.02052888936466641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,6144,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,768,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,768,0.011884444289737277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,10240,0.03074222140842014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,6144,0.16808533668518066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,64,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,256,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,7168,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,32,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,10240,0.007864000068770515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,1024,0.0031475555151700974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,2048,0.01015022231472863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,3072,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,51200,1.0697982576158311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,16384,0.11575644546084934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,64,0.0038977778620190094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,2048,0.015486222174432544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,128,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,5120,0.15655467245313856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,2560,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,768,0.0237137774626414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,5120,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,1536,0.006197333335876465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,51200,0.8027617666456434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,256,0.005663110978073544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,1536,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,16384,0.023185777995321486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,2560,0.07507910993364122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,2560,0.016229333149062265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,6144,0.06729777654012044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,32,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,51200,0.05024977856212192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,256,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,2560,0.05741155809826321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,256,0.002458666761716207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,6144,0.10619999965031941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,2560,0.007089777953094906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,8192,0.006214222146405115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,1536,0.003536000019974179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,3584,0.06185422341028849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,64,0.01990133358372582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,3584,0.06103822257783678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,256,0.009184000392754873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,7168,1.317765341864692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,1024,0.14334755473666722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,1024,0.09504622220993042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,10240,0.19762667020161948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,3584,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,51200,0.112972444958157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,2048,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,2560,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,3584,0.4916204346550836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,512,0.05835022528966268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,10240,0.09490133656395806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,6144,0.020634666085243225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,10240,0.017439110411538016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,768,0.0032319999817344877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,12288,0.016932444439993966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,3072,0.026343110534879897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,2560,0.036221332020229764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,128,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,3584,0.02749600013097127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,65536,0.12113955285814072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,128,0.017631111873520743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,8192,0.1866471105151706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,7168,0.02520711057715946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,512,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,1024,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,1024,0.017858665850427415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,1024,0.008240888516108194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,1536,0.01939377850956387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,65536,3.090665817260742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,2048,0.03624533282385932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,8192,0.011176000038782755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,256,0.03421155611673991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,3072,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,2048,0.05791999896367391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,6144,0.014182221558358936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,512,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,12288,0.02988977895842658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,4096,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,5120,0.08262666728761461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,3584,0.033003555403815374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,6144,0.020285333196322124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,5120,0.03157866663402981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,128,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,256,0.025705777936511572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,64,0.004929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,768,0.05504622062047323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,2560,0.013825777504179211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,5120,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,51200,0.38068177964952254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,1536,0.04115466607941522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,65536,1.692578633626302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,1024,0.0222382214334276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,3072,0.02092355489730835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,3584,0.036045332749684654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,8192,0.02977777851952447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,12288,0.8787608676486545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,768,0.023403555154800415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,512,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,1536,0.028677334388097126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,128,0.00962577760219574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,256,0.010057777994208867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,512,0.039770666095945574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,512,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,3584,0.10785777701271905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,4096,0.042916443612840444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,16384,0.05490844448407491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,65536,0.30560443136427135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,4096,0.008636444807052612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,51200,0.24358754687839082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,16384,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,3584,0.10917511251237656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,4096,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,3072,0.019536000159051683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,2560,0.005379555539952383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,3072,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,2048,0.026294221480687458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,128,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,2048,0.1376711130142212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,3584,0.015449777245521545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,12288,0.06879733006159465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,256,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,8192,0.013468444347381592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,2560,1.0156897438897026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,1024,0.026837334036827087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,51200,0.8613848686218262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,32,0.01964000033007728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,8192,0.07651288641823663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,768,0.026538666751649644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,10240,0.023404444257418316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,128,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,51200,0.06601333618164062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,64,0.023932443724738225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,65536,0.16375288698408338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,7168,0.0200311111079322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,256,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,65536,0.0584684411684672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,7168,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,16384,0.7561288939581977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,1024,0.10832266675101386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,51200,0.31942666901482475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,7168,0.02128355536195967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,3072,0.007527111305130853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,4096,0.1340079969829983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,256,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,3584,0.019480889042218525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,512,0.056458665264977344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,3584,0.029776000314288672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,51200,0.056008888615502246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,10240,0.1923688915040758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,7168,0.09161155753665501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,6144,0.028520001305474177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,128,0.01722755531469981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,3072,0.019472888774342008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,16384,0.019375110665957134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,5120,0.13106844160291883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,1536,0.03672711054484049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,1024,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,5120,0.008745777938101027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,2560,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,768,0.012568888564904531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,64,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,256,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,1024,0.039064887497160174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,768,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,12288,0.10636533631218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,16384,0.02166844407717387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,2048,0.030124445756276447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,3072,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,8192,0.471019532945421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,2048,0.028142223755518597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,3072,4.697174919976129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,3072,0.05810222360822889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,4096,0.016869333055284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,3072,0.06110577450858223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,8192,0.04491022229194641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,256,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,6144,0.011076444553004371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,5120,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,512,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,3584,0.5281573401557075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,2560,0.014303111367755465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,256,0.02127555509408315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,16384,0.04633333285649618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,768,0.022294221652878657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,12288,0.051333334710862905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,6144,0.034698665142059326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,32,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,128,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,65536,0.06958844264348348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,7168,0.01089866707722346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,5120,0.03576177689764235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,65536,0.16594666904873318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,8192,0.024007999234729346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,4096,0.016882666283183627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,16384,0.16467644108666316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,10240,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,6144,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,12288,0.06567288769616021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,12288,0.06088978052139282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,64,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,7168,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,1536,0.06030488676495022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,6144,0.01777866648303138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,51200,0.12070755163828532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,12288,0.011790222591824003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,6144,0.010048888623714447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,16384,0.14330932829115126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,64,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,512,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,6144,0.10486400127410889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,4096,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,768,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,10240,0.05331733491685656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,12288,0.0138568886452251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,7168,0.019318222999572754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,2560,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,51200,0.03798844416936239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,5120,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,32,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,16384,0.009420444567998251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,65536,13.824899461534288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,6144,0.016543111867374845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,1536,0.01776444415251414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,3072,0.006745777610275481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,128,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,65536,4.072957356770833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,32,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,128,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,1024,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,768,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,51200,2.4757715861002603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,2048,0.04142755601141188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,1536,0.339279121822781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,3584,0.037117332220077515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,768,0.02128533356719547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,768,0.020185778538386028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,32,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,6144,0.055083556307686694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,64,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,8192,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,12288,0.018728888697094388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,128,0.027129777603679236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,4096,0.05958400170008341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,8192,0.03808800048298306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,12288,0.013463111387358772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,65536,0.06958578030268352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,7168,0.03365688853793674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,5120,0.008456889126035903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,8192,0.1921724478403727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,3584,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,3072,0.00924622184700436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,3072,0.021662222014533147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,1536,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,2560,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,512,0.020616888999938965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,7168,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,32,0.005817777580685086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,256,0.28908443450927734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,51200,0.19377956125471327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,12288,0.026943999860021803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,65536,0.3970924483405219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,6144,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,2560,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,2560,0.03352088729540507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,2048,0.027172444595230952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,7168,0.018276444739765592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,10240,0.05429600013626946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,10240,0.09656266371409099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,512,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,768,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,8192,0.03369155526161194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,3072,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,5120,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,12288,1.6334702173868816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,7168,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,128,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,3072,0.06610488891601562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,12288,0.08875199821260239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,3072,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,16384,1.5370559692382812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,3072,0.1587795548968845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,256,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,3072,0.036279112100601196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,1536,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,8192,0.007567110988828871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,2560,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,51200,0.06720266739527385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,128,0.04404888881577385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,3584,0.028162668148676556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,3072,0.02273333403799269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,4096,0.026376888155937195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,65536,0.11749422550201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,1024,0.05290311243798998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,128,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,10240,0.026103999879625108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,256,0.005413333161009683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,3584,0.017312889297803242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,2048,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,3584,1.296913782755534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,128,0.002256888896226883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,3072,0.04697688751750522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,10240,1.2085840437147353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,2560,0.006768888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,3072,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,2560,0.007190222541491191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,12288,0.03927733169661628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,12288,0.013972444666756524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,51200,0.04897422260708279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,8192,0.32438309987386066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,51200,0.7040328979492188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,3072,0.014316444595654806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,256,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,64,0.022944000032213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,2048,0.019540444016456604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,3072,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,7168,0.07860977782143487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,1024,0.015450666348139444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,32,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,12288,0.0339902208911048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,65536,0.14517688751220703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,256,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,51200,0.15504088666703966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,3584,0.03730577892727322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,6144,0.02998044424586826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,4096,0.013733333183659447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,256,0.00702400008837382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,16384,0.0818835563129849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,65536,0.5882719887627496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,65536,0.2724977864159478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,1024,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,2560,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,16384,0.044263998667399086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,32,0.025829333398077223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,51200,0.43003021346198184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,12288,0.01346666696998808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,256,0.015135111080275642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,12288,0.12229866451687282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,128,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,16384,0.033548444509506226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,1024,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,5120,0.13143466578589544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,768,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,64,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,10240,0.034781333472993635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,51200,0.605206224653456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,8192,0.8049617873297797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,1024,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,2560,0.08677066697014703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,7168,0.029595557186338637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,5120,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,3584,0.03844889005025228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,65536,2.965163548787435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,768,0.013939556148317126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,1536,0.009479111267460717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,2560,0.023815110325813293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,3584,2.6997386084662547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,256,0.025394666526052687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,3072,0.09357333183288574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,65536,0.06941422488954332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,2048,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,3072,1.1414693196614583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,7168,0.12363999419742161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,4096,0.015443555182880826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,1024,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,4096,0.011525332927703857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,5120,0.013073777986897362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,7168,0.03569777806599935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,65536,1.212550269232856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,4096,0.027621333797772724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,6144,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,768,0.018046221799320646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,8192,0.0961644450823466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,51200,1.578094270494249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,768,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,768,0.03301511208216349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,1024,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,512,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,6144,0.011776000261306763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,12288,0.01717511150572035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,16384,0.02537333303027683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,16384,0.07713955640792847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,256,0.010849778023031024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,1024,0.006162666612201267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,10240,0.005024888863166173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,4096,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,4096,0.12887377209133571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,128,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,65536,0.07120266887876722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,1536,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,4096,0.010021333065297868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,12288,0.03584355447027419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,10240,0.09504089090559219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,8192,0.02330844435426924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,12288,0.06684888733757867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,2048,0.03924977779388428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,5120,0.11656888326009114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,256,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,32,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,5120,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,6144,0.02436977790461646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,12288,0.3919644355773926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,8192,0.07195111115773518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,3072,0.018265777164035372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,4096,0.009221333596441481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,512,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,2048,0.015239111251301236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,7168,0.025058666865030926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,512,0.009242666264375051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,10240,0.048510223627090454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,2560,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,2048,0.08880177471372817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,16384,0.012141333685980903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,1024,0.011227555572986603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,3072,0.20360622141096327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,51200,0.04992444316546122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,2048,0.0188391109307607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,5120,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,2048,0.020768889122539096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,128,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,3072,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,3584,0.3062995539771186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,256,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,128,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,512,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,8192,0.2048355605867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,3584,0.03500266538725959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,8192,0.04127111037572225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,6144,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,16384,1.1959884431627061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,10240,0.19746132691701254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,65536,0.4311031235588922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,6144,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,7168,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,32,0.0021119999388853707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,3072,0.013967111706733704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,16384,0.022954665952258643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,2560,0.015785778562227886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,1024,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,3584,0.10986844698588054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,256,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,128,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,65536,0.11687200599246556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,4096,0.22977866066826713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,768,0.01632000009218852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,1024,0.010259555445777046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,512,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,1536,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,12288,0.01945244438118405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,256,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,7168,0.07256266805860731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,1024,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,256,0.019654222660594516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,65536,0.9236408869425455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,10240,0.04022577736112807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,2560,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,10240,0.03644088904062907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,4096,0.23696443769666883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,7168,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,128,0.020032000210550096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,5120,0.0158897770775689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,256,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,2560,0.025282666087150574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,512,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,256,0.017231111725171406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,768,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,10240,0.02272533377011617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,51200,0.05915911330117119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,256,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,51200,0.18300978342692056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,3072,0.02183288832505544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,5120,0.09562666548622979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,65536,0.2168604400422838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,768,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,5120,0.04650844468010796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,5120,0.015951999359660678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,10240,0.023401778605249193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,51200,0.07545689079496595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,768,0.01829777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,10240,0.01664266652531094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,512,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,2048,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,10240,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,7168,0.10423111253314549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,51200,0.11332444349924724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,512,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,6144,0.14172089099884033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,3584,0.04457599918047587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,51200,0.11671200063493516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,512,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,512,0.037104000647862755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,6144,0.1838444471359253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,6144,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,12288,0.02104355560408698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,128,0.026552889082166884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,2048,0.008421333299742805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,6144,0.022335110439194575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,512,0.019609777463806998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,10240,0.04417333337995741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,8192,0.14428267214033338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,7168,0.03173333406448364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,1536,0.03249955508444045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,12288,0.02200977835390303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,256,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,256,0.017234666479958426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,3584,0.2897111044989692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,1536,0.023743111226293776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,12288,0.018796443939208984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,65536,0.16426755322350398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,2560,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,5120,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,16384,3.1174373626708984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,2048,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,3072,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,768,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,8192,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,65536,0.1138426727718777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,1536,0.011353777514563667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,1536,0.028880000114440918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,128,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,4096,0.008264000217119852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,256,0.0262773334980011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,4096,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,1536,0.0377146667904324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,8192,1.3140044742160373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,65536,0.06725155644946627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,256,0.0032142222755485107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,8192,0.11287644174363877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,6144,0.019175110591782462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,6144,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,512,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,1536,0.025349333882331848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,16384,0.5387919743855795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,256,0.01237777786122428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,16384,1.780825720893012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,8192,0.05202577842606438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,7168,0.045960889922248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,16384,0.03074399961365594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,3584,0.12856445047590467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,1024,0.01696622206105126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,10240,0.03340533375740051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,2560,0.026382222771644592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,32,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,128,0.0019991110182470745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,768,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,6144,0.05202133456865946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,10240,0.050682667228910655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,10240,1.134886211819119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,8192,0.015279110934999255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,1024,0.028421332438786823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,16384,0.26369865735371906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,12288,0.27372445000542533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,32,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,1024,0.016255110502243042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,12288,1.7830924987792969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,512,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,10240,0.018198221921920776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,4096,0.03614933292071024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,256,0.008164444731341468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,3072,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,10240,0.024067555864651997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,512,0.03501155641343858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,1536,0.018975110517607797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,128,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,65536,6.510647243923611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,3584,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,3072,0.06052533123228285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,7168,0.02233244478702545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,32,0.003527111063400904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,3584,0.02030933399995168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,12288,0.10199822319878472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,3072,0.013232000172138214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,51200,0.06169066826502482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,512,0.046866665283838906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,128,0.13716710938347712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,3584,0.13287822405497232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,7168,0.19444799423217773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,32,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,32,0.006549333532651265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,1536,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,1024,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,128,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,10240,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,12288,0.0310204459561242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,3072,0.4067697789933946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,51200,0.07674844397438897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,7168,0.017796445224020217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,8192,0.059671998023986816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,1536,0.02032444377740224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,12288,0.6506497595045302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,5120,0.02680355476008521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,3072,0.017674666312005784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,10240,0.0707226660516527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,10240,0.014545778433481852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,512,0.04927644464704725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,8192,0.016740444633695815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,128,0.01349511080318027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,16384,0.02533866630660163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,8192,0.01107377807299296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,3584,0.16990221871270073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,7168,0.011695110963450538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,65536,0.03594844539960226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,512,0.0742408898141649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,1536,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,3584,0.02812622321976556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,64,0.0015128889224595493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,8192,0.07134044170379639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,12288,0.06996088557773165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,5120,0.04231733414861891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,12288,0.01937955617904663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,3584,0.02313599983851115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,12288,0.008952889177534316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,51200,0.05235200126965841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,4096,0.43827022446526426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,5120,0.030532442861133154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,3072,0.027107555005285475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,6144,0.6179831292894151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,2560,0.07929155561659071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,3584,0.021817778547604878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,256,0.02232444451914893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,1024,0.004210666649871402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,5120,0.02204711072974735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,2560,0.011688888900809817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,2560,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,4096,0.009641778137948778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,512,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,1024,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,2048,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,128,0.001976888833774461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,5120,0.026199999782774184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,2560,0.012181333369678922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,1024,0.029640889830059473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,512,0.990056037902832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,10240,0.03851111067665948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,768,0.0033093332830402586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,65536,0.12235911687215169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,10240,0.12420711252424453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,128,0.015135111080275642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,16384,0.01628177695804172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,4096,0.09663289123111302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,3072,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,10240,0.09847378068500096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,3584,0.058303998576270215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,2560,0.007119110888904995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,16384,0.1869351069132487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,1536,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,2048,0.011219555305110084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,8192,0.06968977716233996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,64,0.0035626664757728577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,32,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,2048,0.0851866669125027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,8192,0.14725777837965223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,128,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,8192,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,16384,0.164701329337226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,12288,0.022663111488024395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,64,0.00962577760219574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,5120,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,2048,0.010804444551467896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,10240,0.06113333172268338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,65536,0.1149982213973999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,6144,0.770900461408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,65536,0.0974853303697374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,5120,0.03915822174814012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,10240,0.036312888065973915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,7168,0.021584888299306233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,65536,0.024691555235120986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,3584,0.010944888823562197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,512,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,1536,0.020463999774720933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,4096,0.050266666544808276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,10240,0.05512711074617174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,5120,0.045716444651285805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,3584,0.03574400146802267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,2560,0.03643555442492167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,51200,0.6807813114590116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,65536,10.7159423828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,768,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,65536,0.08908177746666802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,51200,0.2920435534583198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,1536,0.01126933346192042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,8192,0.02460177739461263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,512,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,6144,0.04037066631846958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,128,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,1536,0.04432533184687296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,3584,0.13381066587236193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,3584,0.05633155504862467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,2048,0.02586311101913452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,512,0.006057777752478917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,2048,0.004855999930037392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,6144,0.03240711159176297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,51200,0.12705600261688232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,256,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,4096,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,4096,0.010772444307804108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,3584,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,3584,0.0125608891248703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,512,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,7168,0.01844800015290578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,512,0.06479644775390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,10240,0.01699911057949066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,12288,0.06366933054394193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,3584,0.0895528859562344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,2048,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,5120,0.00869511150651508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,2560,0.03874311182234023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,7168,0.061958220269944936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,3072,0.02546311087078518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,1536,0.1163955529530843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,7168,0.02163022259871165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,512,0.007045333584149678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,6144,0.021581333544519212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,4096,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,4096,0.07858310805426703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,4096,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,512,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,4096,0.02723466687732273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,51200,0.11400800281100804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,32,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,16384,0.6712026596069336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,512,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,6144,0.029880887932247583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,4096,0.021342222889264423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,6144,0.02237066626548767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,128,0.030332446098327637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,51200,0.05276799864239163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,16384,0.05403822329309252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,4096,0.030249777767393324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,3072,0.020227554771635268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,1024,0.008821333448092142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,128,0.037194665935304426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,6144,0.03943999939494663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,8192,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,4096,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,2048,0.040115555127461754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,8192,0.09974666436513265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,1024,0.016532444291644625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,768,0.01852799952030182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,16384,0.020706666840447318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,2560,0.010939555863539377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,3584,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,4096,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,6144,0.07254488600624932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,1024,0.18517066372765434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,1024,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,128,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,2560,0.021184888150956895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,12288,0.21647822856903076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,10240,0.21192355950673422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,768,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,512,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,3072,0.020780444145202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,32,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,12288,0.05522133244408501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,5120,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,768,0.09336266915003459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,51200,0.17900889449649385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,2048,0.04650933212704129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,8192,0.10935999949773152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,16384,0.07567733526229858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,6144,0.1441013283199734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,10240,0.07486133442984687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,128,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,6144,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,768,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,8192,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,3072,0.019225777851210702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,3584,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,10240,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,4096,0.5615493456522623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,768,0.00646666685740153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,3584,0.0074417781498697065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,10240,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,3072,0.029631111356947158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,6144,0.013040888640615674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,64,0.005664000080691443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,4096,0.16489599810706243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,256,0.025751110580232408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,10240,0.014733332726690503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,2560,0.023971555961502924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,10240,0.03293066554599338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,128,0.045420444673962064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,3072,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,10240,0.02993333339691162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,51200,0.19301511181725395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,32,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,768,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,4096,0.013401777380042605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,768,0.019709333777427673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,7168,0.018184888694021437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,7168,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,12288,0.06787911388609144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,1024,0.10127200020684136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,2560,0.030159112479951646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,7168,0.00889066689544254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,7168,0.15953511661953396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,2048,0.020265777905782063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,5120,0.026750221848487854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,6144,0.3563235600789388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,1024,0.015493333339691162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,512,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,10240,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,2048,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,2048,0.016977777083714802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,1024,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,64,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,51200,0.038035555018319026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,6144,0.00942666663063897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,8192,0.042021334171295166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,6144,0.06563733021418254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,2560,0.016933333542611863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,12288,0.024480000138282776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,128,0.0068808889223469635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,65536,0.07421066363652547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,768,0.01865511139233907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,512,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,7168,0.01607022186120351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,16384,0.2591324382358127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,512,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,65536,1.6346382564968531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,8192,0.021705778108702764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,512,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,512,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,51200,0.04008622301949395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,12288,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,51200,0.1375982231563992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,2048,0.10228177573945786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,12288,0.02812888887193468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,128,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,7168,0.015793777174419828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,10240,0.017853332890404593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,4096,0.014428445034556918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,128,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,3584,0.0232924438185162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,1024,0.0032239999208185407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,10240,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,256,0.009725333087974125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,512,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,1024,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,2048,0.0042257776690853965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,7168,0.2759608957502577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,2560,0.10560088687472874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,2560,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,1024,0.017821333474583097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,2048,2.633756425645616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,5120,0.03783555494414435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,65536,0.3697235584259033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,64,0.019759999381171334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,256,0.02626666592227088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,32,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,128,0.004229333251714706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,6144,0.09026133351855808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,3072,0.41650666130913633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,512,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,4096,0.017796445224020217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,10240,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,4096,0.06751733356051974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,51200,1.055956416659885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,5120,0.11716088983747695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,2560,0.019375110665957134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,10240,0.0181404451529185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,256,0.008351999852392409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,32,0.0019244444039132858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,1024,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,8192,0.015597333510716757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,64,0.0058355554938316345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,128,0.0019217777169413036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,4096,0.03943822118971083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,10240,0.04964088731341892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,2048,0.01239999963177575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,5120,0.008222222328186035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,10240,0.028423110644022625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,10240,0.0867982241842482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,128,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,16384,0.5094079971313477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,256,0.016216889023780823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,768,0.00703111125363244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,32,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,8192,0.009887110855844285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,256,0.03906133439805772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,16384,0.03808177842034234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,4096,0.010330666270520952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,10240,0.42418045467800564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,7168,0.1266311142179701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,1024,0.022591999835438196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,5120,0.04057866666052077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,256,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,10240,0.03580800029966567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,10240,0.049599998527103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,128,0.022058667408095464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,64,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,6144,0.0871466663148668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,8192,0.16650844944847956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,64,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,3584,0.01238044434123569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,3584,0.02528800070285797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,4096,0.022967999180157978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,10240,0.016352888610627916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,65536,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,128,0.43729599316914874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,8192,0.09557777643203735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,6144,0.017263111140992906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,4096,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,3584,0.15486310587988958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,5120,0.019017777509159513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,1536,0.059290667374928795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,65536,0.09195644325680202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,768,0.06902844376034208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,256,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,16384,0.9247670703464084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,2560,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,5120,0.01182933317290412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,2560,0.02611200014750163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,768,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,12288,0.017272000511487324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,51200,1.614540417989095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,6144,0.03076710965898302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,32,0.01089422239197625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,768,0.023195554812749226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,3072,0.07611199882295397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,3584,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,2560,0.013567999833159976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,51200,0.04905866583188375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,6144,0.04787910977999369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,768,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,4096,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,5120,0.01921866668595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,2560,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,16384,0.015899555550681222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,8192,0.01293155550956726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,128,0.09645955430136786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,1024,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,2048,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,768,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,1536,0.027748444014125403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,8192,0.026437333888477747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,6144,0.009376889301670922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,51200,0.06565066840913561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,16384,0.20532355043623182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,128,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,128,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,5120,0.9295209248860677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,512,0.007363555332024892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,768,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,5120,0.009170666337013245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,4096,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,768,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,6144,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,65536,0.08478844165802002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,1536,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,12288,0.01807733376820882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,51200,0.40970577134026426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,768,0.02532533307870229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,1024,0.007706666986147563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,3072,0.8932942284478081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,768,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,6144,0.026400888959566753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,10240,0.008443555898136562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,8192,0.059623108969794385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,512,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,64,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,256,0.01185955521133211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,4096,0.16346577803293863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,6144,0.011216888825098673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,2048,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,51200,0.3595102098253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,12288,0.07967822419272529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,12288,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,65536,0.20954667197333443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,12288,0.054935110939873584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,1024,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,128,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,8192,0.0388302206993103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,16384,0.048151999711990356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,256,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,3584,0.02275733318593767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,4096,0.17118222183651396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,65536,0.06421155399746366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,3072,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,256,0.010454222559928894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,4096,0.019471999671724107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,768,0.003165333428316646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,7168,0.02198222279548645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,7168,0.140630218717787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,512,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,10240,0.4346284336513943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,6144,0.02299733294381036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,5120,0.06124355395634969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,2560,0.0466648903157976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,3584,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,32,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,5120,0.02407022151682112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,5120,0.2190799978044298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,2560,0.0817155573103163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,5120,0.013079110946920184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,32,0.0022035555707083806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,4096,0.007745777567227681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,2048,0.0692053304778205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,51200,0.042209777567121715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,768,0.022890667120615642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,65536,0.1439288854598999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,768,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,4096,0.060518225034077965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,6144,0.020849777592553034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,5120,0.008866666919655269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,4096,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,4096,0.023057777020666335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,768,0.021947556071811255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,8192,0.09865333636601765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,512,0.010328888893127441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,10240,0.015051555302408008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,7168,0.021976000732845728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,2560,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,51200,0.29212623172336155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,3072,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,2048,0.012626666989591388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,3072,0.026398221651713055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,2560,0.007918222082985772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,16384,0.12693956163194445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,128,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,10240,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,51200,0.06314488914277819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,768,0.0074924445814556545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,12288,0.020964443683624268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,4096,0.018061333232455783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,2048,0.012503999802801343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,64,0.007002666592597961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,5120,0.21520000033908418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,1536,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,6144,0.07845511039098103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,3072,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,1536,0.019038221902317472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,51200,0.05378577775425381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,3584,0.04391733474201626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,2560,0.14377333058251276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,2560,0.07490311066309611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,256,0.08627377616034614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,8192,0.04157955447832743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,2048,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,4096,0.011031111081441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,32,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,1024,0.003987555702527364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,2048,0.022776888476477727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,1024,0.0064026667839950984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,512,0.016716443830066256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,512,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,6144,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,12288,0.039368887742360435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,6144,0.01297688898113039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,12288,0.06018755833307902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,2560,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,8192,0.0199626667631997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,256,0.08922133180830215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,1536,0.031743112537595965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,12288,0.053038222922219165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,7168,0.10333689053853352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,64,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,2560,0.031785776217778526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,3584,0.019146665930747986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,2560,0.014607111612955729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,51200,2.001456790500217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,8192,0.03163289030392965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,256,0.06299999687406752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,512,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,4096,0.03979644510481093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,128,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,1536,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,7168,0.012832889126406776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,256,0.03722666700681051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,7168,0.1652497715420193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,65536,0.9605315526326498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,512,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,51200,2.775127198961046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,12288,0.1132311158710056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,8192,0.05177066723505656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,51200,0.10564444462458293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,7168,0.03141333328353034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,1024,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,1536,0.0063324446479479475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,8192,0.06518310970730253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,2048,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,8192,0.019733332925372653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,12288,0.01793955597612593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,10240,0.018162666095627677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,256,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,1536,0.0989697774251302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,768,0.01676177812947167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,512,0.02167911165290409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,768,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,3584,0.019796444310082328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,768,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,16384,0.04038133223851522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,16384,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,8192,0.15035022629631892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,1024,0.13557867209116617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,64,0.0033760000434186724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,3584,0.047761778036753334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,2560,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,2048,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,5120,0.07154577970504761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,4096,0.008734222087595198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,65536,2.0807367960611978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,256,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,6144,0.16512266794840494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,12288,0.04946488804287381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,512,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,5120,0.022250667214393616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,256,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,768,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,7168,0.03538044293721517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,2048,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,7168,0.0967751079135471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,2560,0.02853333287768894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,4096,0.03222489025857713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,5120,0.2140773269865248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,3072,0.07181244426303439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,4096,0.018093332648277283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,1024,0.04389955600102743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,1024,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,10240,0.05179199907514784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,1024,0.018559111489189994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,8192,0.021960000197092693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,16384,0.10823466380437215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,8192,0.02232444451914893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,2560,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,512,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,1536,0.004446222136418025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,12288,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,16384,0.03525066706869338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,8192,0.0595760014322069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,5120,0.03192266821861267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,2560,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,1024,0.3880515628390842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,768,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,1024,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,2560,0.02494488822089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,1536,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,4096,0.0993173321088155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,1024,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,768,0.011028444601429833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,512,0.021436444587177698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,12288,0.3347013261583116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,4096,0.0906791090965271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,256,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,4096,0.12353066603342693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,10240,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,64,0.027265777190526325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,2048,0.010653333531485664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,768,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,2048,0.026025777061780293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,4096,0.014299554957283868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,51200,0.15866488880581325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,32,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,3072,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,7168,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,12288,0.06732000244988336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,10240,0.03068355719248454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,65536,0.03562488820817735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,64,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,51200,1.9364746941460504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,2048,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,3072,0.1478088829252455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,512,0.004061333421203825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,6144,0.03471022182040744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,4096,0.03275911013285319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,1024,0.038089778688218855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,2048,0.014752000570297241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,5120,0.0277511113219791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,16384,0.5212097697787815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,7168,0.1365155511432224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,8192,0.18493333127763537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,6144,0.07793511284722222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,256,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,5120,0.11436711417304145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,6144,0.11412889427608913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,12288,0.13319467173682317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,32,0.0018248889181349014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,2560,0.012816000315878125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,12288,0.05274933245446947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,16384,0.026012443833880957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,32,0.006019555446174409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,65536,0.5169217851426866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,1024,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,1024,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,256,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,1024,0.01711466742886437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,2048,0.020202666521072388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,3072,0.07273688581254747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,12288,0.029672887590196397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,512,0.0177182224061754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,128,0.006103111224042044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,16384,0.03416088885731167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,51200,0.4270213445027669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,128,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,3584,0.006892444358931647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,4096,0.07460533248053657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,3584,0.06139466497633192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,5120,0.03824355536037021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,10240,0.5185395346747504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,1024,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,12288,0.04258222381273905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,51200,0.133513781759474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,10240,0.02342844506104787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,1024,0.01757777730623881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,4096,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,32,0.00564533347884814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,6144,0.10227644443511963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,3072,0.010597333312034607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,7168,0.16644267241160074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,6144,0.021977777282396953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,2560,0.020062221421135795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,16384,0.2766639921400282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,256,0.013397333522637686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,768,0.0032746667663256326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,1536,0.04261777798334757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,1536,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,1536,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,512,0.01694222291310628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,32,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,1536,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,256,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,1536,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,3072,0.00904444439543618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,16384,0.4626764191521539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,8192,0.03898577888806661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,1536,0.026382222771644592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,3584,0.016933333542611863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,1024,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,10240,0.02305511136849721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,256,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,128,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,1024,0.053252445326911084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,768,0.00925244473748737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,12288,0.021288888321982488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,16384,0.21926400396558973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,65536,0.05630844169192844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,7168,0.013697778185208639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,65536,0.37025067541334367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,16384,0.07948622438642713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,768,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,256,0.009383111364311641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,3072,0.010320888625250923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,64,0.03550133440229628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,768,0.005981333139869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,65536,0.08762222528457642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,1536,0.026994667119450037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,6144,0.01847022275129954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,8192,0.04006400042110019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,1024,0.060806221432156034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,128,0.09552888737784491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,5120,0.01759822169939677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,1024,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,5120,0.03404266635576884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,7168,0.08009244336022271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,512,0.007463999920421177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,2560,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,5120,0.04977955420811971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,3584,0.006783999916579988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,2048,0.007320000065697565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,12288,0.18284177780151367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,7168,0.15294044547610813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,51200,0.43667999903361004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,3072,0.02367111047108968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,1024,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,7168,0.11736355887518989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,32,0.008263111114501953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,64,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,7168,0.023964444796244305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,256,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,768,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,12288,0.01315111087428199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,51200,1.078463978237576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,10240,0.022964444425370958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,51200,1.6807840135362413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,65536,0.06677599747975667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,768,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,1536,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,3584,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,3072,0.02608355548646715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,16384,0.01982222166326311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,2048,0.024040000306235418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,1536,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,32,0.03363377849260966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,1024,0.014869333969222175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,768,0.006184000107977126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,2048,0.13149155510796443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,16384,0.178984006245931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,12288,0.048784001006020435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,7168,0.6247270901997884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,256,0.13568711280822754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,6144,0.05547377798292372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,51200,0.09758044613732232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,2048,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,10240,0.05720711416668362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,10240,0.05527199970351326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,32,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,10240,0.04134844409094916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,7168,0.18044088946448433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,7168,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,10240,0.2654399871826172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,7168,0.014859555496109856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,10240,0.012749333348539142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,256,0.03605511122279697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,256,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,2048,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,64,0.008700444466537898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,768,0.22926221953497994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,1536,0.0041759999261962045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,65536,0.07606666617923312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,2048,0.041887111133999295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,128,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,512,0.22900801234775117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,1536,0.02091644373204973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,16384,0.015680889288584392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,3072,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,3072,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,256,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,65536,0.06684799989064534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,51200,15.034842597113716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,16384,0.10281778044170803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,512,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,51200,0.055700441201527916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,65536,3.524492475721571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,65536,0.2168711158964369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,3072,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,65536,0.025621333056026038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,10240,0.162280003229777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,5120,0.02023555503951179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,128,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,65536,0.1969484488169352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,8192,0.0445493327246772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,3072,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,64,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,65536,0.11686755551232232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,2560,0.024855110380384658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,65536,0.5325688785976833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,32,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,512,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,64,0.00600355573826366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,65536,0.20626844300164116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,128,0.0021137777302000257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,512,0.006716444260544247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,7168,0.01332800007528729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,6144,0.020475554797384474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,256,0.019646222392718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,32,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,512,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,1024,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,2560,0.011970666547616323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,51200,0.19026845031314424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,512,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,5120,0.07418577538596259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,65536,0.06286844280030993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,12288,0.033977776765823364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,6144,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,4096,0.010966222319338055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,3584,0.03674044542842441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,3072,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,16384,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,4096,0.08111644453472562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,1536,0.03393155667516921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,1536,0.008215111162927415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,12288,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,1024,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,128,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,51200,1.680574205186632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,256,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,3072,0.018904000520706177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,8192,0.015399111641777886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,2048,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,3072,0.02742133372359806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,16384,0.023375999596383836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,1536,0.010423111418883005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,768,1.1572106679280598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,3072,0.011988444460762871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,512,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,12288,0.016608888904253643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,51200,0.10473866595162286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,768,0.0058293334311909145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,65536,1.812493430243598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,256,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,7168,0.47353331247965497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,51200,0.10643466975953843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,5120,0.08695022265116374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,10240,0.032334221733940974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,10240,0.22133421897888184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,5120,0.10997866921954685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,16384,0.021521778570281133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,768,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,64,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,4096,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,65536,0.30131732092963326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,3072,0.004586666822433472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,12288,0.057928886678483754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,2048,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,1536,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,1536,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,3072,0.019267555740144517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,65536,0.10077777836057876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,8192,0.04555911156866285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,5120,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,65536,3.9732924567328562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,128,0.027543110979927912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,2048,0.04023377762900458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,6144,0.051687111457188926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,16384,0.2785937786102295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,16384,0.11987822585635716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,65536,0.1083279980553521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,3072,0.016509332590632968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,6144,0.035016000270843506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,12288,0.03270577722125583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,512,0.23197955555386016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,12288,0.02865866820017497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,1536,0.01703288820054796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,2560,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,10240,0.007687999970383114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,1024,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,128,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,768,0.010655110908879174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,7168,0.012625777886973487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,8192,0.00962755580743154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,51200,0.4092000060611301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,1536,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,7168,0.09292266766230266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,16384,0.09920089112387763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,256,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,1024,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,128,0.10900977585050796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,65536,4.116950141059028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,1536,0.03329599897066752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,32,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,5120,0.06743022468354967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,768,0.017422222428851657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,16384,0.706350220574273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,1536,0.030641777647866145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,16384,0.038386666112475924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,1536,0.4747040006849501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,64,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,3584,0.05207733313242594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,2048,0.02193333374129401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,16384,0.037772443559434675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,16384,0.03849422269397312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,65536,0.17819289366404215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,51200,0.9813146591186523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,12288,0.08677155441708034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,16384,0.23691288630167642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,6144,0.033518221643235944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,5120,0.16601955890655518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,2048,0.11405511697133382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,10240,0.0440008905198839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,51200,0.07547466622458564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,768,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,12288,0.019117333822780185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,16384,1.0993066363864474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,8192,0.3696906566619873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,8192,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,51200,0.5789662467108833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,128,0.01787911189926995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,768,0.02199555602338579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,4096,0.010388444695207808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,256,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,512,0.11622044775221084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,65536,0.024163555767801072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,2048,0.00741244438621733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,6144,0.02132355504565769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,3072,0.06230755647023519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,12288,0.07835821972952949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,16384,0.020669332808918424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,6144,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,3584,0.008384000096056197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,5120,0.1166728867424859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,128,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,6144,0.027691556347741023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,5120,0.021923555268181696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,6144,0.009723555710580613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,8192,0.012054222325483957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,6144,0.07104888889524671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,8192,1.6168159908718531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,8192,0.20383021566602919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,3072,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,6144,0.02643022272321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,51200,0.6181813346015083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,12288,0.1437022156185574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,768,0.018138666947682697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,12288,0.1437280045615302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,3072,0.008202667037645975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,2560,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,16384,0.03738044367896186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,1024,0.016146666473812528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,32,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,16384,0.2829048898484972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,768,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,3584,0.016372443901167977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,6144,0.12754044267866346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,51200,0.1593288845486111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,12288,0.22911379072401258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,2560,0.010494222243626913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,7168,0.1613866620593601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,7168,0.025252444876564875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,2048,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,12288,0.04113599989149306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,12288,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,3584,0.1289715501997206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,512,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,3584,0.030934221214718286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,8192,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,1024,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,5120,0.014525334040323893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,768,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,12288,0.029520889123280842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,16384,0.06236977709664238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,3072,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,3584,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,3584,0.03191288974550035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,7168,0.012206222448084088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,2560,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,10240,0.039098666773902044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,12288,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,2560,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,128,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,3584,0.09327289130952622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,51200,0.5111626519097222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,12288,0.047853334082497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,2560,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,16384,0.27762844827440053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,16384,0.03381955623626709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,1536,0.02221777704026964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,5120,0.011775111158688864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,4096,0.021734222769737244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,32,0.009880888793203566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,2560,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,8192,0.030524445904625788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,3584,0.2723920080396864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,6144,0.3654168976677789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,1536,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,8192,0.031670222679773964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,6144,0.31490222613016766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,10240,0.10198844141430324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,768,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,4096,0.07354844278759427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,7168,0.4289262294769287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,10240,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,5120,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,8192,0.014849777022997538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,10240,0.026760000321600173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,7168,0.515699545542399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,256,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,768,0.12430755297342937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,7168,0.01146933353609509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,3072,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,3584,0.03423555692036947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,1536,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,6144,0.0324942237801022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,12288,0.04274310999446445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,256,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,10240,0.045343110958735146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,2048,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,12288,0.032808889945348106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,7168,0.013741333451535968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,512,0.12551021575927734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,5120,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,6144,0.019139554765489366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,3584,0.07455288701587252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,6144,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,3584,0.043009777863820396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,65536,0.1706248919169108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,4096,0.0314062237739563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,3584,0.028745777077145044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,8192,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,6144,0.10616888602574666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,51200,0.08406311273574829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,3584,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,256,0.09275555610656738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,5120,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,768,0.018904889623324077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,8192,0.0137155552705129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,1536,0.03819200065400865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,2560,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,10240,0.012695999609099494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,2048,0.011075555450386472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,4096,0.011457777685589261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,4096,0.023461333579487268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,5120,0.02639644510216183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,51200,1.6783093346489801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,3584,0.027845333019892376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,51200,0.1400044494205051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,768,0.012809777425395118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,5120,0.38881421089172363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,3072,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,10240,0.04381600022315979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,1536,0.08986488978068034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,64,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,256,0.005664000080691443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,5120,0.010451555252075195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,65536,0.5244853231641982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,10240,0.06564355558819242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,16384,0.06455377737681071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,128,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,3584,0.04833066794607374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,768,0.0038737778862317405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,7168,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,3584,0.10968266593085395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,10240,0.029345777299669053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,3072,0.05041866501172384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,3072,0.7169511053297254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,3584,0.7507715755038791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,64,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,256,0.045787556303872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,51200,1.0186826917860243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,8192,0.017451556192504037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,1024,0.004305777864323722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,1536,0.004670222186379963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,8192,0.08592444658279419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,2560,1.549085299173991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,6144,0.016467556357383728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,7168,0.025770665870772466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,65536,3.195737838745117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,2560,0.023670223024156358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,768,0.007830222447713217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,4096,0.021048888564109802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,3584,0.047225776645872325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,3584,0.12387911478678386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,12288,0.016809778081046212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,65536,0.05712888638178507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,12288,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,1024,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,32,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,7168,0.02773688899146186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,2560,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,16384,0.048904889159732394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,6144,0.033372445238961115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,16384,0.6606204244825575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,2048,0.003903999924659729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,1536,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,3072,0.014694222145610385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,7168,0.05168266759978401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,2560,0.00721599989467197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,32,0.013205333716339536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,768,0.02272355556488037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,64,0.036297778288523354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,256,0.013287111288971372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,6144,0.11914400259653728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,128,0.005401777724424998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,16384,0.0314408904976315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,16384,0.010963555839326648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,6144,0.020988444487253826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,7168,0.01108977778090371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,3584,0.022614222433831956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,128,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,768,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,65536,4.899400075276692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,1024,0.02283199959331089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,3072,0.020124443703227572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,65536,0.10018133454852635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,3072,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,5120,0.033286223808924355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,2048,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,65536,2.457728068033854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,8192,0.0110880004035102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,2560,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,2048,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,3584,0.008049777812427944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,512,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,12288,0.018562666243977018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,51200,0.09626933601167466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,16384,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,5120,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,1024,0.01385244478782018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,2048,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,7168,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,64,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,7168,0.065884444448683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,64,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,64,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,768,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,256,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,128,0.0170462214284473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,768,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,10240,0.027403554982609216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,16384,0.032251556714375816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,512,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,6144,0.026885333988401625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,51200,2.153779559665256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,1536,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,32,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,1536,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,12288,0.030420445733600195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,3072,0.008560000194443597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,51200,0.06218666500515408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,4096,0.08134933312733968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,256,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,1536,0.017614222235149808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,128,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,2048,0.033344000577926636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,5120,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,128,0.026150221625963848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,256,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,1024,0.020617778102556866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,128,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,3584,0.16625511646270752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,1536,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,128,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,1024,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,2048,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,4096,0.03362933463520474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,16384,0.03917422228389316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,4096,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,12288,0.2697964509328206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,7168,0.014559111661381192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,8192,0.018144889010323417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,12288,0.029719110992219713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,2048,0.05597244368659126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,5120,0.026203556193245783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,1536,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,3584,0.014104000396198697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,512,0.017470222380426195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,2560,0.01145066652033064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,3072,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,512,0.12086844444274902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,12288,0.02290311124589708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,256,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,10240,0.01959377858373854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,3072,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,10240,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,4096,0.0341288877858056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,8192,0.048743999666637845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,1536,0.02015022271209293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,512,0.008436444732877944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,3584,0.01035822182893753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,12288,0.03160533308982849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,3072,0.007432888779375289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,768,0.01714933415253957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,3584,0.04347733325428433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,256,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,7168,0.04255377915170458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,7168,0.03051466743151347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,16384,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,12288,0.10025422440634833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,4096,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,32,0.011761777930789523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,2048,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,7168,0.02659822172588772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,512,0.0032142222755485107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,256,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,5120,0.012256888879670037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,65536,0.5613599883185493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,4096,0.01648177703221639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,4096,0.01796266602145301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,5120,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,51200,0.051114665137396924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,512,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,10240,0.6088026894463433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,5120,0.04686844348907471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,3584,0.021838222940762837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,51200,0.33205689324273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,256,0.013259555730554791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,8192,0.03393422232733832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,32,0.10575733582178752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,256,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,12288,0.03643911083539327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,4096,0.027223111854659185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,2048,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,3072,0.027344889110989038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,10240,0.061675555176205106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,12288,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,768,0.01700444519519806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,512,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,1536,0.011807999677128263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,1536,0.1676044464111328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,8192,0.049125333627065025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,1024,0.03002133303218418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,1536,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,128,0.019606222709019978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,3584,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,3584,0.2730791038937039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,768,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,8192,0.024663999676704407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,8192,0.012049777640236748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,256,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,65536,0.15419822269015843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,768,0.003253333270549774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,64,0.022665777140193518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,12288,0.019664888580640156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,16384,0.10334222184287177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,12288,0.04264444443914625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,3072,0.02628622286849552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,1024,0.0033662221911880705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,128,0.024663999676704407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,256,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,768,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,5120,0.011723555624485016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,128,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,6144,0.013748444616794586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,5120,0.3112444347805447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,3072,0.011363555159833698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,2048,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,10240,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,768,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,768,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,3584,0.007784888976150089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,256,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,8192,0.02897777822282579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,128,0.016220443778567843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,8192,0.19051110744476318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,10240,0.0810826685693529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,4096,0.19021866056654188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,6144,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,12288,0.016840888394249808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,128,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,768,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,5120,0.026353778110610113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,4096,0.7202017572191027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,3584,0.02635555631584591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,16384,0.30572443538241917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,1536,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,256,0.017059556312031217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,8192,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,2560,0.02535199953450097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,256,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,64,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,2048,0.06970489025115967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,128,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,4096,0.039854221873813205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,6144,0.009635555247465769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,4096,0.24356089697943795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,7168,0.01090666651725769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,128,0.015613334046469795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,7168,0.01218577805492613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,16384,0.141975998878479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,1024,0.016332444217469957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,256,0.01851022243499756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,4096,0.012135999898115793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,8192,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,2560,0.01549600064754486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,65536,0.08825333250893487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,32,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,6144,0.057633777459462486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,1536,0.016604445046848722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,1024,0.008442666795518663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,2560,0.027888887458377417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,5120,0.04553066690762838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,8192,0.037816001309288874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,512,0.14160445001390246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,768,0.01893422173129188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,4096,0.1123502254486084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,12288,1.038079155815972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,3072,0.007151111132568783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,5120,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,6144,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,64,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,2560,0.02162133322821723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,6144,0.03662755423121982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,3072,0.03622311022546556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,256,0.005423111220200856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,10240,0.03279822071393331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,2560,0.01586577792962392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,7168,0.3864160113864475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,6144,0.03321866525544061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,3072,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,12288,0.0344835552904341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,65536,0.2377928892771403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,3072,0.04866488774617513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,51200,0.09179911348554824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,256,0.021242666575643752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,65536,0.023964444796244305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,12288,0.020371556282043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,128,0.0038453332251972626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,3584,0.12889688544803196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,2048,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,256,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,3584,0.04864533411131965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,16384,0.043473776843812734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,12288,0.10018133454852635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,1024,0.005705777969625261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,8192,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,128,0.011792889071835412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,65536,0.05764444669087728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,7168,0.013782222237851886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,12288,0.03258400162061056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,3584,0.018238221605618794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,51200,1.19468445248074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,12288,0.0420720014307234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,8192,0.10005955563651191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,512,0.01724177764521705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,768,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,7168,2.5884710947672525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,2560,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,768,0.04467555549409655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,256,0.052384889788097806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,2048,0.06906133227878146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,5120,0.023426666855812073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,2048,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,12288,0.1935422288046943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,768,0.023740443918440077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,8192,0.012432888978057437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,51200,0.13688800070020887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,65536,0.06814755333794488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,16384,0.1320293347040812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,10240,0.25915466414557564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,768,0.005985777825117111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,3584,0.04935822221967909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,51200,0.4369057814280192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,512,0.005646222167544895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,32,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,3072,0.026152888933817547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,12288,0.02407199972205692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,512,0.02854666776127285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,256,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,8192,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,51200,0.05135022269354927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,6144,0.01810933318403032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,5120,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,128,0.05327555537223816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,7168,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,512,0.013129777378506131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,10240,0.06114933225843641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,65536,1.5222862031724718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,16384,0.01791288952032725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,4096,0.0807173318333096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,51200,0.12143466207716201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,1024,0.003479111111826367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,128,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,768,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,256,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,2560,0.9842132992214627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,128,0.028127110666698877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,4096,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,3072,0.04397155510054695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,2560,0.012054222325483957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,768,0.018216000662909616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,12288,0.03647911217477586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,8192,0.09251288572947185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,51200,2.3928070068359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,4096,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,6144,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,12288,0.0244195560614268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,32,0.005653333332803514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,4096,0.02051822178893619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,2560,0.02277155551645491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,1536,0.0074346669846110875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,2048,0.7132382392883301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,4096,0.060864885648091636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,5120,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,7168,0.008427555362383524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,6144,0.2117013269000583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,6144,0.01757066614098019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,7168,0.011986666255527072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,2560,0.024618667032983568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,512,0.10868355300691392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,1024,0.03477422065205044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,3584,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,12288,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,6144,0.024892444411913555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,6144,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,4096,0.033208890093697443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,5120,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,3584,0.08441511127683853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,5120,0.029915554655922785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,512,0.0321395562754737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,768,0.005719111197524601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,5120,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,65536,0.07176711161931355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,5120,0.03933599922392104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,6144,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,7168,0.010275555153687796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,1024,0.013391111459996967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,65536,0.09860444731182522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,12288,0.036284443404939436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,65536,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,12288,0.013578666581047906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,2048,0.017157332764731515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,51200,0.05297155512703789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,256,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,3584,0.5491093529595269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,12288,0.2514897717369927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,256,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,65536,0.3266222212049696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,12288,0.04961599906285604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,128,0.0018275556051068837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,3584,0.004239999999602635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,256,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,3584,0.09575021929211086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,32,0.005437333136796951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,10240,0.02659822172588772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,10240,0.018423111902342904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,12288,0.03507644600338406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,65536,0.9559644063313802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,65536,0.1672888861762153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,2048,0.027501333090994094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,2560,0.09887199931674534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,64,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,1536,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,1536,0.011738667057620155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,128,0.18432354927062988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,32,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,6144,0.03241599930657281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,8192,0.05087022317780388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,10240,0.021635555558734473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,768,0.01831999917825063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,51200,0.048723555273479886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,3072,0.0703315536181132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,1024,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,512,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,2560,0.07438044415579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,7168,0.06722578075197008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,16384,0.037351999017927386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,5120,0.376035557852851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,512,0.00849599970711602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,5120,0.07293244202931722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,64,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,256,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,128,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,128,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,3584,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,3584,0.016864889197879367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,4096,0.023335999912685815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,8192,0.08661333057615493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,1024,0.06258489025963677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,5120,0.010103999740547603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,1536,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,5120,0.031144890520307753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,4096,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,32,0.00425866660144594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,7168,0.2064053350024753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,2560,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,16384,0.02257777750492096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,6144,0.02180533276663886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,16384,0.008348444269763099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,64,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,3072,0.01475288967291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,10240,0.012297777665985955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,2048,0.041103111373053655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,12288,0.07663022147284614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,65536,0.08241955439249675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,2560,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,2560,0.0403404434521993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,3072,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,768,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,2560,0.03303022185961405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,256,0.011006222003036074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,3584,0.5416408644782172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,128,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,16384,0.10689777798122829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,512,0.01726222203837501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,51200,0.4424791071150038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,3072,0.017489777670966256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,7168,0.02027466727627648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,256,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,6144,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,8192,0.027074666486846075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,64,0.011344888971911537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,65536,0.41966933674282497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,128,0.20242488384246826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,7168,0.12023999955919053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,12288,0.018044445249769423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,8192,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,2048,0.0698782205581665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,256,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,256,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,64,0.006031999985376994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,10240,0.012103999654452005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,256,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,3072,0.023944000403086346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,3072,0.015789333317014907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,1024,0.006184000107977126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,5120,0.019129777948061626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,768,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,128,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,3072,0.01367555558681488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,8192,0.08601155545976426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,4096,0.011117333339320289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,6144,0.032680888970692955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,8192,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,768,0.01754844519827101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,51200,0.16792355643378365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,512,0.033352000845803156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,7168,0.053363555007510714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,256,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,512,0.015543111496501498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,10240,2.9556496938069663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,2560,0.03336266676584879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,512,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,7168,1.210788408915202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,64,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,32,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,64,0.009406222237481011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,4096,0.07783289088143243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,4096,0.05848444170422024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,16384,0.3400871223873562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,64,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,2048,0.01904444396495819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,2048,0.06561866733762953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,51200,0.1079288919766744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,12288,0.42724355061848956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,6144,0.01959733333852556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,65536,0.08732977840635513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,7168,0.027016000615225896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,2560,0.036235554350747004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,65536,0.28643467691209584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,1024,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,65536,0.12077511681450738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,64,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,5120,0.4867146809895833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,65536,0.16794310675726998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,1024,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,6144,0.022296888960732356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,128,0.03148622314135233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,768,0.02812533246146308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,128,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,12288,0.04589066571659512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,8192,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,128,0.014836443795098199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,7168,0.038430223862330117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,3584,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,768,0.0027964444210131965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,1024,0.01647111111217075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,768,0.002452444492114915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,64,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,65536,0.4634382459852431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,768,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,16384,0.04135911001099481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,7168,0.06037244531843397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,16384,0.03360177742110358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,1536,0.10088000032636855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,16384,0.055736886130438916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,1536,0.003957333250178231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,512,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,10240,0.08634666601816814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,2560,0.03199733297030131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,5120,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,12288,0.03066844410366482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,256,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,1024,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,2560,0.047967112726635404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,128,0.008858666651778752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,2560,0.05080266793568929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,10240,0.045696000258127846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,65536,0.06818311081992255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,3584,0.025822222232818604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,5120,0.015379556351237826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,6144,0.05572355455822415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,64,0.0017928888814316855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,16384,0.018518222702874076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,12288,0.21036977238125273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,12288,0.06985066996680366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,12288,0.05693155527114868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,64,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,6144,0.023077333966890972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,4096,0.02142488956451416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,51200,0.06673333379957411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,16384,0.057143999470604785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,10240,0.031503111124038696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,6144,0.023917333947287664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,256,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,768,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,51200,0.08324266804589166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,51200,1.122908486260308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,5120,0.015769778026474845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,3072,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,10240,0.7177297804090711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,7168,2.6996578640407987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,2560,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,1536,0.02365422248840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,51200,0.20986666944291857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,1536,0.033412443266974554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,10240,0.04110755523045858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,8192,0.016146666473812528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,12288,0.028938667641745672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,3584,0.011681777735551199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,1024,0.01333688861793942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,2560,0.02049688829316033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,1024,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,65536,4.31875483194987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,4096,0.13235644499460855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,256,0.011063999599880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,256,0.022600000103314716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,1024,0.022668444448047217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,51200,0.06675377819273207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,6144,0.07301688856548734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,65536,0.24736621644761828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,32,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,2048,0.02599999970859951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,8192,0.23108977741665312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,1024,0.007045333584149678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,12288,0.08195822106467353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,64,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,8192,0.07026577658123441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,128,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,16384,0.23548799090915254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,65536,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,256,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,768,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,3072,0.016778666112158034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,7168,0.0399893323580424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,51200,0.06608800093332927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,65536,0.2534968852996826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,8192,0.01866666641500261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,2560,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,1536,0.010110222631030613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,2048,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,128,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,65536,0.16864000426398384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,768,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,10240,0.05494133300251431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,128,0.014521777629852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,16384,0.33975201182895237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,16384,0.05459555652406481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,7168,0.7504675653245715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,1024,0.009972444838947719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,8192,0.03083289000723097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,10240,0.027795556518766616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,6144,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,3072,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,768,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,1536,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,2048,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,5120,0.016718222035302054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,6144,0.025153777665562097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,3072,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,1024,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,10240,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,1024,0.008887111312813228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,1536,0.022454222043355305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,2048,0.01755466726091173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,16384,0.024301333559883967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,1024,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,51200,0.21768089135487875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,512,0.018366222580273945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,3072,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,768,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,16384,0.016120000018013846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,2560,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,2048,0.01853511068556044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,12288,0.5643128818935818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,5120,0.058987557888031006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,65536,0.08106222417619494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,512,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,512,0.02806666824552748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,128,0.0052977779673205475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,512,0.09017866849899292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,768,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,1024,0.017543110582563613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,1536,0.01689511040846507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,32,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,3072,0.014262222581439547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,768,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,10240,0.09946399927139282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,5120,0.05117866728040907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,6144,0.04001599881384108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,5120,0.03277955452601115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,4096,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,7168,0.03603644503487481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,4096,0.013978666729397245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,16384,0.019708444674809773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,10240,0.02881511052449544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,1536,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,12288,0.01630311045381758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,3584,0.09006577730178833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,51200,0.1266311142179701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,16384,0.7255306773715549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,12288,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,5120,0.027797334724002417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,5120,0.04951999915970696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,10240,0.20286311043633354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,6144,0.05625244643953112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,32,0.001785777716173066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,32,0.0039048890272776284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,2560,0.08413155873616536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,5120,0.6335448688930935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,8192,0.26810399691263836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,2560,0.020974222156736586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,5120,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,6144,0.04688622223006355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,768,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,4096,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,32,0.021645334031846788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,2048,0.08347644408543904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,3584,0.07333866755167644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,7168,0.018574222922325134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,3072,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,2048,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,16384,0.1006951133410136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,16384,0.023383110761642456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,2560,0.007485333416197035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,12288,0.016002666619088914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,512,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,1536,0.29658577177259654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,3584,0.009555555880069733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,8192,0.01794933279355367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,512,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,4096,0.30713778071933323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,16384,0.059896886348724365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,4096,0.14069422086079916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,5120,0.009053332938088311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,256,0.01720266706413693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,3072,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,1024,0.013453333742088742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,5120,0.09210844172371759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,2560,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,4096,0.07710666788948907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,7168,0.04275822308328417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,2560,0.02756711178355747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,1024,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,2048,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,65536,0.12413244777255589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,12288,0.25451111793518066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,1536,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,51200,0.0671164459652371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,4096,0.07274755504396227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,2560,0.0810462236404419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,768,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,512,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,768,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,16384,0.11415555742051865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,65536,0.0870328876707289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,1024,0.03939111034075419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,5120,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,128,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,4096,0.009872000250551436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,10240,0.0794266661008199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,256,0.024861334098709956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,12288,0.023959999283154804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,1024,0.11803644233279759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,1536,0.029085334804322984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,12288,0.09959644741482204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,8192,0.7280933592054578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,51200,0.5814364221360948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,64,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,7168,0.03485155436727736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,7168,0.10190577639473809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,2560,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,3584,0.017937777770890128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,6144,0.018544000056054857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,16384,0.019522666931152344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,768,0.01680888897842831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,128,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,3072,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,16384,0.059911111990610756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,51200,0.06830844614240858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,12288,0.0362426671716902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,8192,0.28175467915005153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,64,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,7168,0.4189288881089952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,768,0.0225582222143809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,3072,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,128,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,256,0.03523111012246873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,5120,0.03284622232119242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,512,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,1536,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,8192,0.026383111874262493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,64,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,3072,0.03391644358634949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,1024,0.0873875551753574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,3584,0.04070577687687344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,4096,0.11343467235565186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,1024,0.010031999813185798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,16384,0.04514666729503208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,128,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,768,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,65536,5.962832980685764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,8192,0.0565519995159573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,5120,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,51200,0.058168000645107694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,128,0.006731555693679386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,51200,9.437634785970053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,7168,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,3584,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,10240,0.10623644457923041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,12288,0.8181377516852485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,12288,0.02103022237618764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,12288,0.1174515618218316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,2560,0.08995378017425537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,128,0.015434667468070984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,5120,0.04469511244032118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,3072,0.038183109627829656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,51200,15.741917928059896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,12288,0.026104888982243005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,6144,2.1199253929985895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,16384,0.012925333446926542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,3072,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,512,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,51200,0.06217066446940104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,8192,0.04714488983154297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,8192,0.01980799933274587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,7168,0.03454400102297465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,12288,0.10430755880143906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,64,0.0041857775714662345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,5120,0.03398488958676656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,1536,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,8192,0.12716889381408691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,2048,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,12288,0.014001778430408902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,1536,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,1024,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,7168,0.04287199841605293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,10240,0.3813733259836833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,65536,0.05753599935107761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,3072,0.014833778142929077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,5120,0.09303733375337388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,128,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,3072,0.0306968887646993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,3584,0.07843110958735149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,7168,0.0315324432320065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,12288,0.21849600474039713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,6144,0.5690053304036459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,8192,0.026870222555266485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,768,0.01350844403107961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,7168,0.007327111230956183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,65536,0.07699644565582275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,51200,0.11226399739583333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,2048,0.013560888667901358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,128,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,256,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,65536,0.18104888333214653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,12288,0.15769955846998426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,512,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,2560,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,512,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,128,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,1024,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,5120,0.03930400146378411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,6144,0.009751111268997192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,3072,0.021332444416152105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,8192,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,2560,0.010527999864684211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,1536,0.017858665850427415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,128,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,2048,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,1024,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,65536,0.06177600224812826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,7168,0.03678222166167365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,2048,0.036927110619015165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,3072,0.009211555951171452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,65536,1.5561173756917317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,4096,0.0181351105372111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,4096,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,2560,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,12288,0.04785155587726169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,4096,0.011361777782440186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,7168,0.10486133231057061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,768,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,1024,0.003212444484233856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,768,0.03572888837920295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,2048,0.013849777479966482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,4096,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,65536,1.3995075225830078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,8192,0.12977511352962917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,4096,0.058583113882276744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,12288,0.1989582247204251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,65536,0.07309510972764757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,12288,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,3584,0.013487111363146039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,1536,0.015210666590266757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,5120,0.019268444842762418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,8192,0.012689777546458773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,4096,0.07229866584142049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,768,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,6144,0.014004444082578024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,2048,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,3072,0.01641244524055057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,768,0.002584888910253843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,12288,0.27640265888637966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,768,0.31249332427978516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,2048,0.04327555497487386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,768,0.030471109681659277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,12288,1.4739946789211695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,65536,0.35489688979254824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,5120,0.037689778539869524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,3584,0.029855112234751385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,6144,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,256,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,256,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,256,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,512,0.018217777212460835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,12288,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,5120,0.0581644442346361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,2560,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,51200,0.06688355737262301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,1024,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,512,0.07366577784220378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,7168,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,64,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,768,0.02613688839806451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,7168,0.045765333705478244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,128,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,12288,0.0413697792424096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,65536,0.05609689156214396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,3072,0.035989334185918175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,5120,0.057117336326175265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,7168,0.027013333307372198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,16384,0.03389244609408908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,65536,2.2464887830946183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,12288,0.019490665859646268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,128,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,3584,0.020599111914634705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,1024,0.042325334416495435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,10240,0.031138665146297876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,3584,0.08263199859195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,10240,0.025664889150195654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,1024,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,7168,0.011422221859296164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,64,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,4096,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,3072,0.09270399808883667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,10240,0.028024001253975764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,1024,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,51200,1.497534222073025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,32,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,5120,0.02349688940578037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,10240,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,3584,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,7168,0.03823022378815545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,768,0.03008355696996053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,2048,0.3569093280368381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,2560,0.011208000282446543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,4096,0.0315048893292745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,3584,0.013981334037250943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,7168,0.023511999183230933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,768,0.031361778577168785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,2560,0.03418133325046963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,2560,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,5120,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,12288,0.10464711321724786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,128,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,32,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,64,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,65536,0.34377066294352215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,3584,0.03386133246951633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,4096,0.08806666400697495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,1536,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,1536,0.007330666813585494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,32,0.007177777588367462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,10240,0.3449733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,3584,0.006130666782458623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,8192,0.021740444832377966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,12288,0.03832622369130453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,65536,3.288792928059896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,65536,0.07119111220041911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,3584,1.357078234354655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,2048,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,65536,0.06566933128568861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,3584,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,12288,0.030844443374209937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,6144,0.18535910712348092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,10240,0.04605866803063286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,2048,0.030406223403082952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,12288,0.0818799999025133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,10240,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,256,0.0021928888228204516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,7168,0.01296266665061315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,3584,0.008734222087595198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,10240,0.01572800013754103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,512,0.13866578208075628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,16384,0.11403111616770427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,10240,0.5101431210835775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,4096,0.08709510829713608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,768,0.04589510957400004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,51200,0.1938906643125746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,4096,0.027619555592536926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,51200,0.28028978241814506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,16384,0.02407288882467482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,1024,0.025322667426533167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,1536,0.016905777984195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,7168,0.5071342256334093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,16384,0.03602844476699829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,8192,0.026379555463790894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,1024,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,4096,0.07202488846249051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,256,0.009497777455382878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,7168,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,10240,0.051855110459857516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,1536,0.008732444710201686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,128,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,3584,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,64,0.001822222231162919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,64,0.23806222279866537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,5120,0.2163271109263102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,51200,1.5302889082166884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,10240,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,1024,0.0038275557259718576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,7168,0.01626133256488376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,1536,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,6144,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,51200,0.04711822337574429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,1536,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,16384,0.010652444428867765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,65536,0.11937510967254639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,51200,0.39833511246575254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,768,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,16384,0.1090008881356981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,768,0.047374222013685435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,16384,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,1024,0.00666933341158761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,128,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,32,0.004274666723277834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,4096,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,3584,0.0369395547442966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,2560,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,1536,0.015858666764365304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,2048,0.2601119942135281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,64,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,5120,0.09016266796323989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,3072,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,65536,0.062487112151251904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,768,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,512,0.015460444821251763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,10240,0.0421635541650984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,5120,0.010218666659461128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,64,0.12715821795993382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,3584,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,65536,0.14277777406904432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,5120,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,256,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,7168,0.031207111146714952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,3072,0.03218133250872294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,3072,0.016190222567982145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,7168,0.020841777324676514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,2048,0.07141066922081842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,256,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,2560,0.04768088791105482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,4096,0.028134223487642076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,3072,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,128,0.046889778640535146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,1536,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,5120,0.016146666473812528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,1536,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,768,0.021482666333516438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,5120,0.12267555130852593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,256,0.12373156017727321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,16384,0.02001333402262794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,2048,0.018394667241308425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,64,0.03108000093036228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,2048,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,51200,0.029115554359224107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,6144,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,10240,0.015856888559129503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,2048,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,5120,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,8192,0.014126222994592456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,3584,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,2048,0.006132444573773279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,3584,0.021651556094487507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,1536,0.018927999668651156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,12288,0.03866399990187751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,16384,0.09887466828028361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,64,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,3072,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,8192,0.008648888932334052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,1024,0.15888088279300266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,1536,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,5120,0.027103111147880554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,64,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,1024,0.00351200004418691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,3584,0.06554844644334581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,51200,0.06214400132497152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,51200,0.813843568166097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,1024,0.05499910977151659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,1536,0.004217777815130022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,1024,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,2560,0.0074106670088238185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,4096,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,1536,0.013049778011110095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,8192,0.06157066424687704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,512,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,512,0.010075555907355415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,16384,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,3072,0.011954666839705573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,5120,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,10240,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,2048,0.02649333410792881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,3584,0.018720888429217868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,3072,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,32,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,65536,0.8992595672607422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,128,0.00962044464217292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,256,0.019084443648656208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,3072,0.025386666258176167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,1536,0.0047111109726958806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,512,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,3072,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,256,0.006893333461549547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,3584,0.010128888818952773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,51200,3.8790134853786893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,768,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,7168,0.06812622149785359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,8192,0.024726221958796184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,2560,0.011138666835096149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,3584,0.013491555220550962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,3584,0.052779555320739746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,16384,0.03605155481232537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,4096,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,3072,0.030100444952646892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,8192,0.01833866702185737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,51200,0.05215555429458618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,8192,0.16238578160603842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,65536,0.5620026588439941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,256,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,10240,0.028316444820827905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,6144,0.05566044648488363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,6144,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,128,0.01442577772670322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,512,0.00906666699382994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,1536,0.02494488822089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,768,0.029731555117501154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,3072,0.015217777755525378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,7168,0.02798844377199809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,16384,0.0798906683921814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,8192,0.37919023301866317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,2560,0.023596444063716467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,16384,0.0766871107949151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,128,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,7168,0.03631733192337884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,6144,0.18784711096021864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,1024,0.061871111392974854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,1024,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,3072,0.009172444542249044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,8192,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,1536,0.0038337777886125776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,7168,0.03135200010405646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,2560,0.01882222294807434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,12288,0.20636622111002603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,5120,0.05932266844643486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,12288,0.08600444263882107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,8192,0.023727999793158636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,51200,0.3717484474182129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,3584,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,256,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,8192,2.8890327877468533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,32,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,1024,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,512,0.019293333093325298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,2560,0.022983110613293115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,3584,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,3584,0.033286223808924355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,51200,0.07316799958546956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,6144,0.0092604441775216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,2560,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,7168,0.14258400599161783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,65536,0.010975110861990186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,5120,0.04686844348907471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,4096,0.014496000276671516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,3584,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,64,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,10240,0.01924888955222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,65536,0.15539911058213976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,3584,0.008952889177534316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,768,0.017222222354676988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,5120,0.10172177685631646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,10240,0.8755111164516873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,128,0.015035554766654968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,768,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,7168,0.02529866662290361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,256,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,4096,0.03078311019473606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,7168,0.16407644748687744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,65536,0.06120977799097697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,6144,0.029096000724368628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,16384,5.12852520412869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,12288,0.06422133578194512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,7168,0.11614755789438884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,1024,0.008608889248636033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,51200,0.17930844095018175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,64,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,12288,0.07025777631335788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,8192,0.012784889174832238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,8192,0.017640888690948486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,1536,0.01761244402991401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,16384,0.05544622076882256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,3584,0.02035911050107744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,7168,0.09232266743977864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,2048,0.15692710876464844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,512,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,256,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,12288,0.038155555725097656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,12288,0.18307911025153267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,7168,0.2288284566667345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,7168,0.02639555599954393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,6144,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,3072,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,6144,0.012376888758606382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,65536,0.2942710982428657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,8192,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,8192,0.011376889215575324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,256,0.005070222334729301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,128,0.006199999815887875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,2560,0.007076444725195567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,1024,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,1024,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,1024,0.02777066661251916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,2560,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,512,0.002653333254986339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,64,0.06407466861936781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,2560,0.019311111834314134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,8192,0.022828444838523865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,1536,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,51200,1.4565715789794922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,32,0.006896889044178857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,512,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,768,0.19080266687605116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,32,0.01551644504070282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,256,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,2048,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,6144,0.03756177756521437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,2048,0.1170204480489095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,2560,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,10240,2.2908087836371527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,32,0.007807110746701558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,5120,0.012348444097571902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,8192,0.014711999230914645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,256,0.0633937782711453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,3072,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,3584,0.0090515555606948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,5120,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,12288,0.47934309641520184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,768,0.05158399873309665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,10240,0.4020124541388617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,32,0.02826666831970215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,12288,0.12055467234717475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,12288,0.013926222920417786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,2560,0.005170666509204441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,64,0.1734639936023288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,1024,0.0034355556385384668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,12288,1.2049404780069988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,1024,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,12288,0.03153511219554477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,51200,0.8529724544949002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,3584,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,6144,0.028093334701326158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,256,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,10240,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,256,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,6144,0.018570666511853535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,65536,0.11209511756896973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,7168,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,2048,0.04644977715280321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,10240,0.01847022275129954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,64,0.02345688806639777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,2048,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,64,0.07733688751856486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,64,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,1536,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,768,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,3584,0.038979556825425886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,10240,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,128,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,8192,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,6144,0.1977511114544339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,2560,0.014919999572965833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,32,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,1536,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,3072,0.523074679904514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,5120,0.013825777504179211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,51200,0.014236445228258768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,12288,30.794474283854168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,4096,1.4741652806599934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,51200,0.40663909912109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,65536,0.02460622290770213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,51200,0.05312088794178433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,2048,0.15698488553365073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,2560,0.053802665736940175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,3072,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,1024,0.0034373334298531213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,7168,0.0282008879714542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,12288,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,512,0.011031111081441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,768,0.22716355323791504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,3584,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,512,0.007038222418891058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,10240,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,12288,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,51200,0.18697333335876465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,6144,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,128,0.02019999921321869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,2560,0.05625600285000271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,3072,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,1024,0.052277333206600614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,3072,0.07239822546641032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,3584,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,128,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,7168,0.040431999497943454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,256,0.16215733687082926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,1024,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,6144,0.017273777061038546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,10240,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,51200,0.023973332511054143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,7168,0.04134133458137512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,4096,0.03195110956827799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,256,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,16384,0.5497182210286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,1024,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,3072,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,65536,0.08847555849287246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,2048,0.014889778362380134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,12288,0.15374844604068333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,3072,0.03650577863057455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,512,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,2048,0.023990222149425085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,128,0.005389333185222413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,16384,0.022826666633288067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,65536,0.4249475532107883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,2560,0.030300445026821558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,4096,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,512,0.002616000051299731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,768,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,512,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,768,0.015663110547595553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,2560,0.15213421980539957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,7168,0.021757332815064326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,512,0.003044444446762403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,51200,0.015491555134455362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,3584,0.026661333110597398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,2048,0.015207111835479736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,32,0.01646844380431705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,65536,0.4191511207156711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,7168,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,5120,0.024359999431504145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,5120,0.02427111069361369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,16384,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,3072,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,65536,0.09716977675755818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,3584,0.03137244449721442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,768,0.014520888527234396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,64,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,32,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,16384,0.2344062328338623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,5120,0.048376888036727905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,512,0.02013688948419359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,65536,0.06810666455162896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,12288,0.06867911418279012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,512,0.012736000120639801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,3584,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,2048,0.02590488890806834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,3584,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,64,0.0028880000528362063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,768,0.004540444248252445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,7168,0.0687297781308492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,51200,0.04804177747832405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,8192,0.03910488883654276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,3072,0.046803557210498385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,6144,0.01330311099688212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,64,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,3072,0.02702666653527154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,128,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,768,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,32,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,3584,0.020617778102556866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,5120,0.3117608759138319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,12288,0.021239999267790053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,128,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,7168,0.023804444405767653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,2048,0.08532266484366523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,8192,0.013073777986897362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,256,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,16384,0.01462044484085507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,5120,0.009279111193286048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,128,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,6144,1.311736848619249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,12288,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,2048,0.006170666466156642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,64,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,768,0.016422222057978313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,3072,0.013252444565296173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,256,0.017656000124083627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,8192,0.011644444531864591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,256,0.016352888610627916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,12288,0.017809778451919556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,16384,0.00776088900036282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,128,0.006709333509206772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,6144,0.07221333185831706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,2560,0.02330133318901062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,64,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,32,0.03164533442921109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,5120,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,8192,0.017856000198258292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,3072,0.02760355505678389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,512,0.012888000243239932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,51200,0.05394488904211256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,8192,1.0305155648125541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,51200,0.7547982003953723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,2560,0.016265778077973258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,12288,0.025255999631351892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,12288,0.3234933217366536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,768,0.03744533326890733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,65536,1.109524408976237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,8192,0.08771555291281806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,10240,0.024470221665170457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,5120,0.01588355501492818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,1024,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,2048,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,51200,0.09174755546781753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,2560,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,5120,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,1536,0.09052355421913995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,2560,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,51200,0.1853546698888143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,3072,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,1024,0.012216000093354119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,4096,0.1973315609825982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,3072,0.051033778323067554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,51200,0.7481208907233344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,3072,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,16384,1.8021750979953344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,1536,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,5120,0.020600888464185927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,256,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,12288,0.04864711231655545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,128,0.03421422176890903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,65536,4.749523586697049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,4096,0.01366133325629764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,6144,0.017464000317785475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,1024,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,3072,0.026362667481104534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,2560,0.0070293330483966405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,6144,0.02699377801683214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,2048,0.024002666274706524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,2048,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,2048,0.008711111214425828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,7168,0.012129777835475074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,3072,0.031048887305789526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,1536,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,256,0.0022293333378103045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,2560,0.09454222520192464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,6144,0.05323466658592224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,3584,0.011963555382357704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,8192,0.022666666242811415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,32,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,7168,0.04845688740412394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,65536,0.0700168874528673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,3584,0.01644622286160787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,16384,0.1485982206132677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,65536,0.07041777504814996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,1536,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,10240,0.015288889408111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,7168,0.036100443866517805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,64,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,1536,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,8192,0.24290667639838326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,12288,0.01346933344999949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,1536,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,6144,0.0358551111486223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,6144,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,256,0.028591109646691218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,51200,0.2041706641515096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,128,0.005490666462315454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,51200,1.363276481628418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,3072,0.013528000149461957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,65536,0.04695733388264974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,512,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,768,0.004621333132187526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,128,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,3584,0.021049777666727703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,65536,0.5227208667331272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,256,0.04662755462858412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,2048,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,768,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,5120,0.06878755489985149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,768,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,512,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,128,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,12288,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,65536,0.26260532273186576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,7168,0.060106668207380504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,64,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,7168,0.011887110769748688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,12288,0.06391288836797078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,256,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,3072,0.006673777682913675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,65536,0.16374844974941677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,3072,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,5120,0.009408888717492422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,4096,0.16671289338005912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,4096,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,4096,0.04037955403327942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,3072,0.03813599877887302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,6144,0.02198133369286855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,8192,0.10007199976179336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,2560,0.047797332207361855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,2560,0.01460088955031501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,51200,0.08286577463150024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,64,0.0025742221623659134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,64,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,12288,0.04073422153790792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,3072,0.030759112702475652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,3584,0.010464888479974536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,12288,0.027024888330035742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,64,0.06369688775804308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,128,0.012820444173283048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,51200,0.10842755768034194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,512,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,768,0.017338666650984023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,64,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,768,0.03808622227774726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,16384,0.024902222885025874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,1536,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,7168,0.05166133244832357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,10240,2.555779563056098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,10240,0.014500444134076437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,4096,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,128,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,51200,0.15867200162675646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,16384,0.0806079970465766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,128,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,1536,0.02001333402262794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,2048,0.005653333332803514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,256,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,256,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,2048,0.04905688762664795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,256,0.007339555356237624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,4096,0.07357066869735718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,768,0.025003555748197768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,51200,0.024890666206677754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,65536,0.1252711084153917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,10240,0.037315554089016385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,768,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,768,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,8192,0.00998222248421775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,256,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,8192,0.011000889043013254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,512,0.03443377878930833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,3584,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,128,0.02160444524553087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,5120,0.02609066665172577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,8192,0.017921777235137094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,1536,0.006351111249791251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,1536,0.0196115556690428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,256,0.029150221082899306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,6144,0.00981600003110038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,32,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,5120,0.026383111874262493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,10240,0.0188773340649075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,2560,0.08828977743784587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,512,0.020469332734743755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,768,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,65536,0.045012444257736206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,256,0.03513777918285794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,4096,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,16384,0.05145333210627238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,256,0.009483555952707926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,256,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,7168,0.19288978311750624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,16384,0.8297244707743326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,65536,0.29067378573947483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,512,0.009278222090668148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,10240,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,256,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,256,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,4096,0.03959733247756958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,7168,0.03465244505140517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,2048,0.030164443784289893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,512,0.03683822353680929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,1536,0.006486222147941589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,12288,0.01869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,7168,1.1318115658230252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,32,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,32,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,12288,3.287586636013455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,768,0.009259555902745988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,65536,4.331681569417317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,512,0.028391109572516546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,768,0.012830221818553077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,256,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,7168,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,10240,0.014882667197121514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,768,0.01182044463025199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,256,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,10240,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,32,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,256,0.016598222984208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,10240,0.02358755634890662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,4096,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,16384,0.09124533335367839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,5120,0.09127289056777954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,51200,0.053957333167394005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,10240,0.4967137972513835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,16384,0.07954400115542941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,8192,0.10342133045196533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,12288,0.22607999377780488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,3072,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,6144,0.1728044483396742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,7168,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,51200,2.7052737341986766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,2048,0.5110675493876139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,51200,0.8706017600165473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,512,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,3072,0.049024889866511025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,512,0.011920889218648275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,7168,0.0392506652408176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,10240,0.06895999775992499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,3584,0.04037422272894118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,51200,0.052710221873389355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,12288,0.03686044282383389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,512,0.0025519999778932994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,51200,0.11583556069268121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,10240,0.02051911089155409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,1536,0.05317244595951504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,768,0.1656879981358846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,768,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,1024,0.044249776336881846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,8192,0.018589332699775696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,1024,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,128,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,512,0.011378666592968835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,7168,0.00686488880051507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,1024,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,3072,0.0758320026927524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,2048,0.018357333209779527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,7168,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,1024,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,6144,0.0765777760081821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,4096,0.008153777983453538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,8192,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,12288,0.00958577791849772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,65536,1.3593208524915907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,1024,0.28729245397779674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,512,0.05609155363506741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,65536,0.10069955719841851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,3584,0.010324444207880232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,5120,0.15082489119635686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,1024,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,2048,0.18641332785288492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,10240,0.02714755634466807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,2048,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,12288,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,12288,0.139082670211792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,64,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,10240,0.09475644429524739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,8192,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,16384,1.403998268975152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,65536,1.163815074496799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,16384,0.3789546754625108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,65536,2.2202951643202042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,51200,0.06487022505866157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,10240,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,10240,0.05464888943566216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,768,0.05255022313859728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,6144,0.17387021912468803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,256,0.025033776958783466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,2560,0.021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,512,0.012829333543777466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,1024,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,3584,0.06060710880491468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,6144,0.023004444109068975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,2560,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,32,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,128,0.01389422184891171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,256,0.006888000087605582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,8192,0.012818666795889536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,64,0.027091556125217017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,2048,0.02139733400609758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,51200,0.05365955498483446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,51200,0.051573332813051015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,8192,0.025972444150182936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,12288,0.03237777617242601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,64,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,6144,0.016241777274343703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,65536,3.3420266045464406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,4096,0.4891253577338324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,768,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,4096,0.03682488865322537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,32,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,65536,0.13623289267222086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,256,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,10240,0.09322222073872884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,64,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,12288,0.0689671105808682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,5120,0.05219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,8192,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,1536,0.00851111114025116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,1536,0.014009777042600842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,65536,0.09132799837324355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,7168,0.035976890060636736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,6144,0.05890133645799425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,512,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,12288,0.03964622153176202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,2048,0.02223555578125848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,3072,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,2048,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,65536,0.07423377699322171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,5120,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,10240,0.126129772928026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,5120,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,128,0.00600355573826366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,10240,0.19228443834516737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,65536,1.5969964133368597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,16384,3.553988350762261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,3072,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,8192,0.04762399858898587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,3584,0.011685333318180509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,12288,0.12469600306616889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,2560,0.10273866520987616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,5120,0.010583110981517367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,3584,4.37754143608941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,4096,0.22693332036336264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,6144,1.0925395753648546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,10240,0.2677031093173557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,4096,0.02774133284886678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,256,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,2048,0.08236089017656115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,3584,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,64,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,12288,0.15977599885728624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,128,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,8192,0.13633599546220568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,51200,0.048912889427608915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,1536,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,4096,0.019515555765893724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,8192,0.03968177901373969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,8192,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,3072,0.02239911092652215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,8192,0.013092444174819522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,768,0.012523555921183692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,4096,0.027460444304678176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,128,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,2560,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,2048,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,3584,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,6144,0.049361778630150684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,768,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,768,0.02494488822089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,12288,0.04765511221355862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,512,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,6144,0.14316444926791722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,6144,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,1536,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,128,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,128,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,3584,0.025547555751270715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,51200,1.3290061950683594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,1024,0.015555555621782938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,128,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,16384,0.1013608906004164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,12288,0.03070577647950914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,4096,0.06442488564385308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,5120,0.027906666199366253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,32,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,128,0.0092284447617001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,256,0.02887466549873352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,4096,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,8192,0.04381066560745239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,512,0.0346684455871582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,1536,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,12288,0.11987822585635716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,256,0.004580444345871608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,7168,0.10621955659654404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,51200,0.3061857753329807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,51200,0.7775875727335612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,128,0.022706665926509436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,512,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,3584,0.018232888645595975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,512,0.1754080057144165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,32,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,1024,0.13265244166056314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,64,0.0064497776329517365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,2048,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,4096,0.028784000211291846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,7168,0.5476746559143066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,65536,0.3344382180107965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,1024,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,2560,0.008939555949634975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,5120,0.01927911076280806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,65536,0.06406222449408637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,2560,0.5115795665317112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,256,0.023003555006451074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,256,0.02721422248416477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,12288,0.0907119976149665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,4096,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,3584,0.017307556337780423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,3072,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,512,0.04684355523851183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,128,0.028674665424558852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,768,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,65536,0.09177333116531372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,16384,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,32,0.08096444606781006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,512,0.012869333227475485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,6144,0.05111022127999199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,8192,0.010851555400424533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,5120,0.1282986667421129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,8192,0.017172444197866652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,5120,0.01147377739350001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,768,0.009706666900051964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,256,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,1024,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,5120,0.011902222202883827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,128,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,16384,0.08786311414506699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,16384,0.13351821899414062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,5120,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,2048,0.01312622262371911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,3584,0.01110222190618515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,64,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,51200,0.05297422077920702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,1536,0.008348444269763099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,2048,0.006736889067623351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,1536,0.013560888667901358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,5120,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,2560,0.037568890386157565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,2048,0.012076444096035428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,5120,0.0063182223174307085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,51200,0.13737689124213323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,7168,0.030869334936141968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,64,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,2048,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,6144,1.0991715325249565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,51200,0.059897780418395996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,12288,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,768,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,4096,0.06952177815967135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,2560,0.021282666259341772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,4096,0.007887110941939885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,768,0.012344888514942594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,2560,0.02886577778392368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,6144,0.06463378005557589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,1024,0.39217244254218203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,2048,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,768,0.10851022269990708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,32,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,512,0.015242666006088257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,16384,0.021270222134060327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,3072,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,4096,0.0042151109211974675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,12288,0.026885333988401625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,768,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,1536,0.08045066727532281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,128,0.03775555557674832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,256,0.082241780228085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,8192,0.060361776087019176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,512,0.061214221848381885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,12288,0.01795022189617157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,8192,0.13985955715179443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,2048,0.11679732799530029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,512,0.04028977619277106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,1024,0.031839112440745033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,4096,0.021555556191338435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,6144,0.01759199963675605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,16384,0.13769511381785074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,4096,0.04006577862633599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,7168,0.09472088681326972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,2048,2.1174916161431208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,5120,0.02048177851570977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,1536,0.04672355453173319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,128,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,51200,0.6739057964748807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,6144,0.11223910914527045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,1536,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,768,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,65536,0.081423110432095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,768,0.009363555245929295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,768,0.02534488836924235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,4096,0.012514666550689273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,51200,0.10359822379218207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,512,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,32,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,51200,0.0664106673664517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,64,0.0018684443914228016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,8192,0.21191910902659097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,65536,0.05776355663935343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,64,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,64,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,10240,0.04190311166975233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,8192,0.028265777561399672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,256,0.00673333348499404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,12288,0.03432177835040622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,256,0.03532800078392029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,16384,0.05288177728652954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,16384,0.4090426762898763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,4096,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,1024,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,3072,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,5120,0.016543111867374845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,4096,0.02273511058754391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,32,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,1024,0.08575199709998237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,65536,1.7287439770168727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,16384,0.06993955373764038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,4096,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,6144,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,12288,0.9864702224731445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,2048,0.03200177682770623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,5120,0.015989333391189575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,1024,0.05101066827774048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,2560,0.6224400202433268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,16384,0.022670222653283015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,51200,3.480587641398112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,768,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,1536,0.018383999665578205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,1536,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,3584,0.03574133250448439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,256,0.018173333671357896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,768,0.019093333019150626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,3584,0.0269128895468182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,512,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,7168,0.013172444370057849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,32,0.0038328886859946777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,1024,0.023980443676312763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,4096,0.023100444012218054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,65536,0.12406399514940049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,6144,0.1111253367529975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,5120,0.10646044545703465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,5120,0.05650755431916979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,768,0.00776800016562144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,5120,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,65536,0.08496266603469849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,10240,1.1342159907023113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,3584,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,51200,0.1451173358493381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,3584,0.03802488909827338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,51200,0.014363555444611443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,2560,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,7168,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,3072,0.019001776973406475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,2560,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,51200,2.06276618109809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,65536,0.15840889347924128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,51200,0.7584160168965658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,32,0.024710221423043147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,768,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,8192,0.0352168877919515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,32,0.007214222517278459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,10240,0.028202666176689997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,12288,0.19519733058081734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,3072,0.01824711097611321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,256,0.025310221645567153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,1536,0.026547556122144062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,2048,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,512,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,2560,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,16384,0.1790168947643704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,51200,10.382836235894098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,12288,0.011607111328177981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,8192,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,10240,0.02916533417171902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,128,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,5120,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,768,0.023196443915367126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,2048,0.018035555879275005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,3584,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,3072,0.06105866697099474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,65536,0.5640177726745605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,256,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,1536,0.02980533242225647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,8192,0.078676442305247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,8192,0.010737777584128909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,8192,0.0503840015994178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,1536,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,1024,0.0035004446076022256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,512,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,3072,0.017306667235162523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,2560,0.015105777316623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,1024,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,512,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,51200,0.08256000280380249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,512,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,1536,0.0363377763165368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,64,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,4096,0.03491288754675124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,5120,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,2560,0.007194666398896112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,7168,0.01588355501492818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,7168,0.013610666824711693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,32,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,7168,0.18590933746761748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,12288,0.021959111094474792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,3584,0.040898667441474065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,1024,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,2048,0.06844533152050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,2048,0.004312889029582341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,6144,0.03882044553756714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,1024,0.005681777579916849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,3072,0.017495110630989075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,12288,0.01866666641500261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,6144,0.012180444267061023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,32,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,12288,0.7473386658562554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,10240,0.012368888490729861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,6144,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,51200,0.12978577613830566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,512,0.021333333518770006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,12288,0.281330664952596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,1024,0.028169777658250596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,7168,1.2564284006754558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,16384,0.024472888973024156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,32,0.0026551110463009942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,768,0.020847999387317236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,8192,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,3584,0.017679999272028606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,2560,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,768,0.0070293330483966405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,512,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,8192,0.026153778036435444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,3584,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,768,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,256,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,51200,0.13341511620415583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,1024,0.018801778554916382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,4096,0.009733333355850643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,5120,0.21270399623446992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,8192,0.024140444066789415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,128,0.017885333961910672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,256,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,6144,0.009401777552233802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,2048,0.01365066650840971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,1024,0.13170044951968723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,2560,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,10240,0.39273866017659503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,8192,0.13705155584547254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,2048,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,64,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,768,0.027449778384632532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,4096,0.012331555287043253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,10240,0.03974399964014689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,5120,0.055642664432525635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,128,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,51200,0.11360443962944879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,7168,0.03299644589424133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,3584,0.025407999753952026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,3072,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,4096,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,512,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,12288,0.046535998582839966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,8192,0.06997778018315633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,8192,0.01644355555375417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,5120,0.017948443690935772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,3072,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,16384,0.12572533554501003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,16384,1.54784181382921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,8192,0.16676533222198486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,8192,0.2160186635123359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,4096,0.06832266516155668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,7168,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,5120,0.03507644600338406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,6144,0.010280888941552904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,256,0.008368888662921058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,3072,0.024000888069470722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,2048,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,2560,0.009751111268997192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,3584,0.09255822499593098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,8192,0.28955645031399196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,1536,0.6675857967800565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,7168,0.030058665408028498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,5120,0.028485334581798975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,2048,0.02833777666091919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,10240,0.06224977970123291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,1024,0.03102311160829332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,16384,0.2973751227060954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,4096,0.027183110515276592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,5120,0.0759484436776903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,7168,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,1024,0.05713422430886162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,768,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,5120,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,12288,0.07755288812849256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,8192,0.044291555881500244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,256,0.005306666509972678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,10240,0.052377776967154614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,64,0.003137777869900068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,256,0.02774577836195628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,51200,0.22349421183268228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,1536,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,256,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,3072,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,32,0.008379555410808986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,2560,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,512,0.005776888794369168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,2560,0.046610666645897754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,10240,1.3007386525472004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,16384,0.02334311107794444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,512,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,1536,0.0240639994541804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,12288,0.0569795568784078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,5120,0.02169955604606205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,2560,0.028078221612506445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,64,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,768,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,64,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,256,0.0577768882115682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,2560,0.014345778359307183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,8192,0.023381332556406658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,3584,0.046373334195878774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,1024,0.03916355636384752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,10240,0.15884889496697321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,768,0.054224888483683266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,65536,0.06559022267659505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,2560,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,51200,0.19954133033752441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,4096,0.08963021967146133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,5120,0.012563555604881711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,64,0.008266666697131263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,256,0.021545777718226116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,6144,0.020021332634819877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,2560,0.011694221860832639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,12288,0.025003555748197768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,256,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,6144,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,16384,0.03762844536039565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,2048,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,4096,0.02439022229777442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,65536,0.07874933216306898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,2048,0.0047626665068997275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,64,0.0018737777653667664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,3072,0.05479733480347527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,128,0.012831110921170978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,1536,0.06155911419126722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,8192,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,3584,0.39519821272955996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,1024,0.004643555730581284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,6144,0.042875554826524526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,3584,0.04217866725391812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,6144,0.031874666611353554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,6144,0.010265777508417765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,1024,0.01757244434621599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,3072,0.02372177773051792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,8192,0.03910933269394769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,2560,0.544366200764974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,16384,0.47200621498955625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,4096,0.01978577839003669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,8192,0.024493333366182115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,2560,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,2560,0.1401031149758233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,32,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,1536,0.022375111778577168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,3584,0.017493334081437852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,768,0.011361777782440186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,768,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,2048,0.02791822287771437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,12288,0.044293334086736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,10240,0.16422043906317815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,12288,0.05204088820351494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,6144,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,10240,0.036334223217434354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,7168,0.010187555518415239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,16384,0.01479111115137736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,16384,0.021465778350830078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,1024,0.031172444423039753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,64,0.02330400049686432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,10240,0.23367110888163248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,65536,0.04813777738147312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,4096,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,256,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,16384,0.035663998789257474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,3584,0.01278222186697854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,3072,0.025685333543353613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,6144,0.019149333238601685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,10240,0.04876888791720072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,12288,0.1317866643269857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,65536,0.252691560321384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,768,0.03779822256830003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,32,0.030078222354253132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,3584,0.057373331652747266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,8192,0.01995644470055898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,6144,0.02028622229894002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,51200,0.0664079984029134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,3584,0.08031466934416029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,5120,0.017384888397322763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,12288,0.1637564500172933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,1024,0.016015999846988253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,2560,0.022943110929595098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,8192,0.3421759870317247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,1536,0.05994133154551188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,512,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,51200,0.05353777938418918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,6144,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,256,0.01482222146458096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,3072,0.01588000026014116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,2560,0.016530666086408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,3584,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,2560,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,768,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,1024,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,3584,0.013578666581047906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,32,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,5120,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,1024,0.044249776336881846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,4096,0.014444443914625378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,6144,0.013278221918476952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,4096,0.2926017708248562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,3072,0.03892799880769517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,10240,0.06261422236760457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,2048,0.011004444625642566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,3584,0.06173066960440742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,10240,0.10468533304002549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,128,0.01962222158908844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,8192,0.01571822166442871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,8192,0.11331021785736084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,16384,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,3072,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,1024,0.3288773430718316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,65536,0.12904266516367593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,128,0.015513777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,512,0.08933066659503514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,7168,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,2560,0.018245332770877414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,2048,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,3584,0.022054221895005967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,12288,0.007920000288221572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,2560,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,65536,0.8013031217787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,2560,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,768,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,2048,0.008773333496517604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,2048,0.05211555626657274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,6144,0.01187999960449007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,12288,0.013177778157922955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,4096,0.024375999967257183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,65536,0.20360267162322998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,7168,0.015791111522250705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,2560,0.03997333182228936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,1536,0.02051466703414917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,6144,0.019878221882714164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,7168,0.009654222263230218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,7168,0.04009333252906799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,768,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,4096,0.020278222031063504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,128,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,1536,0.061689780818091504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,512,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,768,0.01749955614407857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,12288,0.8065004348754883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,3072,0.05594755543602837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,5120,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,65536,2.562061309814453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,8192,5.190579732259114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,4096,0.23523378372192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,3584,0.03529333406024509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,6144,0.13377955224778917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,5120,0.03269955515861511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,51200,3.926111009385851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,3584,0.04217244519127739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,65536,1.2237395180596244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,8192,0.01607377827167511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,32,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,5120,0.1130666732788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,12288,0.14964266618092856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,7168,0.019899555378490023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,6144,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,5120,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,2560,0.01497866710027059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,3584,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,256,0.06538310978147718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,7168,0.07264177666770087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,32,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,128,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,6144,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,128,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,7168,0.04015466570854187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,16384,0.28022223048739964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,256,0.011997333003415002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,1536,0.007088888850477006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,8192,0.022984888818528917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,4096,0.02511111067401038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,1024,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,768,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,5120,0.059789333078596324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,10240,0.028063999281989202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,10240,0.036871112055248685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,2048,0.04525244567129347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,256,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,128,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,3584,0.01351199961370892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,51200,0.13087999820709229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,3072,0.059254222446017794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,2560,0.02260622216595544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,1536,0.03675999906327989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,128,0.008246222303973304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,3072,0.028008888165156048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,7168,0.01983022193113963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,5120,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,6144,0.021497777766651575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,10240,0.05298933386802673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,16384,0.05837066968282064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,12288,0.03152088986502753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,7168,0.016537777251667447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,2048,0.04795022143257988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,1024,0.03064089020093282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,10240,0.019253333409627277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,2560,0.09800622198316787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,1024,0.003543111185232798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,7168,0.08194044563505384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,1024,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,6144,0.03309333324432373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,12288,3.5487412346733938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,4096,0.4120275444454617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,64,0.010830221904648675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,5120,0.007530666887760162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,10240,0.12364266978369819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,7168,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,2048,0.004563555535342959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,4096,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,2048,0.36415910720825195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,128,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,2560,0.2373297744327121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,12288,0.0648133357365926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,4096,0.05504444572660658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,6144,0.023952888117896184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,12288,0.016819554898473952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,3072,0.01891911029815674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,32,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,2560,0.023342221975326538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,1024,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,768,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,64,0.01680355601840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,256,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,65536,0.1736053360833062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,128,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,51200,0.053582221269607544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,6144,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,8192,0.2634844515058729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,7168,0.017831110292010836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,7168,0.12775644991132948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,2560,0.0793822209040324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,6144,0.08113333251741198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,1536,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,8192,0.07829066779878405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,3584,0.08330666356616549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,4096,0.01352266636159685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,768,0.025490666429201763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,16384,0.029693331983354356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,6144,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,512,0.016812443733215332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,6144,0.22974220911661783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,51200,0.8574613465203179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,3584,0.00888533310757743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,512,0.0026524443593290118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,1536,0.020907556017239887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,8192,0.018792000081804063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,51200,7.98690202501085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,512,0.009230222139093611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,65536,1.4508239958021376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,1536,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,16384,0.038944890101750694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,16384,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,256,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,2048,0.25683289104037815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,1024,0.004972444640265571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,7168,0.1273768875333998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,256,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,1024,0.053073777092827686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,16384,0.020076443751653034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,2048,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,1024,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,51200,0.3628551165262858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,1536,0.04120622078577677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,512,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,768,0.04247644543647766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,768,0.029836446046829224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,768,0.055324445168177284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,256,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,65536,0.07587733533647326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,1536,0.3076666726006402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,32,0.00907022257645925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,5120,0.7962693108452691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,128,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,3584,0.029719110992219713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,1024,0.027492443720499676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,10240,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,51200,0.062368002202775746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,512,0.039800001515282526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,256,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,2560,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,3584,0.06778399811850654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,32,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,5120,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,256,0.03160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,1536,0.0035537779331207275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,7168,0.013032889200581444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,1536,0.22038666407267252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,128,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,8192,0.1944488949245877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,128,0.021665778425004747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,16384,0.10027644369337295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,16384,0.11248532931009929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,65536,0.06782578097449408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,5120,0.010488889283604093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,4096,0.01426577733622657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,3072,0.0860897766219245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,2048,0.010852444503042432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,16384,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,6144,0.02050133380624983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,2560,0.04586577746603224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,7168,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,2560,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,16384,0.13047556082407633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,16384,0.05331377850638496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,16384,0.04552266663975186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,512,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,512,0.01475822263293796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,256,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,12288,0.032454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,3584,0.014088888963063559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,51200,0.02283199959331089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,1024,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,10240,0.09052799807654487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,512,0.01538222200340695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,7168,0.07556355661816068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,5120,0.012395555774370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,64,0.002773333340883255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,6144,0.01792977750301361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,1024,0.0035111109415690103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,5120,0.023763555619451735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,10240,0.02055999967787001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,6144,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,512,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,10240,0.8908373514811198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,3072,0.07139111227459378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,2560,0.026667555173238117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,4096,0.11989777617984348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,128,0.08732000324461196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,7168,0.012506666282812754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,8192,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,12288,0.03525777657826742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,16384,0.012311999996503195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,1024,0.21345510747697619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,5120,0.13716267214881048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,3584,0.04331733451949226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,2048,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,10240,0.08773777882258098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,5120,0.013741333451535968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,3584,0.02163733376397027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,2560,0.020816889074113634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,768,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,8192,0.01036444471942054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,8192,0.050761779149373375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,16384,0.10264888736936782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,256,0.021070222059885662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,6144,0.045201778411865234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,1024,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,7168,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,3584,0.3707715670267741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,65536,0.7665457725524902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,7168,0.022407111194398668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,8192,0.0292871097723643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,1024,0.13340977827707926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,64,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,8192,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,12288,0.3240275647905138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,51200,1.1357182396782768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,10240,0.030988444884618122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,6144,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,7168,0.08822755681143866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,65536,0.06042489078309801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,10240,0.013365333278973898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,51200,0.10679999987284343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,4096,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,10240,0.0868684450785319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,2048,0.010889777706729041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,256,0.002226666650838322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,512,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,2048,0.02554399934079912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,256,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,3072,0.04455733299255371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,5120,0.10250666406419541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,512,0.013069333301650153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,1024,0.5843457645840114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,2560,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,12288,0.034992890225516424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,3072,0.2554524474673801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,512,0.002944888960984018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,4096,1.2189182705349393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,1024,0.003216888962520493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,1536,0.018558222386572096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,32,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,4096,0.032906668053732976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,3584,0.16399199432796902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,65536,0.434787564807468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,2560,0.012676444318559436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,12288,0.04237777656979031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,16384,0.05842666493521797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,768,0.023697778582572937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,8192,0.10762133200963338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,5120,0.04819733235571119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,6144,1.075160026550293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,7168,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,16384,0.016741333736313712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,12288,0.20342666572994658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,1536,0.0422346658176846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,5120,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,7168,0.16740000247955322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,768,0.003944000022278892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,10240,0.014113777213626437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,2048,0.18650666872660318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,1536,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,2048,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,16384,0.04032622112168206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,2560,0.052336888180838696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,3584,0.007721777591440413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,2048,0.010208889014191097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,3072,0.07060266865624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,65536,0.21961243947347006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,10240,0.1991644435458713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,51200,0.047354665067460805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,5120,0.06724444362852308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,768,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,2560,0.007048889166778988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,3072,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,2560,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,1536,0.07501688930723402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,1024,0.010476444330480365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,512,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,1536,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,1024,0.00813244448767768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,51200,0.05175644490453932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,6144,0.02169422143035465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,1536,0.03264355659484863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,256,0.00342399999499321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,32,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,3072,0.05614488654666477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,2048,0.02405333353413476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,4096,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,1536,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,4096,0.01346933344999949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,7168,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,10240,6.535758972167969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,2560,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,8192,0.07629066705703735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,65536,0.15150666236877441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,32,0.0229120006163915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,3072,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,3072,0.007329777710967594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,16384,0.027419555518362258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,65536,0.08692533440060085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,2048,0.6046844588385688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,5120,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,10240,0.02313066687848833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,65536,0.06387466854519315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,1536,0.04135911001099481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,2560,0.04649244414435493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,8192,0.023406222462654114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,32,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,51200,0.07178933090633817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,128,0.02197866638501485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,768,0.027056889401541814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,512,0.016537777251667447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,2048,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,512,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,3584,0.05855111281077067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,7168,0.47248087988959414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,51200,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,128,0.002895999906791581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,512,0.02107200026512146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,768,0.05214044451713562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,2560,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,3584,0.01794577803876665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,16384,0.28910666041904026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,5120,0.0317191117339664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,65536,0.09442666504118179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,8192,0.1592293315463596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,256,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,256,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,768,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,51200,0.10206222534179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,128,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,5120,0.02570044497648875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,7168,0.0869377785258823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,7168,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,3584,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,1536,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,3072,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,2048,0.02095288866096073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,32,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,4096,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,2048,0.012353777885437012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,128,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,5120,0.02500266664557987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,5120,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,3584,0.15694133440653482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,256,0.06103110975689358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,8192,0.25239022572835285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,8192,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,64,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,128,0.014087110757827759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,10240,0.12745066483815512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,1536,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,64,0.0018995555324686898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,768,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,7168,0.015439111325475903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,10240,0.6396933131747776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,7168,0.186645335621304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,256,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,3584,0.014704888065656027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,10240,0.05736177497439914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,3584,0.03729066583845351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,3584,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,128,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,12288,0.04958488874965244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,128,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,4096,0.020868443780475195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,12288,0.029007110330793593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,10240,0.13991288344065347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,4096,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,32,0.0032195556494924757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,51200,0.13410845067765978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,1024,0.03554844525125291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,65536,0.533531559838189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,4096,0.1823813385433621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,8192,0.05979644589953952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,64,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,6144,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,10240,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,128,0.09021244446436565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,2560,0.014884443746672736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,2560,0.007094221810499827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,8192,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,3584,0.04154133465554979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,768,0.011883555187119378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,10240,0.019544000426928203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,3584,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,8192,0.2713315486907959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,5120,0.01187111106183794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,7168,0.6576604313320583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,6144,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,768,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,2048,0.009749333063761393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,10240,0.28557154867384166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,128,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,6144,0.011600889265537262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,5120,0.0207368897067176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,64,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,6144,0.14979644616444907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,128,0.01551822159025404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,5120,0.0377422206931644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,2048,0.03409333361519708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,128,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,32,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,10240,0.7020426856146919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,12288,0.036279112100601196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,64,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,12288,0.19686222076416016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,128,0.020280889338917203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,6144,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,2560,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,1024,0.010000000397364298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,16384,0.04433688852522108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,1536,0.016354666815863717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,6144,0.08056266440285577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,4096,0.01574844453069899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,1536,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,4096,0.037645333343082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,1536,0.003834666477309333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,512,0.022674666510687932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,3072,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,12288,0.08431733316845363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,7168,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,65536,0.06284355454974704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,32,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,5120,0.019671999745898776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,2048,0.19691733519236246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,64,0.057687997817993164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,3072,0.09638310803307427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,2048,0.02159022291501363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,8192,0.011608888705571493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,3584,0.026047110557556152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,3584,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,10240,0.09331111113230388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,2048,0.01588800052801768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,3072,0.03217333224084642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,2048,0.013296888934241401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,3072,0.02424888809521993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,3072,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,1536,0.02085777786042955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,128,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,1536,0.00423911131090588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,128,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,128,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,6144,0.017883555756674874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,12288,0.032616890139049955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,7168,0.017533333765135873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,1024,0.0188711103465822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,768,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,2048,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,128,0.10450933376948039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,10240,0.02508977717823452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,2560,0.02423111100991567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,768,0.007596444752481248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,5120,0.17747199535369873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,6144,0.05330222182803684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,512,0.024505777491463557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,128,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,4096,0.010695999695195092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,768,0.0026435556097163092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,51200,0.18419910801781547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,32,0.1001288890838623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,16384,0.02641511129008399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,3072,0.01555288831392924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,512,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,3584,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,7168,0.013019555144839816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,64,0.013968000809351603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,10240,0.010290666586822933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,512,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,2048,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,7168,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,16384,3.199289745754666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,51200,1.133896933661567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,10240,0.09427199761072795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,4096,0.044622222582499184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,128,0.018043556147151522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,1024,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,2560,0.0129439996348487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,2560,0.015252444479200574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,3072,0.09147910939322577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,1024,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,3072,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,256,0.01854933301607768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,6144,0.02866311205757989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,768,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,10240,0.01294133315483729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,256,0.010440888504187265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,512,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,51200,0.10399555497699314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,51200,0.27580444018046063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,3072,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,768,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,512,0.00479911102188958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,256,0.0199955552816391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,7168,0.1532426675160726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,1024,0.008106666306654612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,128,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,2048,0.009856888817416297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,16384,0.022456000248591106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,12288,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,65536,0.07764622237947252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,7168,0.021791110436121624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,51200,2.0387634701199002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,512,0.012709333664841123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,1024,0.01773066653145684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,8192,0.10271910826365153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,12288,0.01943111088540819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,7168,0.04634488953484429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,1024,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,1536,0.05299289027849833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,768,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,12288,0.04172799984614054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,1024,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,7168,8.86740027533637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,6144,0.01725955473052131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,51200,0.08525777525371975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,1536,0.01919555498494042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,8192,0.02238399949338701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,5120,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,51200,0.05358577768007914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,2560,0.015589333242840238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,8192,0.020953777763578627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,2048,0.011397333608733283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,512,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,65536,0.2580844561258952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,1024,0.035840888818105064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,16384,0.05583644575542874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,1024,0.06733155250549316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,3584,0.22134489483303496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,2560,0.017878222796652053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,1536,0.21139110459221733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,8192,0.011157333023018308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,256,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,128,0.03425155414475335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,16384,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,512,0.013895110951529609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,4096,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,2048,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,1024,0.025935999221271936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,6144,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,16384,0.07307022147708468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,1536,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,32,0.00902844468752543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,2048,0.02737955583466424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,2560,0.03086844417783949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,3072,0.02230577833122677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,256,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,128,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,2560,0.059205333391825356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,5120,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,1536,0.037011557155185275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,1536,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,65536,0.06538399722841051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,5120,0.008512000242869059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,4096,0.06631555822160509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,768,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,65536,4.326011657714844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,256,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,51200,0.1925626728269789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,5120,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,12288,0.19068088796403673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,16384,0.30196799172295463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,7168,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,256,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,256,0.008604444563388824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,12288,0.036983112494150795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,7168,0.01368799971209632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,16384,0.018352000249756705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,128,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,2048,0.045291556252373584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,128,0.011952888634469775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,2560,0.026890666948424444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,4096,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,7168,0.9101698133680555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,256,0.002748444469438659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,65536,0.0817413330078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,12288,0.019916445016860962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,1536,0.014261333478821648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,256,0.12370310889350043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,6144,0.04031999905904134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,1536,0.02702666653527154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,5120,0.05123555660247803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,4096,0.005976888868543837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,256,0.014856888188256158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,3584,0.0042071110672420925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,128,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,5120,0.0384897788365682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,64,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,64,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,16384,0.04593955477078756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,65536,0.4153333240085178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,128,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,6144,0.054162667857276074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,4096,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,1024,0.013100444442696042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,4096,0.029848890172110662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,2560,0.00924444446961085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,7168,0.17555377218458387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,3072,0.01344622257683012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,128,0.018554665976100497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,2048,0.01183644433816274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,51200,0.356931553946601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,51200,2.426587634616428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,65536,0.6550124486287435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,6144,0.02712533374627431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,3072,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,2560,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,2560,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,10240,0.0462097790506151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,5120,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,1536,0.006441777778996362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,51200,0.09750488731596206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,1024,0.07314222388797335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,51200,0.02916533417171902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,4096,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,128,0.03269244564904107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,2560,0.016763556334707472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,3584,0.08413600259357029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,2560,0.037576887342664934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,512,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,1024,0.14155644840664333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,3584,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,16384,0.040961778826183744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,1536,0.035064889325035944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,12288,0.10345955689748128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,6144,0.11001777648925781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,8192,0.053391112221611865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,512,0.004839999808205498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,3584,0.009932444327407414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,4096,0.14560977617899576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,3584,0.029543111721674602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,6144,0.08124978012508817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,2048,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,256,0.0069057775868309875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,12288,0.04897511005401611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,7168,0.10851199759377374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,2048,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,65536,0.20285422272152373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,128,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,256,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,1536,0.021479111578729417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,512,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,2560,0.00908088849650489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,1024,0.006047111004590988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,12288,0.02162577708562215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,32,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,3584,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,12288,0.5140373441908095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,2048,0.01773155563407474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,64,0.004574222283230888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,5120,0.02772177755832672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,256,0.02717599935001797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,16384,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,256,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,2560,0.07907378011279635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,65536,0.057479110029008657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,2560,0.008408889174461365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,7168,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,12288,2.5934950510660806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,256,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,5120,0.021914665897687275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,7168,0.05267644590801663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,128,0.001977777729431788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,51200,0.7547218004862467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,2048,1.6452746921115453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,32,0.003863999827040566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,65536,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,51200,18.479468451605904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,128,0.008867555194430882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,64,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,128,0.01000088867213991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,7168,0.3285875585344103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,1024,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,51200,0.05155822303560045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,1536,0.02875111169285244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,51200,0.022738666998015508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,51200,0.6410693592495388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,8192,0.056608001391092934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,12288,0.017296888762050204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,7168,0.029845333761639063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,768,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,1536,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,3584,0.009484444227483537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,512,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,3072,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,3584,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,65536,1.317182223002116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,2560,0.009431111315886179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,4096,0.0165928883685006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,7168,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,16384,0.2653440104590522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,51200,0.34079644415113663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,16384,0.12724799580044216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,1024,0.056236445903778076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,6144,0.06814399692747328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,128,0.012910222013791403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,2560,0.014125333891974555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,64,0.008532444636027018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,3072,0.02955199943648444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,51200,0.17441599898868135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,51200,2.7396123674180775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,2048,0.02289688918325636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,1024,0.014921777778201632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,768,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,7168,0.024940444363488093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,64,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,1536,0.05126666691568163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,256,0.01550044450494978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,768,0.017521777086787753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,2048,0.004931555440028508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,5120,0.15550311406453451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,512,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,1536,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,3072,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,8192,0.7041315502590604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,51200,0.780220455593533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,32,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,768,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,4096,0.09782310989167954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,2560,0.02532088922129737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,16384,0.30796978208753795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,1024,0.03380799955791897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,2560,0.09852355718612671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,5120,0.04272622201177809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,2048,0.012425777812798819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,1536,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,2048,0.8012968699137369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,512,0.005487999982304043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,1024,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,12288,0.03186044428083632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,1024,0.05164444446563721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,7168,0.07138044304317899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,16384,0.05401244428422716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,2560,0.017969777186711628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,2048,0.012473777764373355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,65536,0.01736000014675988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,10240,0.1061848865614997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,51200,0.07305688990486993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,65536,0.058871997727288135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,10240,0.030828442838456895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,6144,0.018872888551818002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,4096,0.11167821619245742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,5120,0.00922400007645289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,512,0.02408355474472046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,3584,0.023023999399609033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,4096,0.007817777494589487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,512,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,768,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,64,0.005802666561471091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,12288,0.1918800009621514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,4096,0.008372444245550368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,16384,0.4818853272332086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,768,0.028024888700909082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,6144,0.009745777481132084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,6144,0.042881776889165245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,512,0.02218222286966112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,3584,0.005624889085690181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,6144,0.04138133260938857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,3072,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,2560,0.03936266567971971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,10240,0.15468711323208278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,51200,0.10996355613072713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,128,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,5120,0.06531644529766507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,1536,0.01276622215906779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,64,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,1536,0.04988977644178602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,5120,0.08474222156736587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,65536,0.1486168834898207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,128,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,16384,0.016656888855828177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,768,0.013516444298956128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,4096,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,65536,0.13017777601877847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,1536,0.02917333443959554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,3584,0.018998222218619455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,768,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,3072,0.015933333171738517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,1024,0.01811377704143524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,1536,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,3072,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,64,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,32,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,16384,0.05510222249560886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,12288,0.04747200012207031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,768,0.015101333459218344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,128,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,10240,0.0136817776494556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,128,0.021198223034540813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,16384,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,7168,0.04820800158712599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,2560,0.018449778358141582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,128,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,6144,0.024910221497217815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,1024,0.03144888745413886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,768,0.005646222167544895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,51200,0.08461066749360825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,16384,0.07787999841901991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,2560,0.03670933180385166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,51200,0.046929776668548584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,3072,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,1024,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,1536,0.003640000190999773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,65536,0.1555644406212701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,32,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,4096,0.036664889918433294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,768,0.021362667282422382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,512,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,256,0.01755288905567593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,10240,0.08915999862882827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,768,0.020558221472634208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,3072,0.009434666898515489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,3072,0.0734586649470859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,8192,0.06443822383880615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,1024,0.06217244598600599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,65536,0.1748328871197171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,6144,0.17570222748650444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,512,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,10240,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,2048,0.21872266133626303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,512,0.0031439999325407874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,256,0.011733333269755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,4096,0.013178666432698568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,512,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,64,0.002236444503068924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,4096,0.022554667459593877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,3584,0.08901066912545098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,10240,0.010836444795131683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,256,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,768,0.014534221755133735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,7168,0.06704266866048177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,64,0.048672888014051646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,1024,0.00869511150651508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,2048,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,8192,0.0392159985171424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,10240,0.015878222054905362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,12288,0.059501330057779946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,256,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,10240,0.01329955541425281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,1024,0.025506666964954797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,512,0.05962044662899441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,32,0.04888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,2560,0.005191111316283544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,16384,0.33771112230088973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,128,0.005872888697518243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,4096,0.04059200154410468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,3072,0.01219288922018475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,3584,0.5761146545410156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,16384,0.0362044440375434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,1024,0.0069919998447100324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,5120,0.03346844514211019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,6144,0.04402222235997518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,5120,0.011955555942323474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,512,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,12288,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,10240,0.27376090155707467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,3584,0.04029511080847846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,6144,0.009113777842786577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,65536,1.7258791393703883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,16384,0.08911999728944565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,16384,0.07721333371268378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,7168,0.2460097736782498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,1024,0.05444088909361097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,10240,0.04257333278656006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,6144,0.04394844505521986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,2560,0.039635555611716375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,7168,0.034134222401512995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,65536,0.22165332900153267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,10240,0.06446666849984063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,768,0.05187466740608215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,512,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,5120,1.1916355556911893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,10240,0.13779821660783556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,5120,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,7168,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,65536,1.1341369416978624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,128,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,16384,0.25739823447333443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,64,0.011807999677128263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,3584,0.013910222384664746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,7168,0.013898667362001208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,1536,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,6144,0.011515555282433828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,512,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,2560,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,12288,0.017104888955752056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,5120,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,3584,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,1536,0.028920888900756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,512,0.011063999599880643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,65536,0.5337840186225044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,2048,0.18188977241516113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,65536,0.08059822188483344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,768,0.011335999601417117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,3072,0.05489244394832187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,5120,0.036475555764304265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,10240,0.019063111808564927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,8192,0.07258844375610352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,2560,0.005024888863166173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,4096,0.10718578100204468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,32,0.018797333041826885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,2560,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,4096,0.02437155610985226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,256,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,3072,0.055691553486718066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,3072,0.026769777139027912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,51200,0.05783200263977051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,12288,0.09833600123723348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,51200,0.05149333344565498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,768,0.009967111051082611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,3072,0.00868711123863856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,6144,0.009594666461149851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,5120,0.029425776667065088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,256,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,10240,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,5120,0.023132445083724126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,51200,0.10814666748046875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,8192,0.03871377640300327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,256,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,7168,0.04073866539531284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,51200,0.7012177573310004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,65536,0.08628266387515598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,12288,0.01867377758026123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,128,0.027544000082545813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,64,0.00565422202150027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,32,0.0184879998366038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,65536,0.1578808890448676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,8192,0.011857777833938599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,256,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,2560,0.039472887913386025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,16384,0.432187557220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,768,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,12288,0.7481990920172797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,3584,0.04363377888997396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,5120,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,32,0.0018986666368113623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,3584,0.029706666866938274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,8192,0.1892684433195326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,1536,0.015231110983424716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,65536,0.1523253387875027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,3584,0.010790222220950656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,512,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,128,0.05179288983345032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,2048,0.020953777763578627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,3584,0.0841111143430074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,6144,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,10240,0.14840266439649794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,12288,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,10240,0.08073421981599596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,8192,0.035453332795037165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,3072,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,3072,0.3662889003753662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,16384,0.435175100962321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,512,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,10240,0.3561697800954183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,4096,0.21561422612931994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,51200,0.5490266482035319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,128,0.0022328889204396144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,128,0.02312799957063463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,256,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,4096,0.08469244506624009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,51200,0.055649777253468834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,512,0.053095112244288124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,2048,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,10240,0.01609688831700219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,8192,0.3982168833414714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,4096,0.01979377700222863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,3072,0.037752889924579196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,1536,0.07967911163965861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,12288,0.039585779110590615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,2048,0.016885333591037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,16384,0.05340177814165751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,16384,0.049040890402264066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,7168,0.023210666245884363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,10240,0.0502071115705702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,12288,0.07702844672732882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,7168,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,10240,0.04632711079385546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,4096,0.009902222288979424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,5120,0.04349866509437561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,16384,0.0376835564772288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,32,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,768,0.005691555639108022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,6144,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,5120,0.027121777335802715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,1024,0.016472889317406546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,65536,0.18223910861545137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,2560,0.015121777852376303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,7168,0.0664106673664517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,512,0.0346666673819224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,512,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,51200,6.430984073215061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,2560,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,128,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,7168,0.013398222625255585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,1024,0.01928444372283088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,256,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,1024,0.1629155609342787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,10240,0.1902853382958306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,7168,0.020482665962643094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,128,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,6144,5.947617001003689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,512,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,32,0.011927111281288994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,16384,0.03747999999258254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,5120,0.04513511061668396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,7168,1.1133404837714302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,12288,0.022592888938056097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,51200,0.13406399885813394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,2048,0.014633778068754407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,7168,0.02750844425625271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,512,0.01572444372706943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,2560,0.01703288820054796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,1024,0.020989333589871723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,1536,0.12154489093356663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,5120,0.04662222332424588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,512,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,65536,0.05079911152521769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,512,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,2048,0.01680266691578759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,1024,0.15234133932325575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,5120,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,64,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,768,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,128,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,768,0.028416001134448584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,512,0.07172355386945936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,8192,0.056286222404903836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,65536,10.156126234266493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,65536,1.1271368662516277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,8192,0.15471377637651232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,1536,0.031189332405726116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,65536,1.7748329374525282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,128,0.05723110834757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,51200,0.11543733543819851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,1536,0.10849510961108738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,2560,0.010236444572607676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,3584,0.009642666412724389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,2048,0.06674310896131727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,6144,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,4096,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,256,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,3072,0.013171555267439948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,1024,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,16384,0.5392435391743978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,8192,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,768,0.014547554983033074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,64,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,16384,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,8192,0.0225608895222346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,12288,0.01351288871632682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,768,0.0063182223174307085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,12288,0.5907573170132107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,65536,4.977064768473308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,768,0.04860799842410617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,10240,0.08756888575024074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,3072,0.04932266804907057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,16384,0.009375111096435124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,2560,0.01943555474281311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,3584,0.012830221818553077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,51200,0.15425954924689397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,1024,0.011772444678677453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,16384,0.4325262175665961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,2560,0.0166284441947937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,1024,0.019903111788961623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,5120,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,10240,0.04554488923814562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,65536,0.8014951282077365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,5120,0.016396444704797532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,12288,0.018789332773950364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,10240,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,4096,0.0721377796596951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,6144,0.09619822104771932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,16384,0.10834399859110515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,6144,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,64,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,3584,0.015240000353919135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,4096,0.008783111141787635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,1024,0.005727111051479976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,51200,0.07146400213241577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,4096,0.2086719936794705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,256,0.012447111308574677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,8192,0.10860533184475368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,512,0.013343111508422427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,1536,0.017542221479945712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,2048,0.3640417787763808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,512,0.003922666526503033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,512,0.027279110418425664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,2048,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,51200,0.04658577839533488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,51200,0.2802826563517253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,256,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,6144,0.047240889734692044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,3584,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,512,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,10240,0.06257510847515531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,512,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,10240,0.030164443784289893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,128,0.011027555498811932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,6144,0.0754586656888326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,512,0.031856000423431396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,128,0.05780888928307427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,128,0.019564444820086162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,7168,0.1415662235683865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,128,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,10240,0.0715235537952847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,3584,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,10240,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,2048,0.03432977861828274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,1536,0.016202666693263583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,1024,0.006401777681377199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,3072,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,5120,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,65536,0.1952835586335924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,8192,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,16384,0.09457866350809734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,512,0.015527110960748462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,2560,0.4703991148206923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,3584,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,32,0.04541600081655714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,16384,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,32,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,10240,0.01850666602452596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,10240,0.1438373327255249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,1536,0.05458933446142408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,512,0.45503822962443036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,256,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,256,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,16384,0.08972266647550796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,512,0.04674488968319363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,4096,0.05299199952019585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,2560,0.007100444700982835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,8192,0.010667555862002902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,32,0.0033253334048721525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,256,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,12288,0.3049537870619032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,768,0.009631111390060848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,2048,0.020831111404630873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,8192,0.0199297782447603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,768,0.01920355525281694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,6144,0.00960533320903778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,32,0.005192888693677055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,8192,0.04059022333886888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,51200,0.1440053383509318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,3072,0.046847111648983426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,65536,0.0882106688287523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,3584,0.022293332550260756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,51200,0.4042071236504449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,512,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,65536,0.05952799982494778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,6144,0.07606044742796156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,4096,0.01885955532391866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,3072,0.017690666847758822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,51200,3.4125067392985025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,65536,1.9729075961642797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,7168,0.23878399531046549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,2048,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,3584,0.1928551064597236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,768,0.021941334009170532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,512,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,8192,1.2752017974853516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,2560,0.01754666699303521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,1024,0.004237333519591225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,4096,0.021039111746682063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,2560,0.06497155295477973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,4096,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,3584,0.03552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,32,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,3584,0.019322666856977675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,65536,0.34155821800231934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,16384,3.0202354855007596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,12288,1.2572906282212999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,65536,0.08559022347132365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,10240,0.01891911029815674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,12288,0.2293777730729845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,1536,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,6144,0.015402666396564908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,5120,0.009562667045328353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,4096,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,32,0.007224889265166388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,6144,0.03020711077584161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,4096,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,768,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,10240,0.03030222323205736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,7168,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,128,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,768,0.021165332860416834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,256,0.00563466673096021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,64,0.04410399993260702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,10240,0.13183911641438803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,7168,0.00979288915793101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,6144,0.019018666611777413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,16384,0.11219910780588786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,4096,0.028174221515655518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,51200,0.06435644626617432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,3584,0.024547555380397375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,1024,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,12288,0.0526702238453759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,5120,0.012071111136012606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,512,0.006989333364698622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,512,0.020466667082574632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,768,0.04601600103908115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,1536,0.041746666034062706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,2560,0.027286221583684284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,256,0.01721333298418257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,5120,0.017493334081437852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,256,0.06371911366780598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,256,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,2048,0.5076079898410374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,2560,0.027087110612127516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,65536,0.3363520039452447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,12288,0.03581066595183478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,256,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,6144,0.09837510850694443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,7168,0.3770417902204726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,12288,0.0697724421819051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,1536,0.05033422178692288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,7168,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,2048,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,6144,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,51200,0.011910222470760345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,2048,0.04221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,1536,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,64,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,51200,0.051911999781926475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,5120,0.026199999782774184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,8192,0.12359910541110569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,10240,0.2041217750973172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,768,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,5120,0.009877333210574256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,1024,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,1536,0.28256622950236004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,4096,0.01255555533700519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,10240,0.09101600117153591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,1024,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,1536,0.017122666041056316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,8192,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,2048,0.27371732393900555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,10240,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,12288,0.07030755281448364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,768,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,6144,0.05599111318588257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,10240,0.20114044348398843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,4096,0.05695199966430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,2560,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,512,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,7168,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,512,0.01316088851955202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,4096,0.016550223032633465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,12288,0.013750221994188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,32,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,10240,0.016706667012638517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,12288,0.18168621593051484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,128,0.014921777778201632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,128,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,6144,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,2048,0.03841955463091532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,2048,0.009759999811649323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,51200,0.046153777175479464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,7168,0.01676266723208957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,12288,0.10315377844704522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,51200,0.14756533834669325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,768,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,128,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,51200,0.15975022315979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,256,0.01772888832622104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,512,0.01552177800072564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,12288,0.08095200194252862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,6144,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,65536,0.63009066051907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,512,0.05611644188563029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,128,0.0063164445261160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,6144,0.02875022093454997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,12288,0.02130488885773553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,16384,0.07325955231984456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,4096,0.058113780286577016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,8192,0.021026665965716045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,7168,0.16455911265479192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,1024,0.01957155598534478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,32,0.1966702275806003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,7168,0.012096000214417776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,768,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,7168,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,64,0.06753155257966784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,51200,0.09557244512769912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,256,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,2560,0.26196710268656415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,3072,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,768,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,1536,0.009912889036867354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,3584,0.02903466754489475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,256,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,8192,0.01978577839003669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,51200,0.03610488772392273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,4096,0.019537778364287484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,1536,0.027319111757808264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,256,0.032519999477598406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,51200,0.09961422284444173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,2560,0.08183466725879245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,16384,0.03910488883654276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,1536,0.05226044522391426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,512,0.024496000674035814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,5120,0.06499110990100436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,2560,0.010311110979980892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,7168,0.02719822194841173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,5120,0.00996977753109402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,768,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,2560,0.004942222187916438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,12288,0.0908764468299018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,16384,0.07523288991716173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,768,0.004828444371620814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,8192,0.029987557066811457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,3072,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,65536,0.06026844183603922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,256,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,768,0.008231999973456064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,8192,0.07545689079496595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,6144,0.3612711164686415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,64,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,7168,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,7168,0.034830222527186074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,10240,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,12288,0.037759111987219915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,1024,0.014856000741322836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,1536,0.018176888426144917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,512,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,12288,0.026753778258959453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,3584,0.03880533244874742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,1024,0.0037866665257347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,16384,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,128,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,1536,0.029189334975348577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,5120,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,1024,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,4096,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,7168,0.03437688946723938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,1536,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,3072,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,768,0.048446221484078296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,7168,0.16826755470699736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,512,0.0038088887102074097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,1024,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,10240,0.036855111519495644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,32,0.010854221880435944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,8192,0.03336177931891547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,16384,0.05911644299825033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,2048,0.057722667853037514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,16384,0.16896266407436797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,512,0.015862221519152325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,2560,0.006728888799746831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,65536,0.07618577612770928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,7168,0.0053119998839166425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,12288,0.038888000779681735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,5120,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,3584,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,2048,0.0982888870769077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,256,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,3584,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,512,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,768,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,1536,0.0983511143260532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,3072,0.01813066667980618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,65536,1.5707759857177734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,65536,0.48985958099365234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,128,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,128,0.02592977715863122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,256,0.026004445221689012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,10240,0.017257778180970084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,10240,0.025454221500290766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,16384,3.4674809773763022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,6144,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,2048,0.048469334840774536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,8192,6.537427690294053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,4096,0.6609448856777614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,65536,0.07862844732072619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,512,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,2560,0.01847999956872728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,8192,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,5120,0.008746667040718926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,8192,0.24446845054626465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,16384,0.18181955814361572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,1024,0.31744178136189777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,16384,0.047873778475655444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,4096,12.765111287434896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,16384,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,256,0.02032177812523312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,256,0.01555377741654714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,256,0.11257244480980767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,8192,0.9446070988972982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,3072,0.018216888109842937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,16384,0.44922224680582684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,256,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,65536,5.998446146647136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,6144,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,51200,0.7425662146674262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,4096,0.013827555709415011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,4096,0.2426719930436876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,1024,0.0032151111712058387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,32,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,1024,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,4096,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,2048,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,51200,0.577381345960829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,51200,0.15471377637651232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,2560,0.07648444175720215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,768,0.030444443225860596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,3584,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,7168,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,3584,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,65536,1.8570747375488281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,2048,0.045067555374569364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,128,0.008043555749787224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,512,0.010956444674068026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,16384,0.022384888596004907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,32,0.012649777862760754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,51200,0.40598400433858234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,512,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,2560,0.030327998929553564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,12288,0.0953635573387146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,768,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,1536,0.012311110893885294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,7168,0.009944889280531142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,7168,0.05478755633036295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,5120,0.01699022286468082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,6144,0.023317333724763658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,8192,0.13678844769795737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,10240,0.015486222174432544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,4096,0.02092444399992625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,1024,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,512,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,10240,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,4096,0.008607111043400234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,64,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,7168,0.018186666899257235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,32,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,65536,0.04399199949370491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,1024,0.018811555372344125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,8192,0.022312000393867493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,51200,0.4186835553910997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,2048,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,16384,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,8192,0.04894311229387919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,1024,0.012412444584899478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,128,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,1536,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,6144,0.009406222237481011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,256,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,8192,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,8192,0.14733510547214085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,7168,0.11302755938635932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,8192,0.025980444418059453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,6144,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,65536,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,7168,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,256,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,6144,0.014017777310477363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,7168,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,1536,0.20822400516933867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,1024,0.035767998960283064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,10240,0.015810666812790763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,1536,0.05285333262549507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,6144,0.010606221854686737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,2560,0.024822221861945257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,51200,0.13873777124616835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,1536,0.012122666670216454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,10240,0.08294044600592719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,51200,0.054042667150497437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,8192,0.0340942210621304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,128,0.10085778103934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,1536,0.03188800149493747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,7168,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,16384,0.4437199963463678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,1024,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,16384,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,3072,0.01754399968518151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,2048,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,6144,0.009570666485362584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,1024,0.0031884445084465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,4096,0.007080888582600488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,2560,0.04567111200756497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,6144,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,10240,0.017791999710930716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,7168,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,256,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,256,0.0052968888647026485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,3584,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,64,0.018107554978794523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,65536,0.0625964469379849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,51200,0.06543555524614122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,768,0.016897777716318767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,8192,0.034049779176712036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,4096,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,8192,0.03762400150299072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,64,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,4096,0.020498666498396132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,3584,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,51200,0.06156533294253879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,10240,0.009764444496896531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,768,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,64,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,4096,0.014542222023010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,6144,0.34163112110561794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,4096,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,7168,0.02825511164135403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,64,0.004461333155632019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,1024,0.039020445611741804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,2560,0.012110222544935016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,6144,0.14840532673729792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,128,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,4096,0.03578311204910278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,1536,0.06876533561282687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,2560,0.059030221568213574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,16384,0.020413332515292697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,5120,0.013742222554153867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,3584,0.07731555567847358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,51200,15.685526529947916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,6144,0.00980000032318963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,4096,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,3072,0.013600000076823764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,256,0.027111111415757075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,3584,0.009085333181752099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,8192,0.48972712622748477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,4096,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,16384,0.027845333019892376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,2048,0.1778551075193617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,4096,0.06890133354398939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,12288,0.0317191117339664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,8192,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,7168,0.011807999677128263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,10240,0.13188355498843723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,51200,0.05630489190419515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,8192,0.014030221435758801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,768,0.032962666617499456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,4096,0.008951110972298516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,10240,0.009967111051082611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,7168,0.04491733180152046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,51200,0.08006399869918823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,65536,0.06876355409622192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,12288,0.018124444617165458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,7168,0.0088746663596895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,512,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,3072,0.012034667034943899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,256,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,256,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,1024,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,10240,0.13034489419725206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,256,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,8192,0.03488889005449083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,512,0.056242664655049644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,5120,0.012459555433856117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,3072,0.01612355477280087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,256,0.003248888792263137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,7168,0.026753778258959453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,5120,0.06609333223766752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,4096,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,1536,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,1536,0.08622577455308701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,6144,0.04703022374047173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,5120,0.01884977850649092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,12288,0.11973067124684651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,256,0.0023173333870040043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,512,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,16384,0.02437866727511088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,1536,0.006734222173690796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,3584,0.01873333255449931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,4096,0.01761155492729611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,16384,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,12288,0.36846576796637637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,64,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,768,0.017952000101407368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,256,0.07450489203135173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,16384,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,1536,0.012911111116409302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,4096,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,2048,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,12288,0.02454400062561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,16384,0.1751484473546346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,256,0.032319999403423734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,3072,0.03189599845144484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,12288,0.02436888880199856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,51200,0.31978045569525826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,6144,0.13767022556728786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,2048,0.012536889149083031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,16384,0.02751111156410641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,3072,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,51200,0.0702728894021776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,51200,0.03984355595376756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,4096,0.04961688982115852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,6144,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,32,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,3584,0.013230221966902414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,2048,0.02279288901223077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,6144,0.020290666156344943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,8192,0.05058044526312086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,1536,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,128,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,16384,0.1463866631189982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,3584,0.051517334249284535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,65536,0.14539999432033962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,4096,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,1536,0.02180622186925676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,12288,0.027694221999910142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,2560,0.02757688860098521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,4096,0.11774933338165283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,512,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,1536,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,2560,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,768,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,51200,1.8975270589192708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,12288,0.06667821937137179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,10240,0.027796443965699937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,51200,0.4503991338941786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,3584,0.03667733404371474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,6144,0.02420711186197069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,1536,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,5120,0.04907822277810839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,128,0.017852443787786696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,7168,0.1862408849928114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,3584,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,12288,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,256,0.023019555542204116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,10240,0.055381334490246244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,51200,0.8511599964565701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,10240,0.012540444731712341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,5120,0.028227554427252874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,51200,1.5866017871432836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,1536,0.01724799970785777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,3072,0.025402666793929204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,6144,0.024086222052574158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,10240,0.4691182242499457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,512,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,16384,0.007241778075695038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,768,0.020430222153663635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,128,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,2560,0.05011555552482605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,5120,0.04627911249796549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,4096,0.0380328893661499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,16384,0.22697954707675508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,5120,0.011558222273985544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,1536,0.02807377775510152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,1024,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,6144,0.013011555704805585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,1024,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,3072,0.16190489133199057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,6144,0.16079733106825086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,2560,0.048617776897218495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,2560,0.2256364557478163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,512,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,12288,0.02614044480853611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,2048,0.033195555210113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,1536,0.010433778166770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,256,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,256,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,3584,0.14222933186425105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,1536,0.02555199960867564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,1536,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,128,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,1536,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,2048,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,5120,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,256,0.012774222426944308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,64,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,4096,0.11133955584632026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,51200,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,6144,0.02644355595111847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,8192,0.19223911232418486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,5120,0.009133333133326637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,128,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,10240,0.03577422102292379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,2560,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,512,0.0026559999419583213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,256,0.005316444569163852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,12288,0.023992889457278784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,8192,1.5830524232652452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,7168,0.023641778363121882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,2048,0.07779644595252143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,16384,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,65536,0.6507546636793349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,6144,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,10240,8.43783230251736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,16384,0.02569599946339925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,2048,0.011391110718250275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,65536,0.06810844606823392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,2048,0.020647111866209242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,4096,0.02812266680929396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,16384,0.7061884668138293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,64,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,2560,0.06228444311353895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,768,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,10240,0.02269511090384589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,5120,0.15579022301567927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,1536,0.03798666596412659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,12288,0.03652000096109178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,7168,0.023825777901543513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,16384,0.02110133402877384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,2560,0.04946844445334541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,32,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,512,0.013410666750537025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,256,0.028389334678649902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,51200,0.057167112827301025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,16384,0.036053333017561175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,8192,0.019549333386951022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,768,0.02785244584083557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,768,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,10240,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,256,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,12288,0.04288711150487264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,10240,0.024718221690919664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,10240,0.05286133289337158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,768,0.01702666613790724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,6144,0.9769395192464193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,1536,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,5120,0.010652444428867765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,51200,2.692811542087131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,2048,0.008059555457697975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,5120,0.04979377653863695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,7168,0.03365600109100342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,51200,6.418057759602864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,256,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,2560,0.1340222226248847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,1536,0.03124799993303087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,7168,0.030741333961486816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,3072,0.02492800023820665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,4096,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,2560,0.03785689009560479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,10240,0.01606755620903439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,7168,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,128,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,5120,0.00903555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,3072,0.2731875578562419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,5120,0.017273777061038546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,12288,0.22706845071580675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,5120,0.10571200317806667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,768,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,768,0.005711110929648082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,256,0.002258666687541538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,10240,0.02271555529700385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,5120,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,3584,0.00980888886584176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,16384,0.01623288955953386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,65536,0.1200444433424208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,5120,0.013927110367351107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,7168,0.02436711059676276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,2048,0.021724444296624925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,65536,0.08171111345291138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,51200,0.16818755202823213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,1536,0.0074577778577804565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,51200,0.05142133434613546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,3072,0.03388888968361749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,512,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,256,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,16384,0.020026667250527274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,2048,0.07742400301827325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,3072,0.013762666947311826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,256,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,64,0.0018764444523387486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,1536,0.03815199931462606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,8192,0.28974488046434194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,64,0.039600001441107854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,12288,1.1419866349962022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,4096,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,4096,0.011372444530328115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,8192,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,8192,0.013579555683665805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,12288,0.045624888605541654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,1024,0.23288178443908691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,3072,0.039812445640563965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,8192,0.02511288887924618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,7168,0.1758426692750719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,1024,0.1115466621186998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,128,0.03194222185346816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,2048,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,128,0.17631378438737658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,4096,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,512,0.0070462218589252895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,6144,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,3584,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,12288,0.043493333790037364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,10240,0.04370489054256015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,4096,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,12288,0.017626666360431247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,2048,0.004549333204825719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,5120,0.046923554605907865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,7168,0.5851466920640733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,512,0.028316444820827905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,512,0.006020444548792309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,768,0.013519110778967539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,4096,0.06997066736221313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,256,0.04318133327696058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,10240,0.05044622222582499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,1536,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,8192,0.013570666313171387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,12288,0.009947555760542551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,3072,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,7168,0.018956444329685636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,2048,0.004556444370084339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,16384,0.5824417538113064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,16384,0.09154044257269965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,12288,0.029793779055277508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,3584,0.01665155589580536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,5120,0.012312889099121094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,5120,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,8192,0.20810133881039092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,512,0.025126222107145522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,1024,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,4096,0.026135111848513286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,4096,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,5120,0.4253297646840413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,10240,0.026668444275856018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,16384,0.04377866784731547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,3072,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,3072,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,12288,0.036637332704332136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,768,0.009296889106432596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,12288,0.132189326816135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,8192,0.0159288893143336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,10240,0.015393777026070489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,4096,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,51200,0.03800266649987962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,1024,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,10240,0.021358221769332886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,1536,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,3072,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,51200,0.03780533207787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,51200,0.09851111306084527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,2560,0.025477333201302424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,1024,0.008565333154466417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,768,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,1024,0.020978666014141504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,2048,0.008785777621799046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,1536,0.010292444792058732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,16384,0.14418578147888184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,51200,0.05605510870615641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,2560,0.04680711030960083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,2048,0.06533333328035143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,3072,0.08986577722761367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,8192,0.02052088909678989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,32,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,3584,0.16120710637834337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,65536,0.3120817873213026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,12288,0.03844088978237576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,51200,13.301571316189237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,2048,0.13409245014190674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,3072,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,51200,0.2507670985327827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,3072,0.013157332936922709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,4096,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,2048,0.011681777735551199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,1024,0.030196444855795965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,3584,0.027274666561020747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,65536,0.06770310799280803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,4096,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,128,0.002401777853568395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,5120,0.03929333223236932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,768,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,10240,0.13930488957299128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,4096,0.012090666426552666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,5120,0.111099554432763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,65536,0.538315561082628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,3072,0.007835555407736037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,1536,0.04064355625046624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,16384,0.26489244567023384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,5120,0.12569777170817056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,16384,0.29034490055508083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,7168,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,512,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,64,0.001785777716173066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,512,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,4096,0.01669066647688548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,2048,0.02678311036692725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,16384,0.2625528971354167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,8192,0.035735110441843666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,128,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,1536,0.007394666473070781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,4096,0.023769777682092454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,65536,0.29144265916612416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,2048,0.02030933399995168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,4096,0.030576000610987347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,768,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,2560,0.008934222161769867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,5120,0.05715644359588623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,65536,0.75527556737264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,7168,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,7168,0.01650399963061015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,256,0.005414222263627582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,12288,0.7637386851840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,7168,0.011092444260915121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,3072,0.004570666700601578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,5120,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,10240,0.842298666636149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,3072,0.014547554983033074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,4096,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,6144,0.05731022357940674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,65536,0.8912622663709852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,2560,0.012944888737466602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,5120,0.03177422285079956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,12288,0.020682666036817763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,10240,0.04010488920741611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,65536,0.07549689213434856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,2560,0.01420177850458357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,10240,0.07100622521506415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,2560,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,6144,0.014135110709402295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,5120,0.023552000522613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,10240,0.014735110931926303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,8192,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,51200,0.24394933382670084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,2560,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,65536,0.050287110937966235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,2560,0.006904888898134232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,3072,0.06227199898825752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,512,0.021419554948806763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,16384,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,8192,0.023007111416922674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,512,0.03491377830505371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,2048,0.017451556192504037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,768,0.02660266723897722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,51200,0.02035288843843672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,4096,0.020963556236690942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,8192,0.010688888529936472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,8192,0.016436444388495553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,2048,0.05000977714856466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,65536,0.05813066826926338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,8192,0.015374221735530429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,12288,0.043271111117468945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,6144,0.012291555603345236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,1536,0.006512889017661412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,10240,0.44410043292575413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,6144,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,12288,0.06913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,51200,0.5250435405307347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,6144,0.010115555591053432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,2560,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,512,0.026029333472251892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,32,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,6144,0.011882666912343768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,128,0.0054551110499435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,3584,0.1225937737358941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,4096,0.01831733352608151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,16384,0.04137422309981452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,3072,0.02695288923051622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,12288,0.020385776956876118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,1024,0.02994489007525974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,512,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,768,0.017284444636768766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,1536,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,3072,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,1024,0.054112888044781156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,51200,0.22465599907769096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,128,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,2048,0.01684533390733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,1024,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,512,0.039164443810780845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,2560,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,3072,0.014920888675583733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,10240,0.029806223180558946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,128,0.027752000424597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,128,0.009043555292818282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,16384,0.03367111086845398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,10240,0.5717271169026693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,6144,0.1044800016615126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,16384,0.07885422309239705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,512,0.05104533169004652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,3072,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,10240,2.4756096733940973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,512,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,12288,0.06997955507702298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,6144,0.04340888725386726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,65536,0.14948622385660806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,2560,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,1024,0.026897778113683064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,64,0.01371022231049008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,5120,0.00811288919713762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,2048,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,3584,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,65536,0.1558435493045383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,2560,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,3072,0.022265776991844177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,768,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,65536,0.13423200448354086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,10240,0.02094399929046631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,512,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,2048,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,6144,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,16384,0.48911735746595597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,3072,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,3072,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,2560,0.020265777905782063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,51200,0.6651600201924642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,256,0.020153777466879953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,7168,0.04095200035307143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,65536,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,10240,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,1536,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,5120,0.010054222411579555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,7168,0.022348443667093914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,1536,0.0075777777367168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,12288,0.051389333274629384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,6144,0.017920888132519193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,7168,0.024774221910370722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,128,0.09629778067270915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,64,0.0018995555324686898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,8192,0.6204506556193033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,3584,0.10655999845928615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,65536,0.07769866784413655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,2048,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,16384,0.8835973209804959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,10240,0.0332293344868554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,6144,0.03161777721510993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,128,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,2560,0.005364444520738389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,8192,0.03050400151146783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,1536,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,3584,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,768,0.20375555091434053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,51200,0.0598542226685418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,128,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,32,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,3584,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,16384,0.07854044437408447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,6144,0.03647289011213515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,8192,0.17203911145528158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,65536,0.08705777592129177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,4096,0.01091288857989841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,768,0.04777422216203478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,12288,0.15610844559139675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,10240,0.02853333287768894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,6144,0.17754933569166395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,8192,0.03374844458368089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,12288,0.05424355467160543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,64,0.0041848888827694785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,3072,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,3584,0.027257778578334387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,3072,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,7168,0.013449778159459433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,32,0.014854222536087036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,1536,0.007901333272457123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,3584,0.018522666560279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,2560,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,5120,0.021344888541433547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,12288,0.013285333083735572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,768,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,1024,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,6144,0.07508622275458442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,128,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,51200,0.017145777742067974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,16384,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,1024,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,7168,0.04974222183227539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,256,0.030352890491485596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,4096,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,10240,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,8192,0.010399999717871347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,1024,0.01310755560795466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,6144,0.012990222209029727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,3072,0.026334222820070054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,2560,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,5120,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,10240,0.015806222955385845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,16384,0.1921964486440023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,51200,0.06746577554278903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,768,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,3072,0.06491200129191081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,64,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,1024,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,5120,0.0384133325682746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,512,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,64,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,51200,0.1944648954603407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,65536,1.2656710942586262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,6144,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,7168,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,12288,0.029495110114415485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,12288,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,6144,0.2963457902272542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,8192,0.015821332732836407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,16384,0.0494391123453776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,3072,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,2560,0.010806221928861407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,512,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,32,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,3072,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,7168,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,7168,0.20797510941823324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,3072,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,256,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,10240,0.026385777526431616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,1536,0.014004444082578024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,768,0.003524444583389494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,768,0.0692382189962599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,3072,0.26918578147888184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,128,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,32,0.022103110949198406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,8192,0.550414244333903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,256,0.005408888889683618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,4096,0.033258666594823204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,65536,0.5712711016337076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,3584,0.013627555635240344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,6144,0.11629066202375625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,6144,0.029679109652837116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,2048,0.017467555072572496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,6144,0.04639022217856514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,1536,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,10240,0.05635822481579251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,65536,0.3330373234219021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,16384,0.0498524440659417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,5120,0.19396889209747314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,256,0.021539555655585393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,1536,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,10240,0.19755021731058756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,1024,0.04124888777732849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,64,0.023754666248957317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,8192,0.015962666935390897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,7168,0.014177777700954013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,128,0.00425511101881663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,3584,0.10559199915991889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,32,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,5120,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,6144,0.2856169011857775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,6144,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,8192,0.04954222175810072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,1536,0.11162755224439834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,2560,0.032925334241655134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,51200,23.877710978190105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,51200,0.657950242360433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,1024,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,512,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,5120,0.017080888152122498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,32,0.004173333446184794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,5120,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,7168,0.19654222329457602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,10240,0.04047022263209025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,10240,0.020246222615242004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,128,0.010565333068370819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,2048,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,2048,0.007095110913117726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,768,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,10240,0.03134844369358487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,10240,2.4454417758517795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,2048,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,7168,0.0188737776544359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,16384,0.04036622246106466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,65536,2.579377704196506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,10240,0.038895110289255776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,8192,0.04297955499755012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,64,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,7168,0.01865511139233907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,5120,0.09561333391401504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,6144,0.018427555759747822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,1024,0.0032391111469931076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,3584,0.12957333193884954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,12288,0.1661191119088067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,8192,1.4568942387898762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,7168,0.014891554911931356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,65536,0.44443644417656797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,16384,0.03557599915398492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,256,0.002250666626625591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,128,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,3072,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,1024,0.02773244513405694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,8192,0.019928000039524503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,128,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,12288,0.10955378082063462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,10240,0.16547644138336182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,51200,0.08175644609663221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,51200,0.42098222838507754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,6144,0.3540746635860867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,12288,0.03739555676778158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,3584,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,1536,0.011007111105653973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,2560,0.004804444395833545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,16384,0.018370666437678866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,256,0.01371911085314221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,1024,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,2048,0.07362844546635945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,3584,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,2048,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,12288,0.016709332664807636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,6144,0.035247110658221774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,51200,0.5578933291965061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,5120,0.07090400324927436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,128,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,2560,0.051058666573630444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,1024,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,12288,0.05429777834150526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,1536,0.012897777888509961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,1024,0.011025778121418424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,7168,0.030384000804689195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,4096,0.08570755852593316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,4096,0.06346133020189074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,6144,0.00963288876745436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,1024,0.019710222880045574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,12288,0.04622133241759407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,128,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,512,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,10240,0.028565333949195013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,3584,0.012061333490742577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,3072,0.050683554675843984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,6144,0.0613644454214308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,1536,0.023063111636373732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,51200,0.698912885453966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,5120,0.03118844495879279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,5120,0.11657510863410102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,512,0.008041777544551426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,256,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,5120,0.02100533412562476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,1536,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,2560,0.04439289040035672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,1536,0.022655111220147874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,1536,0.03496355480617947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,256,0.02054755555258857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,8192,0.012836444709036084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,32,0.023455111516846552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,512,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,1024,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,3072,0.11205422878265381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,10240,0.020231111182106864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,64,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,512,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,3584,0.04336888922585381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,2048,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,2048,0.013081777426931592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,1536,0.0034906665484110513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,12288,0.44933334986368817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,12288,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,65536,0.5407582388983833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,8192,0.16596799426608616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,10240,0.13707021872202554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,1024,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,65536,0.15492799546983507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,3072,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,1024,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,51200,0.05424088901943631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,256,0.006395555618736479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,768,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,128,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,2560,0.013738666971524557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,2560,0.017824000782436795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,5120,0.04212533434232076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,7168,0.46241066190931535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,4096,0.03393777873780992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,3584,0.022822222775883142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,10240,0.036710222562154136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,512,0.013905778527259827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,1536,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,65536,0.15259644720289442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,10240,0.02833599845568339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,16384,2.837757322523329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,2048,0.08040089077419706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,1536,0.10882755782869126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,768,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,12288,0.020212444994184706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,1024,0.18251378006405303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,256,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,2048,0.023611555496851604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,768,0.008736000292830998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,4096,0.08612533410390218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,51200,0.17095643944210479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,3072,0.018933332628673978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,1536,0.045888887511359326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,6144,0.07001511255900066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,3584,1.0316319995456271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,1024,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,8192,0.07368355327182345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,128,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,10240,0.13573243882921007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,51200,6.841883765326606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,2560,0.033519999848471746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,2048,0.009233777721722921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,7168,0.022375999225510493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,1536,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,1024,0.023846222294701472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,65536,0.6648337576124403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,8192,0.03849066628350152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,768,0.05193688803248935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,6144,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,4096,0.17217332786983916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,10240,0.12550844086541071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,3072,0.08161866664886475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,2560,0.14069867134094238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,1024,0.020392000675201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,4096,0.07993777592976888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,3072,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,1536,0.06827644507090251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,65536,0.9106782277425131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,16384,0.015850666496488783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,6144,0.04120711154407925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,128,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,4096,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,1536,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,4096,1.1466888851589627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,10240,0.2602248986562093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,16384,0.04901866780387031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,1536,0.03342133429315355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,3584,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,128,0.01144266708029641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,12288,0.22927734586927626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,12288,0.5373635292053223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,2560,0.013479111095269522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,10240,0.19819200038909912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,8192,0.07843377855088976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,3584,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,51200,0.13940710491604275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,1536,0.05180799961090088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,3584,0.007920888562997183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,128,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,3072,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,1536,0.014163555370436775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,256,0.01661600006951226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,6144,0.024767110745112102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,7168,0.04671377936999003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,10240,0.026120889517996047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,4096,0.0326533317565918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,512,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,1024,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,2048,0.03201599915822347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,10240,0.02898933490117391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,512,0.012488889197508493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,2048,0.1129208935631646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,10240,0.03400711218516032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,1536,0.5302382045321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,16384,0.06198755237791273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,768,0.009534222384293875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,3584,0.01698755555682712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,10240,0.05191377798716227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,3072,0.012792000340090858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,12288,0.015048000547620984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,10240,0.028106666273540918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,512,0.011260444091426002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,16384,0.027239999837345544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,7168,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,51200,2.605959150526259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,51200,0.39039111137390137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,768,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,3072,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,12288,0.06255910793940227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,256,0.014524444937705994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,4096,0.014838222000333997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,1024,0.14231999715169272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,7168,0.01959111127588484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,256,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,16384,0.010780444575680627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,8192,0.4478613535563151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,10240,0.1856844425201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,256,0.06572977701822917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,65536,0.34773866335550946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,6144,0.14527111583285862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,2048,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,12288,0.017510222064124215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,16384,0.023839111129442852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,32,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,51200,0.03078311019473606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,1024,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,1536,0.055032001601325146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,2560,0.027266666293144226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,1536,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,16384,0.039565334717432656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,4096,0.04281155599488152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,12288,0.01685422162214915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,128,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,5120,0.01587733295228746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,10240,0.11640889114803737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,32,0.010048888623714447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,7168,0.05067466696103414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,65536,0.07230044735802545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,5120,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,128,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,256,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,5120,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,3584,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,8192,0.20099733935462105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,2560,0.026704889204767015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,1024,0.08789689011043972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,65536,0.49504179424709743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,65536,0.011584888729784222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,128,0.04832177691989475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,12288,0.04031733340687222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,8192,0.1298524406221178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,5120,0.03593600127432082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,8192,0.0433573325475057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,512,0.03461777832772996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,5120,0.021799999806616042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,3072,0.042577776643964976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,128,0.009083555804358589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,1536,0.024718221690919664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,4096,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,8192,0.1732097731696235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,32,0.0019511110666725372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,1536,0.11408444245656331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,12288,0.04790044493145413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,64,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,2560,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,2560,0.20274578200446236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,2048,0.041073779265085854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,65536,0.14616889423794216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,65536,0.06515110863579644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,51200,1.0315377977159288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,16384,0.06038933330112033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,2048,0.032234665420320295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,5120,0.16642667187584773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,512,0.07212444146474202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,3072,0.03523911039034525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,51200,0.16862311628129747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,256,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,2560,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,10240,0.21352977222866484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,1024,0.0597591135236952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,51200,4.6032460530598955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,7168,0.16014844841427273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,5120,0.09837866491741604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,16384,0.029155555698606703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,7168,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,6144,0.23490577273898652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,6144,0.5266088909573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,12288,0.014096000128322177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,512,0.011139555937714048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,65536,0.10116622183057998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,768,0.01683200067943997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,6144,0.03998577925893996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,128,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,12288,0.04609422220124138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,3072,0.021624000536070928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,256,0.010772444307804108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,768,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,51200,0.04043911231888665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,5120,0.028629332780838013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,51200,0.05901955233679878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,768,0.015056000815497505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,5120,0.07616977559195624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,256,0.013791110780504016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,16384,0.062352001667022705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,1536,0.021219554874632094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,128,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,768,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,65536,0.5898595386081272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,3584,0.013265777793195514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,6144,0.011072888970375061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,7168,0.017678222722477384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,32,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,65536,0.9796071582370334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,16384,0.5963217947218153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,256,0.02399377855989668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,51200,0.4353920088873969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,256,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,10240,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,2048,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,12288,0.09348889191945393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,10240,0.1161457830005222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,1024,0.053080889913770885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,51200,0.440720001856486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,2560,0.11534044477674697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,3584,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,256,0.0968559980392456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,5120,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,128,0.10184355576833089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,10240,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,1024,0.01994311147265964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,10240,1.0257529152764213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,16384,0.17918754948510063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,3072,0.03176000052028232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,2048,0.019323555959595572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,65536,0.06570755773120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,768,0.03507466779814826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,32,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,5120,0.013551111022631327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,12288,0.02307022280163235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,6144,0.043439110120137535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,10240,0.010168000227875179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,768,0.039671109782324895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,3584,0.06302755408816867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,128,0.01792266633775499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,1536,0.022110222114457026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,4096,0.040419555372662015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,2048,0.050245334704717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,4096,0.025527111358112756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,12288,0.13992267184787327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,512,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,8192,0.1491546630859375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,2560,0.008719999757077958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,51200,0.3002924389309353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,5120,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,16384,0.28419023089938694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,128,0.010327111515733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,32,0.02085777786042955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,768,0.003228444399105178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,3072,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,12288,0.049192888869179614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,128,0.011767999993430244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,6144,0.10463733143276638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,6144,0.009639110830095079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,3072,0.007243555453088548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,128,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,10240,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,16384,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,1536,0.009892444643709395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,3072,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,128,0.00544088871942626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,3584,0.13605422443813747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,12288,0.03688888748486837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,3584,0.00778311077091429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,3072,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,16384,0.2811173333062066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,4096,0.018969777557584975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,8192,0.04948711064126757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,2560,0.03008177876472473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,512,0.0032462223122517266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,7168,0.3982862101660834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,51200,0.4963724348280165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,128,0.022954665952258643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,256,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,10240,0.023357333408461675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,1536,0.005984000033802456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,8192,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,2560,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,12288,0.08594933483335708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,512,0.02146044373512268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,3072,0.018498667412334018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,1536,0.008400888906584846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,768,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,128,0.07581510808732775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,51200,0.07001422511206733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,3584,0.006728000111050076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,16384,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,5120,0.0096133334769143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,768,0.007341333561473423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,768,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,10240,0.4681848949856228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,12288,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,51200,0.10271377695931329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,1024,0.025438222620222304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,256,0.0023413333627912733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,64,0.005398222141795688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,7168,2.9406799740261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,2560,0.012425777812798819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,65536,0.4656577640109592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,32,0.003859555555714501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,3072,0.019648889700571697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,512,0.015784000356992085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,1536,0.008374222450786168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,768,0.012110222544935016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,16384,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,5120,0.6515919897291396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,256,0.017589333984586928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,5120,0.012272888587580787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,16384,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,1024,0.009528888596428765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,51200,0.05766222212049696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,16384,0.43993245230780703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,7168,0.034874667723973594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,4096,0.025056888659795124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,128,0.010330666270520952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,256,0.02862488892343309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,6144,0.01110222190618515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,4096,0.27511999342176646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,12288,0.013425778183672162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,5120,0.04775466521581014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,8192,0.8257599936591254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,8192,0.022791999909612868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,3584,0.023249778482649062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,7168,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,1024,1.380465825398763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,7168,0.019270223047998216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,2048,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,32,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,3072,0.004264889078007804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,128,0.022280888424979314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,16384,0.47933064566718203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,8192,0.019267555740144517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,10240,0.022227555513381958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,10240,0.046104888121287026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,16384,0.048002666897243924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,512,0.0027431110954946945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,16384,0.10221866766611735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,64,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,6144,0.01166133334239324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,1536,0.006710222197903528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,2048,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,3072,0.029419554604424372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,10240,0.16064533922407362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,2048,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,512,0.01447111037042406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,512,0.02054577734735277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,5120,0.012574222352769641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,512,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,12288,0.019567999574873183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,256,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,128,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,64,0.005528888768619961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,5120,0.025449777642885845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,6144,0.018624888526068795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,128,0.014295111099878946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,16384,0.0868684450785319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,51200,0.04433955417739021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,512,0.0221413340833452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,2048,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,256,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,12288,0.07488622268040974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,256,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,16384,0.6941644350687662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,5120,0.09273688660727607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,3584,0.025423111187087163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,1536,0.004049777984619141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,12288,0.10433955325020684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,2560,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,7168,0.0648204419347975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,1024,0.013569778038395775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,65536,0.3593875567118327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,6144,0.19931999842325845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,1536,0.02504355543189579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,16384,0.01902933253182305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,1024,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,512,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,3584,0.017190222938855488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,128,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,3584,0.03437066740459866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,256,0.0025457777082920074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,7168,0.036914666493733726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,8192,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,128,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,2560,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,7168,0.0517119997077518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,10240,0.02404977712366316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,5120,0.06921777460310194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,1536,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,64,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,2560,0.08541422420077854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,2560,0.02401955591307746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,256,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,2048,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,10240,0.026483555634816486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,7168,0.021176000436147053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,4096,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,5120,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,256,0.006666666517655055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,65536,0.07413599888483684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,16384,0.042374223470687866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,128,0.0022657778528001574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,768,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,64,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,1536,0.0414008895556132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,128,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,10240,0.043459554513295494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,4096,0.010770666930410596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,3584,0.010743111371994019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,64,0.005985777825117111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,7168,0.027432888746261597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,512,0.007053333024183909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,7168,0.04502400093608432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,32,0.012111110819710625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,256,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,65536,0.07298133108350965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,10240,0.018752000398106046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,8192,0.016562667157914903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,16384,0.9137919743855795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,6144,0.03238577644030253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,3072,0.03787555628352695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,2048,0.017141333884663053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,4096,0.02814488940768772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,128,0.02093422247303857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,8192,0.02459466622935401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,2048,0.004287111262480418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,2560,0.03536622060669793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,5120,0.008445333275530074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,3072,0.019110222657521565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,1536,0.027903109788894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,65536,0.06530577606625028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,16384,0.10005778074264526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,128,0.017981333865059745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,16384,0.02887822190920512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,51200,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,12288,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,6144,0.034648001194000244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,10240,0.017059556312031217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,32,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,10240,0.6182977888319228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,10240,0.015103111664454142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,6144,0.6605511241488987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,6144,0.013219555219014486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,768,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,4096,0.008380444513426887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,12288,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,2560,0.018908444378111098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,65536,0.5661537912156847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,7168,0.07656000057856242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,256,0.01111822244193819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,2048,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,6144,0.011751111182901593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,10240,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,2048,0.008565333154466417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,6144,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,1024,0.00509688879052798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,5120,0.0889822244644165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,3584,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,512,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,2560,0.01651022169325087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,1024,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,10240,0.029509332444932725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,128,0.004447111239035924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,8192,0.022115555074479844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,12288,0.3224293390909831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,12288,0.01925244430700938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,2560,0.01332355539004008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,7168,0.03070844544304742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,1536,0.00720355576939053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,32,0.004363555461168289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,1024,0.11557866467369927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,51200,0.06631822056240506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,12288,1.5963066948784723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,768,0.029116445117526587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,8192,0.04667466547754076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,2560,0.03609866566128201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,51200,0.7712835735744901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,6144,0.10393777820799087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,8192,0.004534222185611725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,32,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,768,0.00869777798652649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,4096,0.027496889233589172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,256,0.013156444662147097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,5120,0.1534008847342597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,768,0.014876445134480795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,2560,0.042337778541776866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,51200,10.064066569010416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,2048,0.022379555635982092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,51200,0.7186915609571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,16384,0.3040737840864393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,128,0.018915555543369718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,65536,0.9099440044826932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,4096,0.020679111282030743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,768,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,128,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,256,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,12288,0.10312444633907741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,2048,0.011587555209795633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,6144,0.05413688884841072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,1536,0.4048524432712131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,16384,0.041371554136276245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,5120,0.03882311118973626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,51200,0.03681955403751797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,3072,0.02774488925933838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,65536,0.10584977600309585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,16384,0.05622577667236328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,16384,0.021609778205553692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,128,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,64,0.007744000189834171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,65536,0.08790133396784465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,5120,0.011639110743999481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,256,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,16384,0.058938668833838574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,65536,0.5979679955376519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,128,0.018575999471876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,4096,0.009492444495360056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,16384,0.14711199866400823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,1536,0.017254221770498488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,2048,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,512,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,256,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,1536,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,16384,0.39675201310051816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,2560,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,8192,0.030662222041024104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,1536,0.02970400121476915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,64,0.003140444349911478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,4096,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,128,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,65536,3.3839679294162326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,1024,0.011780444118711682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,6144,0.054455998871061534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,3584,0.04474222328927782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,4096,0.05124533176422119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,7168,0.048482666412989296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,256,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,10240,0.02677333354949951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,256,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,3584,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,4096,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,2048,0.006392000036107169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,10240,0.02778844369782342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,12288,0.9519253836737739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,5120,0.050042665667004056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,768,0.04452088806364271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,6144,0.01624000072479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,4096,0.027055111196306016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,2560,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,10240,0.8216542137993706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,4096,0.027911110056771174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,3072,0.023578666978412207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,768,0.0038319999972979226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,256,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,768,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,8192,0.07924533552593656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,10240,0.055778668986426465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,8192,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,1536,0.01186044431395001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,7168,1.3769031100802953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,2560,0.008217777642938826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,512,0.007148444652557373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,1024,0.05331200030114916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,3584,0.027457776996824477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,1024,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,3584,0.060757332377963595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,256,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,32,0.001791999985774358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,16384,0.019869332512219746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,7168,0.21838843822479248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,768,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,2048,0.006167999986145232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,512,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,2048,0.006554666492674086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,8192,0.11636266443464492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,5120,0.009528888596428765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,7168,0.05340355634689331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,7168,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,3072,0.1636026700337728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,3584,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,10240,0.07384622097015381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,51200,0.04074933462672763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,128,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,7168,0.04128888911671109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,32,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,5120,0.02626222206486596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,512,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,768,0.11425422297583686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,51200,0.4018897745344374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,51200,0.04902044600910611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,256,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,2048,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,5120,0.30575288666619194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,12288,0.48682133356730145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,256,0.010738666686746808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,2048,0.6286471154954698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,1536,0.0038746665749284956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,51200,0.05533155467775133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,256,0.00573333352804184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,10240,0.01200355589389801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,5120,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,51200,0.15893956025441489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,128,0.016743110285864938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,65536,0.8597457673814562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,4096,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,768,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,12288,0.01586577792962392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,3072,0.06521066692140368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,12288,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,1024,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,128,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,51200,8.41507297092014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,12288,0.030028445853127375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,1536,0.1648577716615465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,10240,0.02094488839308421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,12288,0.08897155523300171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,2048,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,65536,0.07728444205390082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,16384,0.2683680057525635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,32,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,3072,0.013927999469969006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,6144,0.05745066536797417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,3584,0.07977510823143853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,65536,3.2719688415527344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,3072,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,768,0.03273244367705451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,3584,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,64,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,16384,0.4977128770616319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,768,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,32,0.0063680000603199005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,2048,0.030973331795798406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,65536,0.08629510800043742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,10240,0.01237511138121287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,5120,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,16384,0.055191109577814736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,3072,0.040582223070992365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,256,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,1024,0.027095110880004034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,1024,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,8192,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,1024,0.0398231115606096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,32,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,128,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,8192,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,768,0.033095998896492854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,7168,0.13198044564988878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,51200,1.2760426203409831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,2048,0.03404622276624044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,3584,0.035339554150899254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,6144,0.05287466777695549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,1024,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,7168,0.033257779147889875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,6144,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,64,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,3072,0.00920800036854214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,8192,0.008959111240175035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,1536,0.003899555653333664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,51200,1.093939569261339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,5120,0.016367110941145156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,2048,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,6144,0.015857777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,32,0.016813332835833233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,512,0.013068444199032254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,7168,0.02625333269437154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,2048,0.011427555647161273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,1536,0.030869334936141968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,3584,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,2560,0.019479110836982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,128,0.11000621981090969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,6144,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,6144,0.012039999994966718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,32,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,1024,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,16384,0.02548711167441474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,64,0.0018888889915413326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,3072,0.4668960041469998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,4096,0.0687004460228814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,256,0.008596444295512306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,64,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,768,0.037309332026375666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,8192,0.4843875567118327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,51200,0.1873617834515042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,10240,0.38454755147298175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,2048,0.02156266735659705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,3584,0.008439111212889353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,12288,0.11780444780985515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,2048,0.010774222513039907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,7168,0.03746577766206529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,128,0.026710222164789837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,1536,0.020523554748959012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,6144,0.0818693306710985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,32,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,8192,0.2684275574154324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,51200,0.11493778228759766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,10240,0.01592088904645708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,3072,0.01035288886891471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,2048,0.017295110556814406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,65536,0.0722293323940701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,1536,0.09466311004426743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,3072,0.08167911238140531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,2048,0.005040888984998067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,3072,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,65536,0.2228684425354004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,51200,0.8518808682759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,32,0.17342844274308947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,1024,0.01553422212600708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,8192,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,1536,0.010905777414639791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,12288,1.0424924426608615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,3072,0.1770319938659668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,256,0.020865778128306072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,4096,0.024643555283546448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,128,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,8192,0.17260532908969453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,256,0.016504888733228047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,6144,0.01332711097266939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,256,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,2048,0.0207004447778066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,10240,0.07184266381793551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,51200,0.09267110957039727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,7168,0.05713511175579495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,768,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,12288,0.049253334601720176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,8192,0.012724444270133972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,512,0.022348443667093914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,6144,0.6943128903706869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,10240,0.034486220942603216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,8192,0.12498044967651367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,51200,0.05606400304370456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,3584,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,4096,0.00924355536699295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,16384,1.9870018429226344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,128,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,6144,0.023686221904224817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,3072,0.05579377545250786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,5120,0.059303998947143555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,1536,0.05147288905249702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,3072,0.13939555486043295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,256,0.005378666851255629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,3584,0.021949332621362474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,8192,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,16384,0.08938488695356582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,2560,0.08790755271911621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,10240,1.2561101913452148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,51200,0.43279110060797793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,8192,0.011207111179828644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,6144,0.030831111801995173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,2048,0.03128355410363939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,3072,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,12288,0.02160266704029507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,65536,0.06741066773732503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,64,0.009730666875839233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,1536,0.01814311080508762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,1536,0.07716000080108643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,10240,0.05528266562355889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,7168,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,3584,0.032050665881898664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,128,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,768,0.018181333939234417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,32,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,512,0.0057297779454125305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,768,0.07067110803392199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,16384,0.09471288654539321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,1024,0.021329777108298406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,7168,0.9534390767415365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,3072,0.024756444825066462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,51200,2.4561644660101996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,32,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,64,0.009303111169073317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,2048,0.09855200184716119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,16384,0.2359671062893338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,12288,0.11355111334058975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,16384,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,256,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,16384,0.033003555403815374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,10240,0.15407910611894396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,512,0.014839111102951897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,128,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,16384,0.06729777654012044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,5120,0.018929777873886954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,51200,5.445017920600043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,10240,0.04351288742489285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,1024,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,6144,0.2399288813273112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,3584,0.03700622253947788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,5120,0.008722667064931657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,4096,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,12288,0.04051822092798021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,12288,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,7168,0.12886844740973577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,4096,0.053469333383772105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,128,0.007100444700982835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,51200,0.019581332802772522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,2048,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,2560,0.01869511107603709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,65536,0.6158711115519205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,65536,0.2531217734018962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,7168,0.023592889308929443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,1536,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,3072,0.026293334033754136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,1536,0.03068711029158698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,128,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,5120,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,256,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,51200,6.680986616346572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,10240,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,256,0.0058648888435628675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,7168,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,4096,0.012084444363911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,3584,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,1024,0.24464000595940483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,3584,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,64,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,16384,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,1024,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,6144,0.01750577820671929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,65536,0.1615040037367079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,2048,0.04636889033847385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,3584,0.011937778029176923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,2560,0.05438311232460869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,3584,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,51200,1.7785672081841364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,3584,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,768,0.1412737766901652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,51200,1.3786719640096028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,256,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,12288,0.025447999437650044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,64,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,128,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,10240,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,6144,0.02221155497762892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,768,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,51200,1.4970568550957575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,8192,0.13017511367797852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,256,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,6144,0.14058844248453775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,10240,0.21723555194007027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,3072,0.016308445069524977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,3584,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,2560,0.011559111376603445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,4096,0.04363200068473816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,2560,0.013584000368913015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,256,0.01484444406297472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,512,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,512,0.007136888802051544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,768,0.0063991112013657885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,12288,0.044758223825030856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,32,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,768,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,8192,0.022612444228596155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,1536,0.05201066533724467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,12288,0.014746665954589844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,51200,0.055408888392978244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,16384,0.01850755512714386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,51200,0.784637345208062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,2560,0.022583999567561682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,4096,0.48134130901760525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,7168,0.028582221931881372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,512,0.012293332980738746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,2048,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,51200,0.06463555494944255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,5120,0.03843644592497084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,2048,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,8192,2.4840337965223527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,10240,0.03409333361519708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,128,0.005705777969625261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,8192,0.09172177977032131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,3584,1.505967140197754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,5120,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,16384,0.24968979093763563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,2048,0.011518222590287527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,6144,0.12707822852664524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,2560,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,5120,0.04179911149872673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,2048,0.08407644430796306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,32,0.004972444640265571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,65536,0.09120355712042914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,2560,0.03310399916436937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,6144,0.043024000194337636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,51200,1.4067049026489258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,64,0.012468444804350534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,1536,0.022451554735501606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,3072,0.014073777529928418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,65536,0.11356177594926621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,6144,0.03191555539766947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,32,0.006675555474228329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,1536,0.028736889362335205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,2560,0.004598222259018156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,51200,0.052469333012898765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,4096,0.011722666521867117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,3584,0.10333866543240017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,5120,0.052757332722345986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,1536,0.07001600000593397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,51200,0.038049777348836265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,1024,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,3072,0.011861333416567909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,8192,0.02511288887924618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,4096,0.02937600016593933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,4096,0.031390223238203264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,12288,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,1536,0.024951110283533733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,5120,0.015579554769727917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,1536,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,8192,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,4096,0.03844088978237576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,3072,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,7168,0.07322488890753852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,1536,0.03706577751371596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,8192,0.24512799580891928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,5120,0.014775110615624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,3072,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,65536,0.09616177611880833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,256,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,1536,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,65536,0.07027110788557264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,5120,0.02158222264713711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,128,0.014631999863518609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,3584,0.00941955546538035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,2560,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,3072,0.01447111037042406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,12288,0.04334311021698845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,128,0.08984266387091742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,256,0.019627556204795837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,12288,0.013425778183672162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,16384,0.07645955350663927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,3584,0.01863822175396813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,65536,1.6818568971421983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,3584,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,12288,0.05521777934498257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,2560,0.01311288856797748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,512,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,5120,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,3072,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,128,0.01182577759027481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,10240,0.11570489406585693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,2048,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,8192,0.03775733378198411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,51200,4.063056098090278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,51200,0.11714577674865723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,10240,0.06467644373575847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,128,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,12288,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,4096,0.030639999442630347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,51200,0.06889866458045112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,65536,2.1426773071289062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,51200,0.10416089163886176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,2560,0.029208001163270738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,5120,0.026399110754330952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,5120,0.561539543999566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,5120,0.04309511184692383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,8192,0.03713599840799967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,8192,0.031383110417260066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,65536,0.1499848895602756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,3584,0.711335129208035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,10240,0.05628355344136556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,2560,0.016674665941132438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,2048,0.04343733191490173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,6144,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,3584,0.038444442881478205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,6144,0.07169244686762492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,1024,0.045838223563300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,3584,0.016923555069499545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,10240,0.016393777396943834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,7168,0.026104888982243005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,1024,0.0038977778620190094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,5120,0.060580445660485156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,8192,0.012512000070677863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,3072,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,8192,0.01493866741657257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,1536,0.019912888606389362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,5120,0.03532355692651536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,65536,1.1064586639404297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,256,0.022648000054889258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,10240,0.07336088683870104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,768,0.004254222330119875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,512,0.06333955791261461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,12288,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,2560,0.010490666660997601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,128,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,512,0.006065777606434292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,4096,0.004563555535342959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,2048,0.00619555554456181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,3072,0.013740444348918067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,16384,1.396167967054579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,768,0.017151110702090792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,128,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,16384,0.018391110830836825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,256,0.011920000116030375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,5120,0.24741066826714408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,1536,0.012829333543777466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,1536,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,512,0.012448888685968188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,3072,0.02164711058139801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,16384,0.3289955457051595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,256,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,4096,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,8192,0.015186667442321777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,128,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,256,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,512,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,8192,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,8192,0.03867288761668735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,65536,0.08385688728756374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,256,0.012440000143316058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,7168,0.023362666368484497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,12288,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,5120,0.32823999722798664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,10240,0.2565164566040039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,2048,0.024863999750879075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,12288,0.3575333224402533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,65536,0.09572444359461467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,1024,0.005994666781690385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,65536,0.34880977206759983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,5120,0.01000711073478063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,128,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,7168,0.8005848990546333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,5120,0.04703377683957418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,6144,0.11927555667029487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,6144,0.029101332028706867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,1024,0.016055110428068373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,1536,0.008086221913496653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,12288,0.029904001288943823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,64,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,12288,0.013829333086808523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,65536,2.5950684017605252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,51200,0.2453626791636149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,8192,0.05300977826118469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,4096,0.01886755559179518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,4096,0.2157448927561442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,768,0.02754488918516371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,1024,0.008932444784376357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,8192,0.03164533442921109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,5120,0.07172355386945936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,3584,0.11529421806335449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,64,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,128,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,512,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,8192,0.07282844516966078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,4096,0.10095733404159546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,3584,0.022036444809701707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,51200,0.023265777362717524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,768,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,12288,0.05008977651596069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,1024,0.008586666650242275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,512,0.012552888856993781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,65536,0.16755199432373047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,32,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,5120,0.04813511172930399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,7168,0.07492799891365899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,12288,0.022022222479184467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,768,0.023615999354256526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,2048,0.007400000260935889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,32,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,51200,0.14085422621832952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,65536,1.9556889004177516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,5120,0.04013777772585551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,256,0.0314995547135671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,2048,0.007239110767841339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,16384,0.02164977788925171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,65536,0.11132178041670059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,32,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,32,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,512,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,32,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,2048,0.12444267008039687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,512,0.018076444665590923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,3584,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,6144,0.022447999980714586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,2048,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,4096,0.028374221589830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,1536,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,64,0.01899377836121453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,7168,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,6144,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,768,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,2048,0.01016533292002148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,32,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,128,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,7168,0.033032000064849854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,128,0.018230222993426852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,2048,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,512,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,65536,0.07072710990905762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,51200,1.110896004570855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,3072,0.03586577706866794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,51200,0.6845919820997449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,8192,0.03490755624241299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,1536,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,1024,0.12173688411712646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,3072,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,3072,0.0739884442753262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,16384,0.03172977765401205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,128,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,3584,0.013057777451144325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,10240,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,256,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,10240,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,256,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,12288,1.1387066311306424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,2048,0.03809955716133118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,65536,0.05934400028652615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,1536,0.020063110523753695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,256,0.010989333192507425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,3072,0.025384000606007043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,65536,10.955041673448351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,5120,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,256,0.03709955679045783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,32,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,10240,0.06608355707592435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,16384,0.02358666724628872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,2560,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,1536,0.006072000082996156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,5120,0.026309332913822595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,10240,1.9315653906928167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,1536,0.01770755483044518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,4096,0.015841777125994366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,3584,0.025229333175553217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,7168,0.03750133183267381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,8192,0.41370222303602433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,3584,0.029727111260096233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,7168,0.243904882007175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,5120,0.010804444551467896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,128,0.020239111449983384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,5120,0.019428445233239066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,7168,0.07548889186647204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,10240,0.038711110750834145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,512,0.17658044232262504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,12288,0.1964319944381714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,3584,0.008019555773999956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,6144,0.1092711091041565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,10240,0.09483555290434097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,2560,0.33682576815287274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,256,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,7168,0.01070577816830741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,2560,0.017883555756674874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,7168,0.1803706619474623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,3584,0.008360889222886827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,16384,0.06986133257548015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,6144,0.018806222412321303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,12288,0.13214488824208578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,3072,0.42855111757914227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,768,0.016703999704784818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,2560,0.011877333124478659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,16384,0.03074044320318434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,16384,0.17388178242577446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,256,0.01957066688272688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,768,0.027067555321587458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,7168,0.1958320008383857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,512,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,128,0.0021928888228204516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,2048,0.050208889775806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,2048,0.012647999657524956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,64,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,10240,0.01586666703224182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,5120,0.047016001409954496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,2560,0.03346666693687439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,16384,0.07746666669845581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,2048,0.00423911131090588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,6144,0.04107022285461426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,6144,0.07129422161314222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,256,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,7168,0.37497689988878036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,512,0.032935112714767456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,4096,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,51200,0.17144266764322916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,65536,0.08850666549470688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,3584,0.009773333039548662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,1024,0.011176889141400656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,3072,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,1024,0.014408000641398959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,10240,0.027832001447677612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,8192,0.021935999393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,2560,0.03047200043996175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,5120,0.033144887950685285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,768,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,5120,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,10240,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,6144,0.03294310967127482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,51200,0.18649510542551676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,12288,0.04019822345839606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,12288,0.5174560017055935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,6144,0.028042667441897925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,3584,0.007817777494589487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,65536,2.1134666866726346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,8192,0.03940088881386651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,32,0.017606221967273288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,12288,0.2606968879699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,7168,0.5234808921813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,256,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,5120,0.01145244472556644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,512,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,256,0.05001422100596958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,768,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,32,0.01958577831586202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,10240,0.04582577612664965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,8192,0.33114221360948354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,12288,0.11042933331595527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,3072,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,3072,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,128,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,2048,0.04417511158519321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,32,0.005684444473849402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,12288,0.05422666668891907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,64,0.011928888658682505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,32,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,65536,4.331957499186198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,3584,0.44083110491434735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,65536,0.07443644603093465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,3584,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,1536,0.013449778159459433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,7168,0.021941334009170532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,32,0.013389333254761167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,1024,0.018906666172875296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,1536,0.23508000373840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,1024,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,10240,0.06914844115575154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,16384,0.03369599911901686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,1536,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,256,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,64,0.002058666613366869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,10240,0.07476533121532865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,8192,0.15811999638875326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,512,0.015663999650213454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,256,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,3072,0.02349244389269087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,8192,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,512,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,5120,0.07032622231377496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,1536,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,51200,0.6159306632147895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,12288,0.06706755691104464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,5120,0.08680799934599136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,768,0.011281777587201862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,12288,0.04428799947102865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,128,0.015234667393896313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,64,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,1536,0.010552888943089379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,768,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,1536,0.012478222449620565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,8192,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,2048,0.08782844411002265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,10240,0.042020443412992686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,5120,0.012885332935386233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,10240,0.014318222800890604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,16384,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,4096,0.04005422194798788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,768,0.03340266810523139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,10240,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,65536,0.06884621911578707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,2560,0.28596356179979115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,12288,0.06793510913848877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,10240,0.014533332652515836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,7168,0.011885333392355176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,51200,3.1785973442925344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,8192,0.03388799892531501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,3072,0.012235555383894177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,51200,0.03850488861401876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,64,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,1536,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,4096,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,3584,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,512,0.05170222123463949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,65536,0.05833066834343804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,2560,0.02407288882467482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,6144,0.02671555512481266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,6144,0.009036444127559662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,7168,0.04292533463901944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,6144,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,1536,0.02171733313136631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,4096,0.014341332846217685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,2560,0.16800444655948213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,768,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,3072,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,3584,0.019854222734769184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,2560,0.03717422154214647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,51200,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,3584,0.012835555606418185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,7168,0.012472000386979846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,10240,0.13655822806888157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,2560,0.01614755557643043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,6144,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,2048,0.036976890431510076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,4096,0.020440888073709276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,3584,0.060356444782680936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,10240,0.025524444050259058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,4096,0.008011555506123437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,1536,0.01901244454913669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,51200,1.0314035415649414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,2560,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,7168,0.016619554824299283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,32,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,5120,0.0191600008143319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,3584,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,3584,0.012259555359681448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,3072,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,2560,0.004918222212129169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,1536,0.028584887584050495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,65536,0.31843021180894643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,6144,0.04426133301522997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,7168,0.0612693362765842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,51200,0.11360266473558213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,6144,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,3584,0.016376000311639573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,16384,0.04254844453599718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,12288,0.016893333858913846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,3584,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,1024,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,1536,0.038423111041386925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,1536,0.026391999589072332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,4096,0.025118221839269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,2048,0.013047999805874295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,128,0.015103111664454142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,7168,0.1229119963116116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,2048,0.03185688787036472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,8192,0.027427555786238775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,1536,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,3072,0.016193777322769165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,32,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,2048,0.009233777721722921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,8192,0.06818311081992255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,6144,0.04136088821623061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,1024,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,4096,0.03414311011632284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,3072,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,3072,0.09768533044391209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,4096,0.14992533789740667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,4096,0.010238221950001186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,32,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,64,0.0018817778262827133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,768,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,12288,0.34438133239746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,512,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,1024,0.0041786668201287585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,4096,0.07097778055402967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,1024,0.02196888956758711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,5120,0.026398221651713055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,51200,0.1417271163728502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,768,0.03012088934580485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,4096,0.025067556235525344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,2560,0.0724817779329088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,768,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,10240,0.025645333859655593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,12288,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,1536,0.04794222116470337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,16384,0.05864355299207899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,16384,0.019370666808552213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,6144,0.11083466476864284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,1536,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,2048,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,32,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,7168,0.01202488856183158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,512,0.004650666481918759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,5120,0.03831377625465393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,1024,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,1536,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,768,0.008753777378135258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,512,0.008346666892369589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,10240,0.04839200112554762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,12288,0.021359999974568684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,65536,0.2815937731001112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,256,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,65536,0.054056889481014676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,1536,0.07866488562689887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,512,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,512,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,256,0.006474666711356904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,10240,0.012686221963829465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,1024,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,16384,0.03360355562633938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,65536,0.12131378385755752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,128,0.012665777570671506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,8192,0.01277511070171992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,5120,0.009635555247465769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,3072,0.019511999355422128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,51200,0.11553066306644016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,256,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,8192,0.1202275620566474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,256,0.015528000063366361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,2048,0.015098666151364645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,5120,0.018236445056067575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,65536,0.10415733522839016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,51200,0.34150489171346027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,768,0.007037333316273159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,10240,0.13208622402615017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,2560,0.1595591174231635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,12288,0.25957687695821124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,65536,0.24833689795600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,65536,0.7194177839491102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,256,0.0038622220357259116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,3072,0.056568887498643666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,6144,0.04149511125352647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,6144,0.013179555535316467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,5120,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,4096,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,65536,0.07362755801942614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,65536,0.3490426540374756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,6144,0.23368088404337564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,256,0.1272559960683187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,5120,0.013092444174819522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,12288,0.03323910964859857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,512,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,2560,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,65536,0.06147911151250204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,1024,0.01742577718363868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,768,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,512,0.017248888810475666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,256,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,6144,0.09186133411195542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,10240,0.1093475553724501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,12288,0.013408888545301227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,12288,0.014064000712500678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,256,0.02598222262329525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,10240,0.03770133190684848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,10240,0.04433600107828776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,16384,0.05774844355053372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,64,0.005991111199061076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,256,0.0026346666531430352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,7168,0.0999564462237888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,512,0.006999111009968652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,65536,0.07189777824613783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,2560,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,3584,0.012296888563368054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,16384,0.05397689011361864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,1024,0.03705599904060364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,1024,0.04312800036536323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,2560,0.008057778080304464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,512,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,256,0.013726222018400827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,6144,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,4096,0.026328000757429335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,2048,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,7168,0.02296000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,6144,0.024031110935741003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,7168,0.16382488939497206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,64,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,4096,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,51200,0.20438933372497559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,1536,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,512,0.03784711162249247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,2560,0.029426667425367568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,5120,0.06015111340416802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,4096,1.5930070877075195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,2560,0.016419554750124615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,8192,0.025690666503376428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,12288,0.5150897767808702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,1024,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,512,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,2048,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,65536,0.059761775864495166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,2560,0.1080933345688714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,12288,0.028544889556037054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,5120,0.07310489151212904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,2560,0.008984888593355814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,16384,0.06788355774349637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,10240,0.05187288920084635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,3584,0.4709679815504286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,3072,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,3072,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,128,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,512,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,4096,0.10726844602160984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,2048,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,8192,0.06432444519466825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,7168,0.027892443868849013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,4096,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,128,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,8192,0.026724444495307073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,128,0.018206222189797294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,16384,0.6250853538513184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,65536,0.03394844465785556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,64,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,1536,0.01993688941001892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,2048,0.03598399957021078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,10240,0.08970933490329319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,1536,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,1536,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,512,0.007531555162535773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,2048,0.007229333122571309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,256,0.025256888733969793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,10240,0.04820266697141859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,8192,0.04003999961747064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,16384,0.13035999404059515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,768,0.007350222104125553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,2560,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,6144,0.012457778056462606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,6144,0.05757066938612196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,10240,0.06324266725116305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,6144,0.14781778388553196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,5120,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,3584,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,6144,0.6845928827921549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,2048,0.02793866727087233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,64,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,128,0.012251555919647217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,5120,0.021526222427686054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,2560,0.02347466680738661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,2048,0.07691644297705756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,65536,5.615851508246528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,12288,0.02718577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,10240,0.033955554167429604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,3584,0.014892444014549255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,6144,0.01893422173129188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,32,0.0034471110751231513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,2560,0.007816889219813878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,2560,0.004604444321658877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,4096,0.009486222432719337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,768,0.0026008888251251644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,16384,0.03208533260557387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,128,0.017142222987280954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,1024,0.0038151111867692736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,7168,0.04355822337998284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,3584,0.07205155822965834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,512,0.002970666728085942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,512,0.012776888906955719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,512,0.01370666672786077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,256,0.020789333515697055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,512,0.0053022222386466126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,2048,0.02901244494650099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,8192,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,3584,0.013815999858909182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,768,0.04878933231035868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,7168,0.22066756089528403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,128,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,65536,12.62749989827474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,51200,3.861914740668403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,65536,1.0898035897148979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,1024,0.018592000007629395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,10240,0.2010791169272529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,5120,0.01294133315483729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,64,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,256,0.0027742222365405825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,51200,0.1039573351542155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,3584,0.022534221410751343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,1536,0.005970666805903117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,3584,0.018927999668651156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,1024,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,768,0.003165333428316646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,12288,0.08496533499823676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,5120,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,8192,0.026668444275856018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,32,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,768,0.01741155485312144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,512,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,3584,0.007838221887747446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,3072,0.03605155481232537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,512,0.0030453333424197305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,2560,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,256,0.009723555710580613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,5120,0.03106844425201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,7168,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,5120,0.013172444370057849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,5120,0.09318844477335612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,5120,0.029442667961120605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,8192,0.08015111419889662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,5120,0.012404444317022959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,65536,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,1536,0.008447111480765873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,10240,0.025389333566029865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,4096,0.008496888809733922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,5120,0.030738668309317693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,5120,0.03837600019243028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,3584,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,2048,0.014640889234013028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,512,0.007875555919276344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,8192,0.07772355609469943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,1024,0.022680888573328655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,512,0.021192888418833416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,1024,0.016400888562202454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,128,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,7168,0.05116800136036343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,8192,0.07026044527689616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,51200,0.06677511003282335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,8192,0.03665511144532098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,512,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,2560,0.01316088851955202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,128,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,1024,0.175254225730896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,2048,0.013225778109497495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,4096,0.02570488883389367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,51200,9.122422960069445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,6144,0.010567111273606619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,5120,0.06595910920037164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,128,0.03783999880154928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,6144,0.03479555580351088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,8192,0.03616711166169908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,5120,0.0203146669599745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,256,0.019582221905390423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,256,0.005672889037264719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,5120,0.018584888842370775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,512,0.029167999823888142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,256,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,6144,0.03202755583657159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,4096,0.041026665104760066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,51200,0.039656887451807656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,7168,0.013577777478430005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,256,0.04717955655521817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,3584,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,64,0.019142222073343065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,10240,0.015120888749758402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,3584,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,2560,0.0075093333919843035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,3584,0.01389422184891171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,2560,0.004936000125275718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,6144,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,1536,0.01015644437736935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,2560,0.03808355662557814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,2048,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,6144,0.049801776806513466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,512,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,16384,0.2011244429482354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,12288,0.028020444843504164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,768,0.006217777729034424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,3072,0.013146667016877068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,65536,1.6347227096557617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,6144,0.008722667064931657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,1536,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,64,0.03564000129699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,65536,0.06965955760743883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,3584,0.018820444742838543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,3072,0.017684444785118103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,128,0.022965333527988855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,5120,0.00889511075284746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,64,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,4096,0.05422755413585239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,4096,0.12077155378129746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,5120,0.014715555641386243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,7168,0.0810319979985555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,32,0.008765333228641087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,5120,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,512,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,10240,0.02792888879776001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,12288,0.01920711166328854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,768,0.026201777988009985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,4096,0.008046222229798635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,4096,0.1158871120876736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,2048,0.02221777704026964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,5120,0.7573368814256457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,8192,0.021159110797776118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,10240,0.032235556178622775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,2048,0.09830577505959405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,1536,0.12129688262939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,65536,0.1307377815246582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,3072,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,10240,0.0657822224828932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,256,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,7168,0.1323653327094184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,51200,0.06397600306404962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,768,0.0958062211672465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,51200,0.34180710050794816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,1024,0.0807173318333096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,256,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,256,0.02053777707947625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,2560,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,128,0.02546399997340308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,6144,0.08829955259958903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,6144,0.06957688596513537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,4096,0.22605954276190865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,10240,0.09882666667302449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,512,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,7168,0.022287999590237934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,768,0.009098666409651438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,2560,0.20857244067721895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,5120,0.025451555848121643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,32,0.02089333368672265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,1024,0.00850222259759903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,12288,0.023008000519540574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,2560,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,12288,0.10900266965230306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,8192,0.09938133425182766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,2560,0.007415999968846639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,4096,0.01789422167672051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,256,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,128,0.02404622236887614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,16384,0.11030311054653591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,1024,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,7168,0.03304533494843377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,768,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,5120,0.01070755554570092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,16384,0.0887155532836914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,12288,0.05358933409055074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,4096,0.027520888381534155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,3072,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,512,0.05583288934495714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,64,0.0017991111510329777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,10240,0.01145511120557785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,512,0.04270844327078926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,4096,0.021923555268181696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,3584,0.030459556314680312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,256,0.017553778158293832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,768,0.012889777620633444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,51200,0.05511200096872118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,64,0.06817066669464111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,7168,0.044890665345721774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,1024,0.04537599947717455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,256,0.011369778050316704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,12288,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,8192,0.05677244398328993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,7168,0.09530755546357896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,2048,0.014297778407732645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,12288,3.576735178629557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,51200,0.016199999385409884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,256,0.039346665143966675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,3072,0.010991999672518836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,2048,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,1024,0.27245865927802193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,51200,0.052192889981799655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,768,0.04556355542606778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,1536,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,12288,0.03168711066246033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,768,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,256,0.03985244366857741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,256,0.015295111470752291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,16384,0.0555022226439582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,512,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,128,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,2048,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,2560,0.012610666453838348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,12288,0.09348089165157741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,512,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,8192,0.0749173363049825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,12288,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,2048,0.014840000205569796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,1536,0.026084444589085046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,7168,0.013740444348918067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,2560,0.015629332926538255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,256,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,10240,0.03145600027508206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,512,0.016555555992656283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,7168,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,2048,0.05757421917385525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,4096,0.028802666399214003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,768,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,6144,0.05776355663935343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,1024,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,65536,0.06310400035646227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,12288,0.025399999486075506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,32,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,2048,0.007820444802443186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,1536,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,7168,0.03085422184732225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,1536,0.06804088751475017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,10240,0.5510604646470811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,12288,0.06652000215318468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,1024,0.031678222947650485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,10240,0.011880888707107969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,7168,0.09894577662150066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,128,0.00810222244924969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,1536,0.007897777689827813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,1536,0.004580444345871608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,16384,0.02715555495686001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,7168,0.0363377763165368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,128,0.02331999937693278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,6144,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,2560,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,10240,0.016600888636377122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,512,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,5120,0.005312888986534542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,7168,0.07367110914654203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,7168,0.01956355571746826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,3072,0.06974311007393731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,16384,0.18686489264170328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,16384,0.07146666447321574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,1024,0.023578666978412207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,4096,0.00870488915178511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,12288,9.8726806640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,7168,0.010643555886215635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,16384,0.2939128875732422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,1536,0.003972444269392225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,65536,0.6649244626363119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,3072,0.021750221649805706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,256,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,2048,0.010993777877754636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,4096,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,128,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,256,0.02825688984658983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,2048,0.027928001350826685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,512,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,51200,0.040031999349594116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,3072,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,768,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,2048,0.12599555651346842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,3584,0.05824177794986301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,65536,0.14383110735151503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,512,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,16384,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,768,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,4096,0.008198222352398766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,65536,0.21333777904510498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,768,0.016415999995337594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,3072,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,51200,0.3090035650465223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,5120,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,512,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,51200,0.5740888913472494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,16384,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,2560,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,128,0.010280888941552904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,768,0.007749333149856991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,12288,0.11400532722473145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,1024,0.010129777921570672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,3584,0.008672888908121321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,7168,0.009746666583749983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,12288,0.027628444963031348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,256,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,10240,0.3022097746531169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,512,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,1536,0.03805689016977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,768,0.02202488813135359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,512,0.07006400161319308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,16384,0.06707288821538289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,512,0.011769777370823754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,8192,0.12899733914269343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,4096,0.0486968888176812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,6144,0.22874132792154947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,3072,0.007337777978844113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,51200,0.6225146717495388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,16384,0.18587377336290148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,8192,0.16777777671813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,4096,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,128,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,12288,0.03337511089113023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,1536,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,2560,0.012473777764373355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,4096,0.025166221790843542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,1024,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,2560,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,3072,0.010935111178292168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,512,0.010463111102581024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,1536,0.029659556017981634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,768,0.014134221606784396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,3072,0.01071466671095954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,2048,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,1536,0.055439111259248525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,32,0.06918222374386258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,768,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,768,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,8192,0.02016533414522807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,3072,0.015228443675571017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,768,0.045591112640168935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,3584,0.046720888879564076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,16384,0.017904000149832833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,8192,0.04033777780003018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,512,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,3072,0.03693777653906081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,3584,0.017495999733606975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,128,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,12288,0.17726400163438585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,4096,0.08185333013534546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,128,0.020492444435755413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,10240,0.02588711182276408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,7168,0.05346666773160299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,6144,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,65536,0.16653154955969915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,8192,0.04542666673660278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,8192,0.0676764448483785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,64,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,6144,0.023669333921538457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,10240,0.09299200110965306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,5120,0.03489066825972663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,3072,0.011980444192886353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,5120,0.06968800226847331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,1024,0.020021332634819877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,5120,0.009845332966910468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,3072,0.01087288889620039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,3584,0.043599112166298755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,3584,0.028952889972262915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,7168,0.0352497763103909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,51200,0.12085955672793919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,51200,0.7375769085354276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,256,0.007908444437715743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,4096,0.06588177548514472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,2048,0.05528266562355889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,2048,0.014058666096793281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,2048,0.025573333104451496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,7168,0.01459377838505639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,10240,0.04669955703947279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,256,0.0029493332323100832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,2048,0.06198666493097941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,7168,0.03467466764979892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,128,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,2560,0.058323555522494845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,128,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,2048,0.019317333896954853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,12288,0.10983466439776951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,1536,0.04362577862209744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,1536,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,51200,0.0517173343234592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,5120,0.04558666547139486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,2560,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,65536,0.20161955886416963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,4096,0.026399999856948853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,512,0.11385599772135417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,2048,0.033702221181657575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,7168,0.03906755646069845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,2560,0.02252888845072852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,128,0.0018151110659042995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,8192,0.226036442650689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,256,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,7168,0.014225777652528552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,65536,0.0944711102379693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,16384,0.1813786692089505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,8192,0.004930666751331753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,3072,0.022738666998015508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,1536,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,3584,0.022640888889630634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,51200,0.05384088887108696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,4096,0.027012444204754297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,64,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,65536,0.06900799936718412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,16384,0.02090666691462199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,7168,0.016424889365832012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,5120,15.873799641927084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,2048,0.07731822464201185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,2048,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,32,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,512,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,2048,0.024780443973011438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,16384,0.019890666007995605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,3584,0.04400533437728882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,64,0.0034968890249729156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,64,0.004855111241340637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,5120,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,1536,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,65536,19.963172064887154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,5120,0.05896266963746813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,768,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,1536,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,6144,0.1412702269024319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,51200,1.0927439795600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,10240,0.031939556201299034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,3072,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,128,0.011711111499203576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,2048,0.01198933356338077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,16384,0.016467556357383728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,10240,0.04212355613708496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,128,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,65536,0.07218133078681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,256,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,512,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,51200,0.0403386652469635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,3072,0.02528444429238637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,128,0.0032222221295038858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,5120,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,7168,0.02256711158487532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,64,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,10240,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,2560,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,256,0.04042488998836941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,1536,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,128,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,4096,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,8192,0.020605333977275424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,3072,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,512,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,4096,0.012034667034943899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,4096,0.3133804533216688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,128,0.021352888809310064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,768,0.015041778484980265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,1536,0.02141600019401974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,3584,0.12759999434153238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,768,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,3072,0.09736355808046128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,128,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,2048,0.005482666608360078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,768,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,4096,0.031038221385743883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,65536,0.076127999358707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,256,0.014900444282425774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,65536,0.06418222188949585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,64,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,6144,0.04900977677769131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,512,0.1589475605222914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,2048,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,2560,0.05100355545679728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,5120,0.019718221492237516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,4096,0.1482764482498169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,2560,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,512,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,1536,0.0690222250090705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,2560,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,256,0.016781333420011733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,768,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,7168,0.14317688677046034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,51200,0.7762702306111654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,32,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,2048,0.02164266672399309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,768,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,65536,0.138157327969869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,32,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,3584,0.7710275650024414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,6144,0.2208026647567749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,12288,0.029558221499125164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,7168,0.05039911137686836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,256,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,8192,0.04089422358406915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,128,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,2560,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,2560,0.0074826669361856245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,12288,0.02091466718249851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,32,0.006796444455782573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,2560,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,8192,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,64,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,8192,0.010654222634103564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,16384,0.033887111478381686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,8192,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,1536,0.023608888188997906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,5120,0.027300443914201524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,10240,0.05865866608089871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,4096,0.05543555484877693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,65536,0.09197066889868842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,3584,0.016050666570663452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,6144,0.07572888665729098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,1536,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,8192,0.02684266699684991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,3072,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,16384,0.07823910978105333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,256,0.147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,65536,0.14791022406684026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,2560,0.03288355469703674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,8192,0.1481048928366767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,3072,0.043531556924184166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,2048,0.020222221811612446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,768,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,1024,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,12288,0.13621421655019125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,7168,0.0738088885943095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,3072,0.0038311113086011675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,65536,0.0585048860973782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,2560,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,128,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,32,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,4096,0.04711822337574429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,1024,0.007786666353543599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,64,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,65536,0.17513867219289145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,65536,0.054344889190461904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,1536,0.006490666833188798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,12288,0.0395413339138031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,256,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,10240,0.19168355729844835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,10240,1.0553919474283855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,7168,0.03836089041497972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,65536,0.12654933664533827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,65536,0.0612693362765842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,65536,0.2599119875166151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,16384,0.7337573369344076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,64,0.00566133318675889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,1536,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,64,0.10164711210462783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,4096,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,3072,0.01551822159025404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,32,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,4096,0.05039733317163256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,10240,0.1316559977001614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,7168,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,2560,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,2560,0.026404443714353774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,12288,0.13666400644514295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,8192,0.016565332810084026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,128,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,2560,0.01236355553070704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,1536,0.03611110978656345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,512,0.01920000049802992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,16384,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,16384,0.2845768928527832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,7168,0.012250666817029318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,10240,0.05582844548755222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,10240,0.27833154466417104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,3072,0.007583111524581909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,256,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,8192,0.02023555503951179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,7168,0.01386044422785441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,4096,0.06362577941682604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,7168,0.03336800138155619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,2560,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,12288,0.029232889413833618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,10240,0.0980222225189209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,1536,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,7168,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,16384,0.13373243808746338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,3072,3.023993810017904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,16384,0.04454044500986735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,3584,0.08120089107089572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,2048,0.01923111081123352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,768,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,7168,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,5120,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,32,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,1024,0.017507554756270517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,4096,0.30829066700405544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,1024,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,3584,0.008052444292439355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,3072,0.04495733314090305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,3072,0.012152888708644442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,16384,0.0433991120921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,7168,0.02092711130777995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,256,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,51200,31.813337537977432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,3584,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,64,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,1536,0.019426667028003268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,1536,0.018459555175569322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,3584,0.012402666939629449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,2560,0.015196444259749519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,3584,0.40505867534213597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,65536,0.02755733331044515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,51200,0.05065155691570706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,768,0.017490666773584154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,1024,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,16384,0.03233688738611009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,1024,0.01829155617290073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,3072,0.012678222523795234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,6144,0.0685102211104499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,2560,0.015063999427689446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,64,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,3584,0.040377779139412776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,1024,0.01961688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,12288,0.966742197672526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,8192,0.0842462248272366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,1024,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,8192,0.32079821162753636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,512,0.003568888952334722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,51200,0.057447108957502574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,3584,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,6144,0.02033866610791948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,16384,0.21702933311462402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,1536,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,65536,0.060269335905710854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,1024,0.02554133368863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,6144,0.012248888611793518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,512,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,65536,0.2568559911515978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,32,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,128,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,5120,0.8564764658610026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,51200,0.054056889481014676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,256,0.002983111060327954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,3072,0.03212622139188979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,256,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,12288,0.13989867104424372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,4096,0.014666666587193808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,8192,0.07249866591559516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,10240,0.02329155637158288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,512,0.020872889293564692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,768,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,12288,0.07854933208889432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,1536,0.08697599834865993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,2560,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,10240,0.11377778318193223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,6144,0.08619644244511922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,256,0.08176889022191365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,128,0.014831110835075378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,10240,0.02915999955601162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,12288,0.03628177775277032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,768,0.033378667301601835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,32,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,5120,0.014095111025704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,32,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,3584,0.01869244376818339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,32,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,768,0.018018666240904067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,7168,0.14250400331285265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,768,0.13954933484395346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,16384,0.06625688738293119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,128,0.015184000134468079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,256,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,32,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,64,0.0017946666727463405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,128,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,12288,0.018567999203999836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,2048,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,10240,0.14089066452450222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,3584,0.047057777643203735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,16384,0.6691884464687772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,65536,0.2506133450402154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,16384,0.02027466727627648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,16384,0.06116977665159437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,256,0.011953777737087674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,2560,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,5120,0.009090666969617208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,51200,0.8582533200581869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,256,0.010271111296282874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,3584,0.07250489128960504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,6144,0.014017777310477363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,16384,0.3408835464053684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,2560,0.026702221896913316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,5120,0.11443911658393012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,256,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,256,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,51200,0.10979644457499187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,128,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,1536,0.006196444647179709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,4096,0.12832178009880915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,7168,0.012854222622182635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,51200,0.049610667758517794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,6144,0.02810844447877672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,5120,0.02686400049262577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,51200,0.061669336424933545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,8192,0.4059182273017035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,10240,0.046332445409562856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,10240,0.12347200181749131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,3072,1.1852151023017037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,6144,0.027075555589463975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,3072,0.06435644626617432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,768,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,3072,0.008612444831265343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,5120,0.00980799976322386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,256,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,8192,0.01828622155719333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,2560,0.016150222884284127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,12288,0.07733955648210314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,256,0.011561777856614856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,1024,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,65536,0.06608800093332927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,6144,0.0976675550142924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,768,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,8192,0.03769333495034112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,64,0.007517333659860823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,4096,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,1024,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,32,0.015104000767072042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,256,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,10240,0.027808000644048054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,512,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,128,0.05700622002283732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,768,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,12288,0.02279288901223077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,1024,0.05915288792716133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,512,0.015015999476114908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,16384,0.030258665482203167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,5120,0.02347022294998169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,6144,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,65536,0.44434489144219297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,7168,0.004993777722120285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,3584,0.018718222777048748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,4096,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,5120,0.021651556094487507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,1024,0.008073777788215214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,2048,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,2560,0.1377777788374159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,10240,0.030231111579471167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,4096,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,1024,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,4096,0.020594666401545208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,512,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,1024,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,32,0.014019555515713163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,256,0.006072000082996156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,10240,2.2584027184380426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,65536,0.029940442906485662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,128,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,5120,0.015771556231710646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,6144,0.015981333123313058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,4096,0.02273955610063341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,5120,0.010637332995732626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,256,0.04098844528198242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,51200,1.1436701880560982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,128,0.0090844440791342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,32,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,10240,0.010475555227862464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,1536,0.0038764443662431505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,768,0.002691555561290847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,1536,0.006803555621041193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,12288,0.6828906801011828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,12288,0.05435555511050754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,128,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,3072,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,2048,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,3584,0.030065778228971694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,768,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,768,0.09339200125800239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,2560,0.024635555015669927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,8192,0.13125156031714544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,3072,0.03770133190684848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,1024,0.02196088929971059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,2560,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,6144,0.15922133127848306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,7168,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,7168,0.012304888831244575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,6144,0.030299554268519085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,32,0.013036444783210754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,6144,0.0051893335249688895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,512,0.020118221640586853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,256,0.026939556002616882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,8192,0.013388444152143268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,2048,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,3584,0.06230311261283027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,512,0.002581333327624533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,1536,0.023535999986860488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,512,0.11514311366611057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,1536,0.010528888967302112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,64,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,5120,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,4096,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,64,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,4096,0.02756888833310869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,16384,0.39328445328606504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,6144,0.02404266595840454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,3072,0.013559111290507846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,8192,0.013242666920026144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,512,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,128,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,1536,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,7168,0.04961155520545112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,6144,0.03565511107444763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,1024,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,128,0.03669688767857022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,16384,0.07169066535101996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,5120,0.18551022476620146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,2560,0.010622222390439777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,256,0.016514667206340365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,7168,0.2649742232428657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,3072,0.07664177815119426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,1024,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,2560,0.03289066751797994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,128,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,16384,0.015461333923869662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,10240,0.03576888971858554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,16384,1.6148160298665364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,1536,0.012800888882742988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,5120,0.12541955047183567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,64,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,65536,0.09836000204086304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,6144,0.3966648843553331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,16384,0.04693866769472758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,512,0.058428446451822914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,128,0.06616889105902778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,512,0.02825777729352315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,512,0.021846221552954778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,768,0.00998933364947637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,128,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,256,0.09731022516886394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,3072,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,6144,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,10240,0.030368000268936157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,128,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,4096,0.034761776526769005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,128,0.050606220960617065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,10240,0.10127200020684136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,5120,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,768,0.010741333166758219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,64,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,51200,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,5120,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,256,0.013208889298968844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,1024,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,1536,0.016982222596804302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,32,0.024375110864639282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,1024,0.01608088943693373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,65536,0.14816444449954563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,256,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,1536,0.008813333180215623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,1024,0.006341333190600078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,2048,0.00656444455186526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,51200,0.0714577767584059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,64,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,3072,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,1536,0.006325333482689328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,6144,0.037316444847318865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,1536,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,10240,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,5120,0.03070844544304742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,1536,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,256,0.004312889029582341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,12288,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,512,0.0034835553831524322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,256,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,10240,0.0173688895172543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,6144,0.021753778060277302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,12288,0.016696888539526198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,3584,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,64,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,128,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,512,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,16384,0.05296711126963297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,12288,0.013696889082590738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,1536,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,3072,0.020851555797788832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,2048,0.0074826669361856245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,7168,0.1221911112467448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,16384,0.06575377782185872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,2560,0.032277332411872015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,3584,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,2048,0.007230222225189209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,65536,0.29942043622334796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,16384,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,6144,0.01792711185084449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,128,0.026574222577942744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,7168,1.3305590947469075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,65536,0.023940443992614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,512,0.036489778094821505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,4096,0.007726222276687622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,16384,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,32,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,12288,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,2560,0.07179822524388631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,16384,0.4590657552083333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,768,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,128,0.013345777988433838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,4096,0.02825688984658983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,4096,0.27618667814466685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,3584,0.016599999533759225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,1536,0.016691555579503376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,6144,0.3116817739274767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,4096,0.05078311098946465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,65536,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,1024,0.005671111245950063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,5120,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,3072,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,3072,0.07779110802544488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,16384,0.04421777857674492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,256,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,5120,0.13924088742997912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,768,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,3072,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,4096,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,16384,0.022817777262793645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,256,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,65536,0.38029156790839297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,6144,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,4096,0.11243732770284016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,16384,0.05485066771507263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,51200,0.3162879943847656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,4096,1.0027155346340604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,128,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,768,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,10240,0.012304888831244575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,4096,0.010312889185216691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,64,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,1024,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,4096,0.008488888541857401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,3072,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,4096,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,7168,0.04626133375697666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,1024,0.013153778182135688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,7168,0.06851910882525973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,3072,0.0232595553000768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,32,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,3072,0.3233288923899333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,32,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,2048,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,1024,0.011544889046086205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,12288,0.021355556117163763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,2560,0.19026311238606772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,6144,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,6144,0.5395661989847819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,1024,0.02534666657447815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,5120,0.08591288990444607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,4096,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,7168,0.01885511146651374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,51200,1.617613368564182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,7168,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,7168,0.17433867189619276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,256,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,512,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,128,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,128,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,3072,0.02313511073589325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,3072,0.020781333247820537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,1024,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,5120,0.025288888149791296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,6144,0.28304534488254124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,2048,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,10240,0.03960977660285102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,51200,0.12126221921708848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,2048,0.06765511300828722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,256,0.016534222496880423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,32,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,1024,0.03193511234389411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,128,0.022856000396940444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,5120,0.012515555653307172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,2560,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,6144,0.05474755499098036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,1536,0.004335999902751711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,256,0.0443795555167728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,7168,0.022308443983395893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,10240,0.048787554105122886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,65536,0.9507591459486219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,10240,0.6187466515435113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,2048,0.014249778456158109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,512,0.05581421984566582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,128,0.010993777877754636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,10240,0.07064000103208753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,65536,0.8495377964443631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,4096,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,10240,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,12288,0.6801306406656901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,1536,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,2560,0.013220444321632385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,1024,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,1536,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,65536,0.12296356095208062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,2560,0.025996444953812495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,51200,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,3584,0.2047004434797499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,256,0.003909333298603694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,4096,0.015001777145597668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,65536,0.07253511084450616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,65536,0.2363342179192437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,1024,0.050379554430643715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,256,0.013559111290507846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,3072,0.05516622132725186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,12288,0.14848889244927302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,1024,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,32,0.01924977699915568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,1024,0.1802124447292752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,16384,0.7444471253289117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,12288,0.017143110434214275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,1536,0.017991110682487488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,256,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,1024,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,256,0.012644444074895648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,6144,0.02535199953450097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,10240,0.1443448861440023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,12288,0.06200622187720405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,3584,0.010997333460383944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,4096,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,768,0.00703288863102595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,16384,0.025798221429189045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,2048,0.017314665847354464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,65536,0.06269777483410306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,3072,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,2560,0.007291555404663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,51200,0.17703378200531006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,128,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,12288,4.449763403998481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,65536,0.1344755490620931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,7168,0.039990223116344877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,51200,0.1653831137551202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,10240,0.14110399617089167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,2048,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,7168,0.020669332808918424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,2048,0.07833244403203328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,1536,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,2048,0.007556444240940942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,8192,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,32,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,64,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,10240,0.01686044368478987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,512,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,7168,0.02774311105410258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,128,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,32,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,64,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,65536,0.04977866676118639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,7168,0.022273777259720698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,3584,0.014195554786258273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,8192,0.05676800012588501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,51200,0.22596089045206705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,128,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,2560,0.019077334139082167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,16384,0.20288444889916313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,6144,0.02465244465404087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,8192,0.041835556427637734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,2560,0.006211555666393704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,768,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,3072,0.013524444566832649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,128,0.027679110566775005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,65536,0.0632293356789483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,1536,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,1024,0.01829777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,4096,0.04544711112976074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,65536,1.6183848910861547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,512,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,512,0.01888533267709944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,1024,0.14155377282036677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,768,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,1536,0.12107555071512859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,256,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,2560,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,64,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,1536,0.014128888646761576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,256,0.02552000019285414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,8192,0.06275466415617201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,6144,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,6144,0.10383377472559611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,8192,0.03049422303835551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,2560,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,1536,0.06560622321234809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,3072,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,3072,0.016513778103722464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,64,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,8192,0.02074399921629164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,2560,0.010088000032636855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,65536,0.0630319979455736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,3584,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,4096,0.05561511384116279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,512,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,7168,0.028648889727062646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,4096,0.012429333395428128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,64,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,7168,0.05059733324580722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,128,0.03205600049760606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,512,0.0025964443468385274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,6144,0.09669955571492513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,32,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,2048,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,16384,0.042747557163238525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,10240,0.5372737778557671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,2560,0.2508942286173503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,512,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,2048,0.007000889215204451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,3072,0.03194400005870395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,2560,0.014656000667148165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,4096,0.018587556150224473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,5120,0.05212355653444926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,65536,0.06503911150826348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,12288,0.03449066811137729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,2560,0.006017777654859755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,3072,0.008447111480765873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,768,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,128,0.014849777022997538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,2560,0.054773333999845714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,51200,0.21935999393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,3584,0.05814844369888306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,16384,0.04064711266093784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,768,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,256,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,4096,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,65536,0.13136800130208334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,8192,0.015628443823920358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,7168,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,16384,0.0664106673664517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,4096,0.02603911028967963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,512,0.011773332953453064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,8192,0.1173022190729777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,1536,0.006113777971929974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,4096,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,10240,0.03224800030390421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,2560,0.016893333858913846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,1024,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,512,0.04679377873738607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,768,0.29345867368910045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,8192,0.027767111857732136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,16384,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,3584,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,4096,0.008933333059151968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,512,0.010841777755154504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,3072,0.3014106750488281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,16384,0.08629866441090901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,65536,0.09697066413031684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,51200,0.969936900668674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,2560,0.052813334597481616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,12288,0.15802400641971165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,2048,0.017647999856207106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,16384,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,3584,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,64,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,16384,4.829887813991971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,12288,0.028802666399214003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,1024,0.03761955433421665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,5120,0.00776977754301495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,2560,0.01144977741771274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,10240,0.087991111808353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,51200,3.271126217312283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,2048,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,2560,0.014634667171372307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,512,0.015089778436554803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,1536,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,5120,0.009478222164842818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,4096,0.03472355670399136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,768,0.01681155628628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,12288,0.43484089109632706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,5120,0.266827556822035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,1024,0.018226666582955253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,2560,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,2560,0.009072889056470659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,3584,0.15351288848453098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,10240,0.4771146774291992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,5120,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,256,0.0188417782386144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,65536,1.49163818359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,3584,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,6144,0.031594667169782854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,16384,0.22389777501424155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,5120,0.02933599882655674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,3072,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,3584,0.09021600087483723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,1024,0.015779554843902588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,3584,0.00776444458299213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,1536,0.00701333334048589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,1536,0.005728888842794631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,512,0.014440889159838358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,256,0.11444800429873997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,16384,0.015623110863897534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,128,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,16384,0.03115733464558919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,4096,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,16384,0.0920942227045695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,2048,0.004856889032655292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,256,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,10240,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,8192,0.021252445048756067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,3584,0.09568978018230861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,2048,0.015360888507631091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,1024,0.00926755534278022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,2048,0.04578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,2560,0.006501333167155583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,16384,0.2411795457204183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,51200,0.9558382034301758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,128,0.015136000182893542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,768,0.018163555198245578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,4096,0.017094223035706412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,2560,0.007336888876226213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,12288,0.10257422261767918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,51200,0.5079920026991103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,256,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,1024,0.016175111134847004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,1536,0.012694222231705984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,8192,0.10138489140404595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,4096,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,16384,0.033358222908443876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,6144,0.011039111349317761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,16384,0.02568799919552273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,512,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,128,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,65536,0.38811998897128636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,5120,0.07528799772262573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,64,0.012558221817016602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,7168,0.039685332112842135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,128,0.015388444066047668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,128,0.04186311033036974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,16384,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,3072,0.016081778539551627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,128,0.02197066611713833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,12288,0.06695466571384005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,8192,0.15813510947757298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,6144,0.11915822823842366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,7168,0.256930669148763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,256,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,5120,0.0448035564687517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,12288,0.04222755630811056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,1024,0.03404266635576884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,2560,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,51200,0.06975822316275702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,2560,0.00701422244310379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,32,0.0058248887459437055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,7168,0.012312889099121094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,5120,0.15068177382151285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,16384,0.0432968901263343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,768,0.05778400103251139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,12288,0.332007990943061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,1536,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,12288,0.04404533240530226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,3584,0.029959112405776978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,6144,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,256,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,3072,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,1536,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,8192,0.015621332658661736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,51200,0.3165626790788439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,16384,0.0302817788388994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,32,0.012481778032249875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,768,0.05727377865049574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,256,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,3072,0.015808888607554965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,2560,0.019966221517986722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,256,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,1024,0.03313333458370633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,5120,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,512,0.014206222361988492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,8192,0.055594669448004834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,64,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,51200,0.10311466455459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,1024,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,16384,0.14834845066070557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,10240,0.038262221548292376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,6144,0.016952888833151925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,6144,0.044236444764667086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,6144,0.011101333631409539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,512,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,1536,0.057229333453708224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,6144,0.026543110609054565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,12288,0.15605066882239446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,65536,0.29310043652852374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,3072,0.02724444369475047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,65536,0.03492622243033515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,768,0.0028151110228565005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,8192,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,51200,0.2079999976687961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,16384,0.14903199672698975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,51200,0.7929529084099664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,128,0.027316444449954565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,128,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,6144,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,64,0.018211555149820115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,2048,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,1536,0.03329866793420579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,512,0.04179199867778354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,65536,0.08408977588017781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,512,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,2048,0.021923555268181696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,3584,0.034252444903055825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,16384,0.04199377695719401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,12288,0.03204177816708883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,6144,0.09445155329174465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,65536,1.373879114786784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,65536,17.275609334309895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,7168,0.02237155536810557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,3072,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,6144,0.01201777739657296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,2048,0.06880444288253784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,6144,0.05518044365776909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,6144,0.015644444359673392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,1024,0.032229334115982056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,7168,0.5024888780381944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,4096,0.010031999813185798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,4096,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,32,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,2048,0.0069955554273393415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,8192,0.2330568896399604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,768,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,32,0.013824889229403602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,2048,0.011034666664070554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,2560,0.015917332635985482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,8192,0.020624000165197585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,65536,0.3301431073082818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,51200,0.12168622016906738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,1024,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,12288,0.4840675459967719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,2048,2.474623998006185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,12288,0.01646577815214793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,4096,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,768,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,12288,0.03666399916013082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,64,0.010119111173682744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,1024,0.022849778334299724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,5120,0.008838222258620793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,65536,0.1346071163813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,1024,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,5120,0.06120177772310045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,1536,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,51200,0.04934666554133097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,1536,0.007747555772463481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,10240,0.01997066703107622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,8192,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,8192,0.0159599996275372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,128,0.005208888815508948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,10240,0.025826666090223525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,8192,0.5305555661519369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,256,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,768,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,5120,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,8192,0.011014222270912595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,32,0.07072177860471937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,512,0.0563875569237603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,64,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,4096,0.02010311186313629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,4096,0.0495066675874922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,3072,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,51200,0.15194933944278294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,6144,0.018062222335073683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,65536,0.07560000154707167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,65536,0.04818400078349643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,12288,0.3499155574374729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,16384,0.030604445272021826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,8192,0.048319111267725624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,5120,0.009276444713274637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,16384,0.12020356125301784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,12288,0.1083573301633199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,2048,0.14190844694773355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,768,0.021115556359291077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,5120,0.34310221672058105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,2560,0.043856001562542386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,512,0.012395555774370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,256,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,128,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,8192,0.13890133963690862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,1536,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,16384,0.1470764478047689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,3072,0.7917875713772244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,3584,0.22321867942810059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,8192,0.05481688843833076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,768,0.024800888366169397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,128,0.03751555416319106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,8192,0.057504885726504855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,256,0.02994577752219306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,7168,0.027432888746261597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,256,0.03545866741074456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,2560,0.019454222586419847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,8192,0.07133066654205322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,1024,0.003036444385846456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,3584,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,8192,0.05956800116433037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,10240,0.0236151119073232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,64,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,7168,0.11539110872480605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,7168,0.009494221872753566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,768,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,3584,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,51200,0.05000533329115974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,768,0.004252444538805219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,16384,30.67324490017361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,128,0.01293244461218516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,65536,0.3011893431345622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,16384,0.024473778075642053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,2560,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,51200,0.04782755507363213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,2560,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,768,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,6144,0.018904889623324077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,7168,0.04206755426194933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,6144,0.012498666842778524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,7168,0.030249777767393324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,256,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,10240,0.1759271091885037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,65536,0.9229164123535156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,7168,0.07140977515114678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,65536,0.5840373569064671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,2560,0.044898665613598294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,4096,0.038607110579808555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,7168,0.05232533481385973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,512,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,7168,0.015608888533380298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,65536,0.39745955997043186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,7168,0.12223377492692734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,128,0.0019804444164037704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,1536,0.06754399670494927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,7168,0.07373778025309245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,3072,0.03673422336578369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,10240,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,65536,41.532470703125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,8192,0.051022221644719444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,16384,0.04215555720859104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,65536,0.09219377570682102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,256,0.040752000278896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,768,0.019537778364287484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,4096,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,5120,0.00674044465025266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,65536,0.062275555398729115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,1024,0.020964443683624268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,768,0.0500311123000251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,32,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,8192,0.041146665811538696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,10240,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,2560,0.1994640032450358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,8192,0.09250133567386204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,7168,0.3396737840440538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,3584,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,768,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,1536,0.10564621951844956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,7168,0.011000889043013254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,256,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,16384,0.04064355625046624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,3584,0.046683556503719754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,8192,0.15839378039042154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,3584,0.022264889544910852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,5120,0.07049066490597196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,2048,0.14574133025275335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,256,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,64,0.001783999924858411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,51200,0.19124889373779297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,512,0.006659555352396435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,51200,0.0488755570517646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,512,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,8192,0.07354933023452759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,3072,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,3072,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,4096,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,4096,0.026829333768950567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,512,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,64,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,1536,0.014112888111008538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,2048,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,1024,0.05333866675694784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,6144,0.05920977724923027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,3584,0.05765866571002536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,3584,0.022023111581802368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,16384,0.022843556271659002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,3584,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,512,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,7168,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,512,0.026319111386934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,6144,0.060863998201158315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,16384,0.022651554809676275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,128,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,65536,0.37778488794962567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,768,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,8192,0.06708621978759766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,256,0.009911999934249455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,2048,0.07232622305552165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,7168,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,512,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,32,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,1536,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,12288,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,51200,0.22153688801659477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,128,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,6144,0.05571733580695259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,10240,0.012253333297040729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,256,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,4096,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,65536,0.2822977701822917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,51200,0.038923554950290255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,64,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,128,0.00942133367061615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,4096,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,1536,0.011348444554540845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,12288,0.1412151124742296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,10240,0.020928000410397846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,768,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,2560,0.006776888751321369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,2560,0.13631733258565268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,2048,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,6144,0.03996710975964864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,1024,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,3072,0.007515555454625024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,256,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,65536,2.517480002509223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,768,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,7168,0.0087333329849773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,10240,0.04984711276160347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,2560,0.012446222205956778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,16384,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,2560,0.00489066665371259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,12288,0.028837333122889202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,12288,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,3072,0.015011555618709989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,2048,0.028208888239330713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,1536,0.0314284430609809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,4096,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,256,0.024941333466105994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,10240,0.024417777856190998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,3584,0.091575112607744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,64,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,1024,0.0063786668082078295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,8192,0.025418667329682246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,1024,0.010395555860466426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,5120,0.04417599903212654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,2560,0.046738667620552905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,768,0.06282311015658908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,7168,0.03718488746219211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,1536,0.006392000036107169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,128,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,8192,0.013229333692126803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,7168,0.01400888959566752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,51200,0.19229066371917725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,8192,0.14046843846638998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,3584,0.010334221853150262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,3072,3.0653243594699435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,51200,0.11874843968285455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,65536,0.09751644399431016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,12288,0.01850222216712104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,64,0.012050666742854647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,3584,0.010327999790509542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,256,0.01218933363755544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,512,0.003195555466744635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,8192,0.09245866537094116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,6144,0.030294222964180842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,64,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,10240,0.014932443698247274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,4096,0.26267732514275444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,1536,0.027623999449941847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,12288,0.343738661872016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,64,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,5120,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,65536,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,6144,0.01036533299419615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,3584,0.025060445070266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,65536,0.8973911073472766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,768,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,10240,0.23989865514967176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,7168,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,4096,0.06976266702016194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,12288,0.035022222333484225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,2048,0.030255112383100722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,3072,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,4096,0.0239955551094479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,768,0.05413600140147739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,65536,0.06268533070882161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,128,0.004918222212129169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,256,0.014158222410413953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,7168,0.30618132485283744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,256,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,10240,0.02160533269246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,2048,0.02718577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,2048,0.24929422802395293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,1536,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,128,0.01753155555990007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,51200,0.046217779318491616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,8192,0.1455493370691935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,12288,0.08163999848895602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,768,0.04769600099987454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,32,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,2048,0.005699555493063397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,8192,0.38456000222100156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,64,0.00812000036239624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,1024,0.019546666079097323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,65536,0.06582044230567084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,1024,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,7168,0.021916444102923077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,2048,0.02221600049071842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,4096,0.026127111580636766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,3072,0.04167022307713827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,2048,0.07702844672732882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,16384,0.041433778074052595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,5120,0.07054577933417426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,2560,0.07851822508705987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,4096,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,7168,0.05403111047214932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,768,0.03229599859979417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,2048,0.004179555508825514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,8192,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,1536,0.3988124529520671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,3072,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,1024,0.007429333196745978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,2048,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,7168,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,1024,0.24330311351352266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,256,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,5120,0.029678222205903795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,32,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,3072,0.013402666482660504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,5120,0.029294222593307495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,2048,0.028006222512986925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,65536,0.14354399840037027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,5120,0.024057777391539678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,1536,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,8192,0.037300444311565824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,256,0.01620533400111728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,768,0.055592887931399874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,1536,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,512,0.03010222315788269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,128,0.01272266689274046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,6144,0.053595556153191455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,768,0.014087110757827759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,7168,0.05124622252252367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,4096,0.00815288888083564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,12288,0.0695084465874566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,64,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,128,0.006796444455782573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,3584,0.024477332830429077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,2560,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,1536,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,16384,0.4363733397589789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,5120,0.028063111835055884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,512,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,12288,0.03219555483924018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,2560,0.016322667400042217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,16384,0.06855377886030409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,8192,0.13220000267028809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,256,0.01756977703836229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,2560,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,64,0.008263111114501953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,2560,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,65536,4.017961714002821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,64,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,2048,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,51200,0.7048320240444608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,2560,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,10240,0.8237386809455024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,768,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,2560,0.0069955554273393415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,128,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,5120,0.009944889280531142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,3584,1.2147466871473525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,2560,0.014133334159851074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,64,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,2560,0.011215111447705163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,2560,0.014268444644080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,1536,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,64,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,3584,0.02776177724202474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,5120,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,8192,0.014339556296666464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,768,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,512,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,128,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,4096,0.024900444679790076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,12288,0.028037332826190527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,1024,0.019296000401178997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,256,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,12288,0.055296887954076133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,128,0.018357333209779527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,256,0.003559999995761447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,256,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,4096,0.01668355531162686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,2560,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,7168,0.061520000298817955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,512,0.020204444726308186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,1536,0.24537600411309135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,65536,0.09668978055318196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,3072,0.015808888607554965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,7168,0.013626666532622443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,12288,0.034167998366885714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,7168,0.011934222446547614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,16384,0.05392711030112373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,5120,0.12792889277140299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,768,0.00611200018061532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,5120,0.01088444474670622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,64,0.018166222506099276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,5120,0.01791200041770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,1024,0.06229777468575371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,512,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,51200,0.5921617613898383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,10240,0.13672088252173528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,7168,0.20804799927605522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,16384,0.12009421984354655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,3584,0.01184266640080346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,8192,0.08230133189095391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,128,0.006811555474996567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,4096,0.09761777851316665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,3072,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,3584,0.017464000317785475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,256,0.012422222230169507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,1024,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,5120,0.004550222307443619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,12288,0.20880533589257133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,256,0.015589333242840238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,768,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,1536,0.010423111418883005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,256,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,1536,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,16384,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,12288,0.10302578078375922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,256,0.7086515426635742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,768,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,256,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,4096,0.025052444802390203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,6144,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,768,0.017844445175594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,2560,0.017151999804708693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,51200,0.021729777256647747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,16384,0.03621866636806064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,51200,0.09237777524524265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,10240,0.09392888678444757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,12288,0.020942222740915086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,8192,0.12529155943128797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,10240,1.6605741712782118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,65536,0.2395333449045817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,12288,0.01442399952146742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,12288,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,1536,0.01428711083200243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,768,0.09637244542439778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,12288,0.08067911201053195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,4096,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,65536,0.06758044163386027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,5120,0.04399111204677158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,64,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,1024,0.0038693332009845306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,5120,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,51200,0.08111911349826388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,10240,0.032525334093305804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,512,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,5120,0.029401779174804688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,1024,0.05330488748020596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,16384,0.809824890560574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,51200,2.6427688598632812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,51200,1.4959253735012477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,32,0.1616426706314087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,1024,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,6144,0.08174222045474581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,51200,1.7738346523708766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,1024,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,4096,0.017072889539930556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,32,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,2048,0.015360888507631091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,8192,0.012778666284349231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,6144,0.013717333475748697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,2048,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,10240,0.16903466648525664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,10240,0.37744445270962185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,128,0.006451555424266391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,3072,0.025535111625989277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,4096,0.7348408699035645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,6144,0.019664888580640156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,4096,0.02723377777470483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,7168,0.05126755436261495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,3584,0.01958311100800832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,2048,0.013215999636385175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,6144,0.021931555536058214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,7168,0.06890577740139432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,3584,0.01920977731545766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,3584,0.024125332633654278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,10240,0.02979644470744663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,2560,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,51200,0.35116177135043675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,512,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,2048,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,1024,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,12288,0.09296177493201362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,2048,0.02016266683737437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,8192,0.025379555092917547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,768,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,65536,0.1690124405754937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,256,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,3584,0.021013332737816706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,512,0.01278222186697854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,7168,0.047190222475263804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,7168,0.2922702100541857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,5120,0.05939200189378527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,16384,0.016514667206340365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,64,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,1024,0.018792000081804063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,768,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,2560,0.027609777119424608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,16384,0.02478666603565216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,6144,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,6144,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,51200,0.04969777663548788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,128,0.01924444403913286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,10240,0.08021333482530382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,4096,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,12288,0.012534221841229333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,2560,0.02641244398223029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,4096,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,4096,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,51200,0.30521776941087514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,256,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,4096,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,768,0.003597333199448056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,10240,0.2046391169230143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,2560,0.0074106670088238185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,10240,5.850543975830078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,16384,0.048635555638207324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,6144,0.0384844442208608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,16384,0.12875289387173122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,1024,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,512,0.012859555582205454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,3584,0.011379555695586734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,4096,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,6144,0.042620443635516696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,768,0.019279999865425956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,64,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,1024,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,1024,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,2048,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,2560,0.04272088739607069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,1536,0.00869511150651508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,3072,0.01183733344078064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,1024,0.015986666083335876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,1536,0.0038862224254343244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,8192,0.0696444445186191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,1536,0.01958755486541324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,512,0.0038515557017591265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,3584,0.01349688900841607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,6144,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,2048,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,1024,0.0788595543967353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,10240,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,1536,0.05658666955100166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,10240,0.04278844594955444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,16384,0.04523466693030464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,65536,0.11132800579071045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,2560,0.2606675624847412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,8192,0.0654275549782647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,3584,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,256,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,4096,0.05471377902560764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,6144,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,8192,0.4127475685543484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,5120,0.02405511173937056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,128,0.02187999917401208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,65536,0.0872924460305108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,512,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,10240,0.362590233484904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,51200,0.2599262131585015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,32,0.08320177925957574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,512,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,768,0.03305066625277201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,5120,0.01218933363755544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,4096,0.035445332527160645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,128,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,51200,0.1129244433508979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,64,0.004535111288229625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,3072,0.269399987326728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,8192,0.04909066690338982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,10240,0.01272266689274046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,16384,0.3840133349100749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,512,0.006073777874310811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,10240,0.09989333152770996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,1536,0.025793777571784124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,1536,0.011334222224023608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,3072,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,2560,0.014514666464593677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,6144,0.2679813437991672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,5120,0.024903110331959192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,3584,0.008565333154466417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,3584,0.006213333457708359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,4096,0.029558221499125164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,1536,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,3072,0.05717688798904419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,6144,0.042451557185914784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,3584,0.009740444521109263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,1536,0.01331733332739936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,3584,0.017723555366198223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,3584,0.03329955538113912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,4096,0.003906666818592283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,51200,2.554879930284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,8192,0.010809777511490716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,5120,0.014885332849290637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,1536,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,4096,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,10240,0.21372000376383463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,10240,0.01889244384235806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,2048,0.018548443913459778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,1536,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,65536,0.20270310507880318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,3584,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,1536,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,51200,0.5981129010518392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,65536,0.2816809018452962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,768,0.0057546666098965546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,12288,1.3273084428575304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,3072,0.019683554768562317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,64,0.030375109778510198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,1536,0.05353155401017931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,7168,0.02091022166940901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,256,0.012342222034931183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,3584,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,128,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,51200,0.21459021833207872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,256,0.026475555366939966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,12288,0.01568177839120229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,2560,0.02624888883696662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,768,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,65536,0.1360657744937473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,1024,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,768,0.007495111061467065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,6144,0.018816888332366943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,768,0.016827555166350473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,1024,0.007713778151406183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,16384,0.0463031099902259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,64,0.007687110867765214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,3072,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,16384,0.044622222582499184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,1024,0.028935111231274072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,32,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,5120,0.7835937605963813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,16384,0.036538667149013944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,3584,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,3584,0.10055022107230292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,128,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,1536,0.25648710462782115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,12288,0.056009776062435575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,51200,0.26965422100490993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,1536,0.035058667262395225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,3584,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,7168,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,2560,0.011848889291286469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,3584,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,128,0.08302577998903063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,32,0.0058702222175068325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,2048,0.02069244450993008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,12288,0.19550044006771514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,8192,0.021925333473417494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,768,0.04849777619043986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,512,0.01370666672786077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,3584,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,1024,0.04463466670778063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,64,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,5120,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,256,0.07603200276692708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,2048,0.06539644135369195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,16384,1.190640025668674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,8192,0.19222577412923178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,3072,0.1531120008892483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,2560,0.008725333544943068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,128,0.01761244402991401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,2560,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,51200,0.04445777667893303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,1024,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,1024,0.030268443955315485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,64,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,2048,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,1536,0.06020622120963203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,1024,0.008028444316652086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,3584,0.3446417649586995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,3584,0.08318666617075603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,4096,0.18166489071316191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,16384,0.012789333032237159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,256,0.028590222199757893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,51200,0.043192889955308705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,10240,0.2719342178768582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,1536,0.006315555423498154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,768,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,512,0.006136000156402588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,3584,0.015843555331230164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,256,0.033759110503726535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,3584,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,1024,0.01954844428433312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,2560,0.05436799923578898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,256,0.09363555245929295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,3584,0.013732444081041547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,3072,0.009170666337013245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,768,0.10573066605461968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,2048,0.010421333213647207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,51200,0.6182586881849501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,1536,0.025044444534513686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,12288,0.07766666677263048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,5120,0.18576533264584014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,3584,0.04632711079385546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,5120,0.00869066682126787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,10240,0.03263466556866964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,51200,0.008610666626029545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,10240,0.12832889291975233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,3072,0.05520800087187025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,2048,0.011863999896579318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,3072,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,1536,0.01575911045074463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,4096,0.014336888988812765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,512,0.032496889432271324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,65536,0.09003555774688721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,512,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,1024,0.06186844242943657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,7168,0.21902755896250406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,2048,0.03845244314935472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,4096,0.019334221879641216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,12288,0.1764284504784478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,8192,0.0772213339805603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,5120,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,512,0.16951999399397108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,512,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,768,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,65536,0.01572711103492313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,2560,0.004521777646409141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,4096,0.020908445119857788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,5120,0.020967110991477966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,2560,0.11994043986002605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,10240,0.012480888929631976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,1536,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,128,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,7168,0.024188444018363953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,512,0.0035742223262786865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,512,0.013225778109497495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,8192,0.024488000406159297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,8192,0.05132888754208883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,5120,0.04941866795221964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,12288,0.7295546531677246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,4096,0.016000888413853116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,10240,0.02295644415749444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,64,0.004230222354332606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,256,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,2560,0.03559199968973795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,3584,0.2933644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,3072,0.007337777978844113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,512,0.17624266942342123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,512,0.029451555675930444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,8192,0.055494222376081676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,16384,0.2969751093122694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,3072,0.021278222401936848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,64,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,12288,0.06956177949905396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,1536,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,10240,0.3474888801574707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,32,0.005389333185222413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,5120,0.02725155486000909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,51200,0.11095022492938572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,5120,0.026055999928050574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,4096,0.034372445609834455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,4096,0.04770044485727946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,65536,0.17350755797492134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,12288,0.04591111010975308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,2560,0.010775999890433418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,2048,0.39887022972106934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,32,0.008093333078755273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,51200,0.0661333335770501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,512,0.01479111115137736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,512,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,2048,0.008214222060309516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,8192,0.010772444307804108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,10240,0.0474275549252828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,65536,0.08213422033521864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,10240,0.10537066724565293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,16384,0.07750933037863837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,3072,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,5120,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,256,0.009891555541091496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,768,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,1536,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,8192,0.07190489106708102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,7168,0.05431644452942742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,256,0.014259555273585849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,12288,1.8981342315673828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,6144,0.027464000715149775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,10240,0.02835644284884135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,6144,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,1536,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,7168,0.2001662254333496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,12288,0.020938666330443487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,16384,0.11256889502207439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,768,0.03096710973315769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,4096,0.02163022259871165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,2560,0.022445332672860887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,1536,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,8192,0.15405689345465765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,10240,0.03748622205522325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,5120,0.01016711112525728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,128,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,1536,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,768,0.03660888804329766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,6144,0.08714933527840509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,512,0.035515556732813515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,256,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,32,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,256,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,768,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,51200,0.03813066747453477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,256,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,128,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,3072,0.022336888644430373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,7168,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,512,0.0025884444928831523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,12288,0.730679988861084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,768,0.02335111134582096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,10240,0.04386666748258802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,5120,0.02573066618707445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,4096,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,256,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,65536,0.6797768804762098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,3584,0.018978666928079393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,12288,0.4570639928181966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,5120,0.016883555385801528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,1024,0.007189333438873291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,16384,0.0251528885629442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,65536,0.31158754560682506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,768,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,4096,0.07560533285140991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,12288,0.045888887511359326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,1536,0.0446648895740509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,1536,0.020787555310461257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,2048,0.033891555335786604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,2048,0.019984000258975558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,32,0.006215999937719769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,128,0.06759644216961332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,10240,0.016711110870043438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,2560,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,6144,0.019923556182119582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,2560,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,10240,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,6144,0.019205333458052743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,51200,0.09254488680097793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,16384,0.10760888788435195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,1536,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,128,0.014485332700941296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,7168,0.04418844315740797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,8192,0.035918222533331975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,3584,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,128,0.005713777823580637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,16384,0.021888888544506494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,8192,0.03288977675967746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,1024,0.0317573348681132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,7168,0.011676444775528379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,65536,0.35624265670776367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,3584,0.015039111177126566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,65536,0.41886043548583984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,12288,1.9788729349772136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,12288,0.017384888397322763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,16384,0.03453333179155985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,512,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,128,0.02123466630776723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,512,0.05472977624999153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,65536,0.8875430954827203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,3072,0.008787555827034844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,51200,0.09304622146818374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,32,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,2560,0.030392001072565716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,5120,0.5826675626966689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,512,0.009128889275921715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,1024,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,7168,2.1513875325520835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,32,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,3584,0.026572444372706946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,8192,0.07139822509553698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,3072,0.023337778117921617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,51200,0.11883021725548638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,256,0.005722666780153911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,12288,0.014718222949239941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,128,0.0018746666610240936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,512,0.0081395556529363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,32,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,8192,0.06826221942901611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,1024,0.14098132981194392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,51200,0.12047733200920953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,1024,0.023642667465739783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,65536,0.18456710709465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,1024,0.021968000464969214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,5120,0.060848891735076904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,4096,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,10240,0.18190311061011422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,10240,0.045349333021375865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,8192,0.033598221010631986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,128,0.005665777872006099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,6144,0.010240889257854885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,64,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,16384,0.32435377438863117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,65536,0.16444355911678737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,512,0.0042160000238153665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,512,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,3072,0.011699555648697747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,4096,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,6144,0.10919644435246785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,3584,0.05034844411744011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,512,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,5120,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,1024,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,3072,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,16384,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,512,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,5120,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,3584,0.043572445710500084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,1536,0.013471999930010902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,1536,0.003536000019974179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,10240,0.0251528885629442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,3072,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,2048,0.026870222555266485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,768,0.03637511200375027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,768,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,3072,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,16384,0.18137599362267387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,16384,0.15333155790964761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,4096,0.049660444259643555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,8192,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,8192,0.17391022046407065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,2048,0.037852442926830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,3072,0.006507555643717448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,65536,0.06338489055633545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,16384,0.03601422243648105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,5120,0.02942044536272685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,2560,0.07638488875495063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,3584,0.08772266573376125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,10240,0.02195022172398037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,1536,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,512,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,8192,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,256,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,4096,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,768,0.013328889177905189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,768,0.0075395554304122925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,65536,0.09962666696972317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,51200,0.4263795481787787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,10240,0.09947999980714585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,12288,0.2268142170376248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,4096,2.4333207872178817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,512,0.05966577927271525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,64,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,5120,0.16792533132765028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,8192,0.03714399867587619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,512,0.06492088900672065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,512,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,2560,0.017522666189405654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,768,0.0041697778635554844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,4096,0.06716889142990112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,1536,0.008143111235565608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,256,0.0041511112617121804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,1536,0.019810666640599568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,65536,0.15017599529690212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,64,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,512,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,128,0.01498044364982181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,16384,1.1170507007175021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,1536,0.061342219511667885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,1024,0.01921866668595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,3584,0.012835555606418185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,32,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,3584,0.01426666643884447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,6144,0.009197333620654212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,12288,0.03810933232307434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,4096,0.00792622235086229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,768,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,2048,0.15115643872155085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,3584,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,12288,0.011104889214038849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,1536,0.013722666435771518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,256,0.01794933279355367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,5120,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,2560,0.054804444313049316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,6144,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,256,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,65536,3.914536794026693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,65536,0.010815111299355825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,768,0.022915555371178523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,1024,0.018260444204012554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,6144,0.05667555332183838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,51200,0.04615733358595106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,512,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,64,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,2048,0.017874666386180453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,3072,0.018558222386572096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,8192,0.07861422167883979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,3584,0.01590133375591702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,256,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,12288,0.032540443870756365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,2048,0.20845066176520455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,4096,0.007881777981917063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,2048,0.125782224867079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,5120,0.031094223260879517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,3584,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,8192,0.009376000199053023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,10240,0.06710400183995564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,10240,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,65536,0.06766666968663533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,4096,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,51200,0.052276445759667285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,1536,0.05941333373387655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,3072,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,4096,0.024235554867320593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,1024,0.0028435554769304064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,3072,0.02937244375546773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,5120,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,1536,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,65536,0.07611822419696383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,8192,0.1777520047293769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,16384,0.024937777055634394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,1024,0.03156177865134345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,16384,0.016509332590632968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,51200,1.8694053226047094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,12288,0.25647910435994464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,6144,3.0232586330837674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,3584,0.024655999408827886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,6144,0.013394667042626275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,1024,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,12288,0.01715288890732659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,4096,0.07757155762778388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,12288,0.02927288744184706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,10240,0.013376000026861826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,1024,0.024018666810459558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,12288,0.030902223454581365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,65536,0.7228373421563042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,1024,0.01275111072593265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,5120,0.03219199842876858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,16384,0.029820445511076186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,12288,0.029948443174362183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,768,0.030209776428010728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,3072,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,2048,0.1605102221171061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,3584,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,256,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,1024,0.0063440000845326324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,51200,0.396816889444987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,3072,0.053080889913770885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,16384,0.024396444360415142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,64,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,10240,0.08492177724838257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,4096,0.03962044583426581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,1536,0.03147911032040914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,1024,0.02569244470861223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,5120,0.008691555923885768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,65536,0.31583910518222386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,768,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,768,0.02417244513829549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,768,0.09771733151541816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,3072,0.03269777695337931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,32,0.0028373334142896864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,3072,0.02269866731431749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,6144,0.07255822420120239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,3584,0.031191110610961914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,10240,0.15660533640119764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,2048,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,2560,0.007342221836249034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,16384,0.1169368823369344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,51200,0.15129155582851833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,51200,13.876639472113716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,768,0.02551022171974182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,2560,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,2048,0.025937777426507738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,4096,0.10665333271026611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,1024,0.008053333395057255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,8192,0.1249119970533583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,51200,0.06128978066974216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,10240,0.023269333773189124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,768,0.027427555786238775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,512,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,2560,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,16384,0.01221155540810691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,10240,0.11422577169206406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,2048,0.014891554911931356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,1024,0.19011822011735705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,12288,0.32547110981411403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,16384,0.04445599847369724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,4096,0.018183110488785636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,256,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,1024,0.004548444516128964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,3072,0.019780443774329293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,12288,0.06160177787144979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,5120,0.03320088982582092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,128,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,3584,0.005036444299750858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,5120,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,51200,0.059769776132371694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,128,0.003531555748648114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,5120,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,128,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,3584,0.006701333241330252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,65536,0.06839555501937866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,65536,1.7444818284776475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,3584,0.21471199724409315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,4096,0.10770666599273682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,4096,0.01419911119672987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,6144,0.048384000857671104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,7168,0.012824888858530255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,12288,0.13299466503991023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,256,0.017376888129446242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,32,0.1370097796122233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,16384,0.04878488845295376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,2560,0.02867199977238973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,8192,0.010710222025712332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,65536,0.3370062245262994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,256,0.017091555727852713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,32,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,12288,0.02609511050913069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,4096,0.030829333596759375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,3584,0.8086222012837728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,7168,0.05414755476845635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,64,0.009275555610656738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,7168,0.02499466637770335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,32,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,768,0.007820444802443186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,256,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,4096,0.04355555441644457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,1536,0.026769777139027912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,12288,0.21342755688561332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,4096,0.02351733379893833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,768,0.015471110741297403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,64,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,5120,0.05544266766972012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,7168,0.02156977852185567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,8192,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,8192,0.015297777122921415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,10240,0.04762844575775994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,16384,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,3584,0.24785065650939941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,7168,0.04737777842415703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,512,0.009198221895429824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,4096,0.048912889427608915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,1536,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,65536,0.06781422429614596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,5120,0.0247653325398763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,6144,0.04734666811095344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,64,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,32,0.005518222020732031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,32,0.015427556302812366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,1536,0.04788000053829617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,256,0.06144089168972439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,32,0.007824000385072496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,1024,0.022484444909625586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,8192,0.1622017754448785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,768,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,51200,0.12781066364712185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,51200,0.6374070909288194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,6144,0.03527199890878465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,3584,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,7168,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,512,0.0027884443600972495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,3584,0.026007999976476032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,10240,0.01643733349111345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,2048,0.02049333353837331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,3072,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,2048,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,1024,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,1024,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,3072,0.02272622287273407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,51200,6.072842491997613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,5120,0.022397332721286353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,3584,0.008098666866620382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,3072,0.03858933183881972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,7168,0.02811466654141744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,65536,0.5469626850552028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,7168,0.017849778135617573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,10240,0.035657776726616755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,10240,0.046499556965298124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,2560,0.02681688964366913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,16384,0.010985777609878115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,5120,0.047092444366878934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,64,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,768,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,8192,0.062881780995263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,4096,0.017662222186724346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,6144,0.19495822323693168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,512,0.01921066641807556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,6144,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,2560,0.05917511383692423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,10240,0.03405688868628608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,32,0.003301333429084884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,2560,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,12288,0.21055732833014595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,7168,0.26547643873426646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,256,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,3584,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,2560,0.045648889409171216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,3072,0.026022222306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,3072,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,2560,0.06294844547907512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,128,0.014867555763986377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,8192,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,16384,0.3002977901034885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,7168,0.023383999864260357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,64,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,2048,0.014886221951908536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,12288,0.13410754998524985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,7168,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,16384,0.08279199732674493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,8192,0.016001777516471017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,3072,0.019660444723235238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,2560,0.007672888537247975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,12288,0.02390755547417535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,16384,0.06796266635258992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,2048,0.03722044494416978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,7168,0.04576177729500664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,6144,0.014464888307783337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,5120,0.04107822312249078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,2048,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,10240,0.06641333632998996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,2048,0.026212443908055622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,6144,0.08468711376190186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,2560,0.01641688909795549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,2560,0.014919110470347934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,51200,0.11309599876403809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,51200,1.2277652952406142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,10240,0.7516195509168836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,1536,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,8192,0.20398933357662627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,51200,0.20454133881462944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,32,0.006433777925040986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,3072,0.031268444326188825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,6144,0.18591288725535074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,32,0.018576888574494254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,10240,0.02641600039270189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,128,0.015241778559154935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,7168,0.027789334456125896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,256,0.020787555310461257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,5120,0.029466665453381006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,512,0.004011555678314633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,512,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,1536,0.04043289025624593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,3584,0.044723557101355664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,2048,0.10167111290825738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,768,0.014583999911944071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,51200,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,4096,0.016237333416938782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,16384,1.3231102625528972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,16384,0.04485511117511325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,3584,0.03915111223856608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,2048,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,10240,0.041725334193971425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,16384,0.3509688907199436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,4096,0.02809422214825948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,3584,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,2560,0.026935999592145283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,65536,0.21254044108920625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,7168,0.6050497690836588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,51200,0.4382879998948839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,51200,0.10609866513146295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,1024,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,768,0.003565333369705412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,51200,0.061781333552466504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,51200,0.06747821966807048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,4096,0.29155288802252877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,16384,0.05099466774198744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,7168,0.2071831093894111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,65536,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,3584,0.033153778976864286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,51200,0.26145778761969674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,6144,0.022287110487620037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,5120,0.039292444785436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,128,0.04526755544874403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,1536,0.02427111069361369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,6144,0.012356444365448423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,1536,0.02442933287885454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,5120,0.008935111264387766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,256,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,12288,0.11091288593080308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,12288,0.00961511085430781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,16384,0.2639457649654812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,10240,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,16384,1.2809581756591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,768,0.008941333327028487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,6144,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,6144,0.05808977948294746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,16384,0.05003733436266581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,3584,0.04353333181805081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,64,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,3584,0.04208711120817396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,3584,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,5120,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,7168,0.014926221635606555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,512,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,3584,0.006803555621041193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,2560,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,1024,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,768,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,5120,0.020424000091022916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,10240,0.012304888831244575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,8192,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,8192,0.13326222366756862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,10240,0.029532445801628962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,7168,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,128,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,16384,0.12185866302914089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,7168,0.029184887806574505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,512,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,1536,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,65536,0.19052978356679282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,8192,0.024665777881940205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,10240,1.0602116054958768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,1024,0.08204444249471028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,3584,0.023924445112546284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,6144,0.04440977838304308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,128,0.022265776991844177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,5120,0.21031199561225045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,4096,0.023962666591008503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,2048,0.023448000351587932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,5120,0.04851822058359782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,6144,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,3072,0.11505866050720215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,5120,0.033414221472210355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,2560,0.01772533357143402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,51200,1.8890372382269964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,1024,0.07686577902899848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,32,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,8192,0.10108088784747654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,6144,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,51200,0.9138346778021919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,1536,0.008566222257084316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,6144,0.026177777184380427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,5120,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,3584,0.021653332644038733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,7168,0.013399110900031196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,768,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,51200,0.8065297868516711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,4096,0.0793386697769165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,65536,0.44146132469177246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,65536,1.201314714219835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,2048,0.016567111015319824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,2048,0.09334311220380996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,256,0.016703110602166917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,3072,0.012749333348539142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,3584,0.16195021735297307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,5120,0.01690933273898231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,16384,0.07512355513042875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,7168,0.583479086558024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,1024,0.020943111843532983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,65536,0.06692533360587226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,8192,0.0978142221768697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,16384,0.06524799929724799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,256,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,32,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,2048,0.025622222158643935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,1536,0.014529777897728814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,256,0.013789333403110504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,6144,0.598815123240153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,256,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,4096,0.020455999506844413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,256,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,5120,0.024996444582939148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,65536,0.20495200157165527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,7168,0.014392889208263822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,256,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,7168,0.012206222448084088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,5120,0.253057771258884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,3072,0.3026506635877821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,256,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,10240,0.017539555827776592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,32,0.018951111369662814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,8192,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,256,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,65536,0.14792177412245008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,3584,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,7168,0.19505955113304985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,512,0.011504000259770287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,6144,0.16026133961147732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,6144,0.008216000265545314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,51200,0.09499466419219971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,512,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,1536,0.015966221690177917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,7168,0.019654222660594516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,3072,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,768,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,16384,0.057890666855706104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,3072,0.03727111220359802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,51200,0.07720355855094062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,768,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,256,0.017551110850440133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,6144,0.021406221720907424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,16384,0.16650488641526964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,3072,0.05833333068423801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,7168,0.03526755505137973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,3072,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,512,0.01756177842617035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,10240,0.016466667254765827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,2048,0.004201777693298128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,8192,0.03130577670203315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,3584,0.050795555114746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,7168,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,1536,0.013047999805874295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,5120,0.1456071138381958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,3072,0.06636889113320245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,3584,0.49115467071533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,12288,0.011395555403497485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,1536,0.0251511103577084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,6144,0.0351164440313975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,16384,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,256,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,2560,0.00960533320903778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,768,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,12288,0.0882800022761027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,768,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,768,0.03256977929009332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,2560,0.02314222190115187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,2048,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,3584,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,10240,0.24349954393174914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,8192,0.014575999644067554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,8192,0.13120178381601968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,8192,0.0391937792301178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,7168,0.019564444820086162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,6144,0.05505511164665222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,7168,0.039974222580591835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,16384,0.08122577932145861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,2560,0.007231111327807109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,5120,0.022447999980714586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,1536,0.035691556003358625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,768,0.0031884445084465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,10240,0.06777244144015841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,6144,0.06042489078309801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,10240,0.013199110825856527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,768,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,51200,7.191168891059028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,1536,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,256,0.07151377863354154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,128,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,3072,0.022288888692855835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,2048,0.21834399965074328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,768,0.017290666699409485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,256,0.019768888751665752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,256,0.023062222533755835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,3584,0.09100977579752605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,3584,0.011004444625642566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,7168,0.032687111033333674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,4096,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,65536,0.19013422065311006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,512,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,256,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,7168,0.014342221948835584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,7168,0.2153786685731676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,1024,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,51200,0.16676089498731825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,4096,0.011515555282433828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,3072,0.030418667528364394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,1536,0.01424888935354021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,512,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,8192,0.015896888242827523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,16384,0.03501244386037191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,256,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,10240,0.01918044520748986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,65536,0.2904177771674262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,3584,0.010429333481523724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,65536,0.07149244679345025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,768,0.12042310502794053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,8192,0.06910044617123075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,5120,0.04807022213935852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,8192,0.015159999330838522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,256,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,768,0.016384000579516094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,768,0.02792444494035509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,64,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,256,0.03853777713245816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,1536,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,4096,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,16384,0.021924444370799597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,4096,0.08896533648173015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,128,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,512,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,1536,0.01969688965214623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,768,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,512,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,51200,0.41218845049540204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,3072,0.015271999769740634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,256,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,51200,0.057971556981404625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,1536,0.12986132833692762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,8192,0.056214219994015165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,3072,0.028436445527606543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,512,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,51200,0.09836088948779637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,65536,0.08075111442142062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,128,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,64,0.003273777870668305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,3072,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,16384,0.09134755531946818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,512,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,8192,0.08311111397213407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,12288,0.3643137878841824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,51200,1.151842647128635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,5120,0.035331557194391884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,3072,0.01014311114947001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,65536,0.06696089108784993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,7168,0.06325066751903959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,1536,0.044897778166664966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,8192,0.7460000250074598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,2560,0.04777600036727058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,3072,0.03056266572740343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,65536,0.07130666573842366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,7168,0.027637332677841187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,1536,0.004513777792453766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,4096,0.013023110727469126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,12288,0.07288889090220134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,65536,0.2372008959452311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,128,0.01440177857875824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,7168,0.056158224741617836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,1024,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,4096,0.017123555143674214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,3584,0.013927999469969006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,2560,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,51200,2.743732452392578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,1024,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,1024,0.009381333159075843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,128,0.015673778123325773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,51200,0.22060888343387178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,1024,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,7168,0.821824868520101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,16384,0.6261306868659126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,5120,0.02212000058756934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,16384,0.05011466807789273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,512,0.014570666684044732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,64,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,6144,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,7168,0.03802311089303758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,128,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,4096,0.01758399936887953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,256,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,51200,2.5078035990397134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,3584,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,3584,0.008060444560315874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,7168,0.06504088640213013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,512,0.01110666659143236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,256,0.013747555514176687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,4096,0.020629333125220407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,2560,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,65536,0.18989600075615776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,1024,0.010480889015727572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,6144,0.08696266677644517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,8192,0.051719999975628324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,1536,0.036192887359195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,51200,3.8138321770562063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,3584,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,3072,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,2560,0.11670577526092529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,16384,2.132449679904514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,1536,0.19427822695838082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,1024,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,3072,0.01440000037352244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,10240,0.037288000186284385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,512,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,51200,0.2675244543287489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,3072,0.08255733384026422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,2048,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,2048,0.018944000204404194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,12288,0.0147724449634552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,5120,1.9529476165771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,5120,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,768,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,51200,0.10616088575787014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,64,0.011396444506115384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,768,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,16384,0.3355991045633952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,2048,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,512,0.02271644439962175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,512,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,5120,0.01735377808411916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,2560,0.07137599918577406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,65536,0.352908452351888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,51200,0.05136977632840475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,3584,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,768,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,6144,0.01015466699997584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,6144,0.019479999939600628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,2560,0.03503555390569899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,1024,0.0772320032119751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,1536,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,2560,0.007522666619883643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,8192,0.10143999920950995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,5120,0.019384000036451552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,768,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,2560,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,7168,0.13276888264550105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,512,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,51200,0.021125333176718816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,768,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,51200,0.1188124418258667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,65536,0.4740284283955892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,1536,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,256,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,512,0.009943111075295342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,10240,0.011982222398122152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,1024,0.20487556192609999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,7168,0.05013155606057909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,8192,0.026525333523750305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,10240,0.03209244542651706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,65536,0.22971910900539824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,65536,0.19219556119706896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,6144,0.006165333506133821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,128,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,128,0.022650667362742957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,10240,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,16384,0.036975112226274275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,1024,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,12288,0.1324639982647366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,6144,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,5120,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,64,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,512,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,12288,0.02776266634464264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,10240,0.013169777890046438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,1536,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,1536,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,16384,0.16741332742902967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,65536,0.13445511129167345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,3072,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,3072,0.22688444455464682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,65536,0.27319465743170845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,1536,0.026499556170569524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,512,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,7168,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,768,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,10240,0.19955733087327746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,3072,0.02607111136118571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,64,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,12288,0.131221334139506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,512,0.003567999849716822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,5120,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,1536,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,12288,0.1003040009074741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,64,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,512,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,6144,0.03186310993300544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,10240,0.022821333673265245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,10240,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,3584,0.7787111070421008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,10240,0.1194302241007487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,3072,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,65536,0.0668319993548923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,1024,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,1024,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,6144,0.03055822186999851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,5120,0.016419554750124615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,5120,0.020805334051450092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,256,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,6144,0.027856889698240492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,2560,0.02471911079353756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,2048,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,3072,0.09813421964645386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,12288,0.04053422146373325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,4096,0.01108088923825158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,2560,0.10532266563839382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,65536,6.518309275309245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,5120,0.01924533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,128,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,1024,0.004067555483844546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,4096,0.07446577813890246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,7168,0.060271998246510826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,3584,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,256,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,10240,0.05852711200714111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,1024,0.009718221922715506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,128,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,32,0.05978666411505806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,7168,0.2332391209072537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,7168,0.041542222102483116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,3584,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,7168,0.013006221916940479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,16384,0.07471022340986463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,7168,0.010452444354693094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,1536,0.03258488906754388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,51200,0.31333687570359975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,768,0.029120001527998183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,6144,0.02641422218746609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,2560,0.07415911224153307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,1024,0.01738577749994066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,32,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,7168,0.019536000159051683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,2048,0.017236444685194228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,5120,0.012530666258600024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,5120,0.016919111212094624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,3072,0.03573333223660787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,16384,0.02037244372897678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,4096,0.05752178033192953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,8192,0.04279199904865689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,3072,0.087318221728007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,4096,0.020395555429988436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,128,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,512,0.015287111202875773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,10240,0.13009156121148005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,3072,0.0064497776329517365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,3584,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,3072,0.023295111126369897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,65536,0.8999395370483398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,768,0.02770044406255086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,1536,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,128,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,5120,0.019066666563351948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,3584,0.03068711029158698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,8192,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,4096,0.05132800009515551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,5120,0.2566026581658257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,32,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,768,0.032849778731664024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,65536,0.649793783823649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,1024,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,6144,0.02772444486618042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,3072,0.5428951051500108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,8192,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,512,0.013974222871992322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,16384,0.054458667834599815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,128,0.020404444800482858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,32,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,65536,0.23558489481608072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,512,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,8192,0.07792088720533583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,8192,0.03370399938689338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,64,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,2048,0.006551110910044775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,7168,0.010328888893127441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,512,0.0812915563583374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,64,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,32,0.001984888894690408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,16384,0.06350488795174493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,8192,0.015306666493415833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,4096,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,256,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,256,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,8192,0.06283999813927545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,64,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,2560,0.035720000664393105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,3072,0.0809697773721483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,10240,0.05227288934919569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,8192,0.03808177842034234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,64,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,1536,0.08067911201053195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,1536,0.041386667225095965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,3584,0.14403377638922796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,128,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,12288,0.9155315823025174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,2560,0.03003022074699402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,768,0.009062222308582729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,2048,0.05010133319430881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,6144,0.06061689058939616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,10240,0.4547644191318088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,51200,0.14052977826860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,8192,0.03563288847605387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,5120,0.049978666835361056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,128,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,1024,0.015215999550289579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,51200,0.29496532016330296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,2560,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,65536,0.06087733639611138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,768,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,2560,0.00812977800766627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,2048,0.004580444345871608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,1536,0.011404444773991903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,128,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,65536,0.05967289209365845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,768,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,768,0.07264888948864408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,2048,0.02459466622935401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,3072,0.014025777578353882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,1536,0.04752888944413927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,7168,0.015824000040690105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,512,0.0026524443593290118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,10240,0.07176533010270861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,12288,0.18905866146087646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,65536,0.08429600132836236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,64,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,51200,0.050992889536751636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,10240,0.24379822942945692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,16384,0.5559280183580186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,51200,0.11491200659010146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,16384,0.3937831189897325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,256,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,5120,0.0314666661951277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,51200,0.17277155982123482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,51200,0.24235910839504668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,32,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,51200,0.8184426625569662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,3584,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,16384,0.026760000321600173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,3072,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,512,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,2048,0.03407466742727492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,65536,0.13189244270324707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,3584,0.023327999644809302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,2048,0.0611635579003228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,65536,0.04723111126157972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,65536,0.5637804667154948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,16384,0.27258843845791286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,2048,0.016996444927321542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,128,0.040639999839994646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,6144,0.016488000750541687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,1536,0.0064151109092765385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,7168,0.046872887346479625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,128,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,4096,0.02500088844034407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,6144,0.009335111412737105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,512,1.779430177476671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,1024,0.014148443937301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,3584,0.10276977883444892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,51200,0.20205244753095838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,2048,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,1024,0.006510222123728857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,512,0.0041271112859249115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,256,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,51200,0.05196711089875963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,16384,0.4628435770670573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,2048,0.010482666393121084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,65536,1.3134124543931749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,32,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,4096,0.14339289400312635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,3584,0.05415999889373779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,7168,0.02590755621592204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,3072,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,3072,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,1024,0.017989334132936265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,5120,0.06007644202974108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,8192,0.014682667122946845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,768,0.01923733287387424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,12288,2.857912063598633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,768,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,64,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,768,0.005660444498062134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,256,0.9360311296251086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,6144,0.016833777228991192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,65536,0.07963555388980441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,512,0.06145422326193916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,6144,1.8758267296685114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,8192,0.015423999892340766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,3584,0.008953777452309927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,8192,0.033484445677863225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,5120,0.013064888616402945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,2560,0.007696888513035244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,3072,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,256,0.004264889078007804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,3584,0.01310755560795466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,256,0.02872977654139201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,2560,0.03791288865937127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,10240,0.04830577638414171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,2560,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,64,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,2048,0.020278222031063504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,5120,0.026919111609458923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,5120,0.019907555646366544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,4096,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,12288,1.1323857837253146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,128,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,10240,0.007838221887747446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,7168,0.42391112115648055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,12288,0.029657777812745836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,10240,0.02920177910063002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,7168,0.02772355576356252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,3584,0.009523555636405945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,5120,0.014741332994567024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,768,0.027816000911924574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,1024,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,65536,0.22400710317823622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,128,0.0035439998739295532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,256,0.037631111012564764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,5120,0.016747555798954435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,128,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,10240,0.016327111257447135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,2048,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,10240,0.011694221860832639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,51200,0.30867822964986164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,16384,0.03917510973082648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,10240,0.02964622113439772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,7168,0.03974044322967529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,64,0.02663999961482154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,65536,0.21521600087483725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,16384,0.039268443981806435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,12288,0.02596355477968852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,512,0.006062222023804982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,1536,0.020392888122134738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,51200,0.4339475631713867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,6144,0.339121765560574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,512,0.004297777596447202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,3584,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,3072,0.01899111105336083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,7168,0.24211200078328451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,1024,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,6144,0.2928684552510579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,10240,0.11319643921322292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,512,0.007328000333574083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,512,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,10240,0.011640888949235281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,768,0.07293777995639376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,7168,0.013086222112178802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,12288,0.40146401193406844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,1536,0.029156443145540025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,8192,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,6144,0.019409777389632333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,65536,0.1758791075812446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,65536,0.9233919779459635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,32,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,5120,0.049397332800759204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,51200,3.9010433620876737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,4096,0.038034667571385704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,5120,0.008544888761308458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,7168,0.010238221950001186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,768,0.018648000227080453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,10240,0.05898400147755941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,3072,0.03773066732618544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,128,0.0324942237801022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,65536,0.2827564345465766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,4096,0.009737778041097853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,768,0.02569155560599433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,16384,0.04100266761249966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,256,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,7168,0.0311119986904992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,6144,0.02496799992190467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,1536,0.3489697774251302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,16384,0.24376177787780762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,7168,0.02329066726896498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,1024,0.008422222402360704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,256,0.0063573333124319715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,7168,0.023361777265866596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,512,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,4096,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,64,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,4096,0.12084799342685276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,4096,0.033281776640150286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,2560,0.020455110404226515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,51200,0.07105955812666151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,1536,0.019316444794336956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,12288,0.40236443943447536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,16384,0.0166631109184689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,6144,0.004549333204825719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,5120,0.018409777018758986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,3072,0.012082666986518435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,64,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,1536,0.006002666635645761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,5120,0.4521502388848199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,10240,0.14904088444179958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,256,0.005872000008821487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,51200,1.9093981848822699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,2048,0.0064080001579390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,7168,0.017330666383107502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,2048,0.033921778202056885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,5120,0.07207110855314466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,1536,0.19185511271158853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,65536,0.0778871112399631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,6144,0.028713779317008123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,2048,0.04676889048682319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,4096,0.09692000018225776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,16384,0.05180444320042928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,1536,0.042931556701660156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,32,0.0025439999169773525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,768,0.01957066688272688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,512,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,8192,0.011476444701353708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,2560,0.03652088840802511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,4096,0.048546665244632296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,51200,0.04762044548988342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,8192,0.10646133290396796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,51200,0.5703528722127279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,7168,0.20553155740102133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,3584,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,32,0.00868711123863856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,3072,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,12288,0.0185199992524253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,1024,0.01000711073478063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,128,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,65536,1.0163902706570096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,65536,0.8408737712436252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,1024,0.020742222666740417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,256,0.0129439996348487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,16384,0.026496888862715826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,12288,0.01886044442653656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,51200,0.12416888607872857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,512,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,4096,0.05193066596984863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,5120,0.07601422071456909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,4096,0.0882871084743076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,1024,0.005700444595681296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,10240,0.15394222736358643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,768,0.030942221482594807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,2560,0.14717688825395372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,8192,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,10240,0.1326391167110867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,768,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,2048,0.007343110938866933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,12288,0.022672888305452134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,1024,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,768,0.0862364437845018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,128,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,16384,0.03228622343805101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,3072,0.029975109630160864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,2048,0.015617777903874716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,32,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,51200,0.16135644912719727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,3584,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,2048,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,3072,0.06872889068391588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,512,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,64,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,256,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,1536,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,256,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,12288,0.013376889129479727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,2048,0.00699733363257514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,10240,0.04080266753832499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,512,0.013600889179441663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,128,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,5120,0.03731911049948798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,8192,0.08515199687745835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,512,0.018532445033391316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,6144,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,5120,0.02535999980237749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,3072,0.04337244563632541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,1024,0.024381332927280005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,256,0.006160888820886612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,1024,0.03240622083346049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,7168,0.23650132285224068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,128,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,51200,15.874838935004341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,65536,0.14009333981408015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,5120,0.015987555185953777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,768,0.04146044452985128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,5120,0.020728000336223178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,2560,0.010092444717884064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,64,0.005530666559934616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,2560,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,32,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,6144,0.027447111076778833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,10240,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,4096,0.04098666707674662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,16384,0.016524444023768108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,768,0.028759999407662287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,12288,0.5127866533067491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,5120,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,2048,0.004597333156400257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,7168,0.03628888726234436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,7168,0.015855999456511605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,32,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,2048,0.00850933293501536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,6144,0.01145866678820716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,2560,0.015791111522250705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,512,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,3584,0.08636622296439277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,12288,0.8380648824903699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,1536,0.0063271112740039825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,16384,0.12666932741800943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,4096,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,256,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,51200,1.8640302022298176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,768,0.011196444431940714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,16384,0.3230160077412923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,12288,1.2744390699598525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,8192,0.028063999281989202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,3584,0.012240889171759287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,51200,0.4869582388136122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,10240,0.027373333772023518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,7168,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,7168,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,3072,0.07774133152431913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,12288,0.08451466427909003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,2560,0.05360355642106798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,1024,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,51200,0.0692231125301785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,4096,0.021393777595625982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,768,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,8192,0.018889778190188937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,10240,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,1536,0.006808888994985157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,6144,0.07102488809161715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,12288,0.016935999194780987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,1536,0.010869333313571082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,51200,0.514212449391683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,2560,0.05288444293869866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,128,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,8192,0.030187557140986126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,10240,0.16589332951439753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,32,0.01755644381046295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,5120,0.02405422263675266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,768,0.03523822294341193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,2048,0.10271644592285156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,1024,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,32,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,16384,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,5120,0.03669155637423197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,3584,0.03049600124359131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,768,0.04016266597641839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,10240,0.03625866770744324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,1536,0.013454222016864352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,8192,0.011003555523024665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,1024,0.02295200030008952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,51200,0.1922453376981947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,6144,0.44835466808742946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,3584,0.015285332997639975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,2048,0.04858577913708157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,768,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,3072,0.3253493309020996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,3072,0.04873422119352552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,3584,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,8192,0.08673422204123603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,128,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,512,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,3072,0.052892443206575185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,1536,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,3072,0.018405333161354065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,6144,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,2048,0.02019733356104957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,51200,0.17148799366421172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,1536,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,768,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,3072,0.050789333052105375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,10240,0.8050862418280708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,6144,0.12155289120144314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,32,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,10240,0.041949331760406494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,128,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,51200,0.9784604178534614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,10240,0.058034665054745145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,51200,0.9333217408921984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,65536,0.05636622508366903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,1536,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,16384,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,7168,0.040707555082109236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,256,0.006515555497672822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,512,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,2048,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,256,0.006777777853939269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,65536,0.09088889095518325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,65536,0.8718498018052844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,65536,0.6815262370639377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,2560,0.020103999310069613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,1024,0.02825066778394911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,3584,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,2048,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,128,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,4096,0.21640088823106554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,2048,0.03635999891493056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,64,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,65536,0.1377155515882704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,2560,0.010258666343159145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,512,0.007970666719807519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,65536,26.79909939236111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,10240,0.15247378084394667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,16384,0.023963555693626404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,51200,0.04658222198486328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,5120,0.10650310913721721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,64,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,51200,5.985616048177083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,3072,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,12288,0.01130488928821352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,512,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,10240,0.03667555583847894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,2048,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,16384,0.0236124445994695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,256,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,8192,1.5997777514987523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,2560,0.009528000321653154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,4096,0.03108711043993632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,16384,0.22788800133599174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,1536,0.00870488915178511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,2048,0.003944888710975647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,32,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,3072,0.019139554765489366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,3072,0.021941334009170532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,10240,0.026906667484177485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,5120,0.09557066361109416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,5120,0.027429333991474573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,7168,6.732858869764541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,51200,0.054710222615136035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,8192,0.07988533046510485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,32,0.0028817777832349143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,8192,0.018774222996499803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,3584,0.014704888065656027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,3072,0.011607111328177981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,3072,0.02421244482199351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,6144,0.2244266668955485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,8192,0.04615911179118686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,2048,0.08273066414727105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,1536,0.01868888901339637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,3072,0.025916443930731878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,8192,0.01236533373594284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,768,0.00452533322903845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,2048,0.04274310999446445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,8192,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,10240,0.041702220837275185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,8192,0.014314666390419006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,16384,0.016488000750541687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,6144,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,51200,0.09223377704620361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,6144,0.010305778019958073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,12288,0.03490044342146979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,2560,0.05845599704318576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,6144,0.01144266708029641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,2048,0.011844444606039258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,128,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,65536,0.19446044498019746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,768,0.018040888839297824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,7168,0.050832887490590416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,1024,0.012440000143316058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,768,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,2048,0.023450666003757056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,1536,0.080549332830641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,65536,0.02755466600259145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,8192,0.03131822082731459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,8192,0.022784888744354248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,16384,0.04618844389915466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,512,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,7168,0.018561777141359117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,3584,0.008476444416575963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,10240,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,16384,1.8641821543375652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,3584,0.01128888875246048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,256,0.027469333675172593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,2560,0.03882311118973626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,1536,0.06512977679570515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,1024,0.008035555481910706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,3584,0.031599998474121094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,6144,0.030200888713200886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,7168,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,51200,1.314752048916287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,1536,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,32,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,1024,0.053651554716957935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,3072,0.07141866948869494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,16384,0.024225778049892847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,768,0.026791110634803772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,8192,0.03967466619279649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,4096,0.06110577450858223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,1536,0.0196142229768965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,6144,0.14810844262441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,8192,0.020584000481499564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,65536,0.07791733079486422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,4096,0.09872355726030137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,51200,0.07948978079689874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,7168,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,10240,0.07609155442979601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,4096,0.007907555335097844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,6144,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,64,0.005126222140259213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,16384,0.05167822374237908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,128,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,6144,0.06849422057469685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,1024,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,512,0.07912711302439372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,2048,0.048228445980283946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,1024,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,3584,0.035162667433420815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,2560,0.06489955716662936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,7168,0.009922666682137383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,65536,0.06820978058709039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,256,0.017854221993022494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,2560,0.6030009057786729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,256,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,10240,0.01716088917520311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,2048,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,2048,0.02562933332390255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,8192,0.02869422237078349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,65536,0.5782897737291124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,1024,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,6144,0.10003199842241074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,8192,0.06109244293636746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,8192,0.011100444528791638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,65536,0.12966577212015787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,5120,0.021147555775112573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,4096,0.05237866772545708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,1024,0.00452622233165635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,2560,0.07937244574228923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,4096,0.083624886141883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,3584,0.03958044449488322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,512,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,51200,0.053344001372655235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,1024,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,6144,9.703870137532553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,512,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,256,0.030964444080988567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,10240,0.019887111253208585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,256,0.012498666842778524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,8192,0.02225333286656274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,3584,0.15869777732425266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,768,0.003891555385457145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,5120,0.6429253154330784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,4096,0.04475733306672838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,51200,0.08517599768108791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,1536,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,65536,0.17377155356936982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,256,0.015375110838148328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,768,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,1024,0.007334222396214803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,256,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,768,0.11721866660647923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,2560,0.012690666649076672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,16384,0.016050666570663452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,2560,0.004553777890072929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,1536,0.0035946667194366455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,512,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,512,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,8192,0.03000266684426202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,6144,0.006802666518423293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,8192,0.013249778085284762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,4096,0.025880000657505457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,5120,0.4661537806193034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,256,0.01818933255142636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,512,0.011419555379284753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,768,0.1586115492714776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,51200,0.11446577972835964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,256,0.020638222495714825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,3584,0.04468355576197306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,768,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,512,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,7168,0.17987288369072807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,768,0.020798222886191476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,512,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,10240,0.01457155578666263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,768,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,128,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,6144,0.09869688749313354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,7168,0.010375111467308467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,1536,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,32,0.0018977777411540349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,64,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,16384,0.21768621603647867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,1024,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,16384,0.02090488870938619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,4096,0.02643555568324195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,51200,0.06846844487720065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,1536,0.02827377782927619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,128,0.014575110541449653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,65536,0.3002444373236762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,7168,0.010501333408885531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,51200,0.1290782160229153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,1024,0.017868444323539734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,5120,0.04965244399176704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,10240,0.016573333077960543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,1024,0.012496888637542725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,5120,0.03220266766018338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,1536,0.003996444245179494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,32,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,64,0.005982222242487802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,2048,0.0074240002367231585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,2560,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,3072,0.055016001065572105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,7168,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,16384,0.04383111000061035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,128,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,2048,0.026742221580611333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,1024,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,1536,0.010402667025725046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,64,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,3072,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,3072,0.011844444606039258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,2560,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,51200,0.10702311330371433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,2048,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,1024,0.01622222198380364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,512,0.04828622274928623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,51200,0.12531556023491755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,3584,0.14292177889082167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,2560,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,3584,0.03999911083115472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,16384,0.07327377796173096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,12288,1.4573955535888672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,2048,0.03720266620318095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,2048,0.011535999675591787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,512,0.0068764446510208984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,51200,0.052728000614378184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,1536,0.004038222134113312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,65536,0.3299066755506727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,256,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,3072,0.03801155421468947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,3584,0.05676266882154676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,128,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,16384,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,65536,0.18580710887908936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,256,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,12288,0.039148443275027804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,2560,0.01906133360332913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,128,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,4096,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,4096,0.008420444197124904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,12288,0.6124764548407661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,16384,0.04706844356324938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,32,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,256,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,2560,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,8192,0.017833777599864535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,64,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,7168,0.059883554776509605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,256,0.0284862220287323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,2048,0.020524443851576913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,256,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,128,0.8784355587429471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,4096,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,768,0.016914667354689706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,3072,0.018605333235528734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,5120,0.019142222073343065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,32,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,256,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,4096,0.008905777500735389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,768,0.030776888132095337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,4096,0.10649422142240737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,51200,0.15154400136735705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,256,0.0022248888595236675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,3584,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,7168,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,10240,0.01805688937505086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,3072,0.010456889039940305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,51200,0.5964568985833062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,6144,0.009956444303194681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,5120,0.023933332827356126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,2048,0.01237777786122428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,3072,0.044066667556762695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,4096,0.02536800007025401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,3584,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,3072,0.04531288809246487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,1024,0.013001778059535556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,2048,0.049361778630150684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,3072,0.011396444506115384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,128,0.027049778236283198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,512,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,128,0.001889777680238088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,10240,0.050379554430643715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,65536,10.906067742241753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,768,0.0454293323887719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,768,0.018848000301255118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,2560,0.005026666654480828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,16384,0.024334222078323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,64,0.05967377954059177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,6144,0.05961510870191786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,128,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,65536,0.08703555663426717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,512,0.00507377791735861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,3072,0.012806221842765808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,2560,0.012468444804350534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,6144,0.034585777256223894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,256,0.015082667271296183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,1536,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,6144,0.008718222379684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,2048,0.10210044516457452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,10240,0.7728115717569987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,256,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,64,0.013859555953078799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,16384,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,3584,0.19244444370269775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,12288,0.03474311033884684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,51200,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,51200,3.55165163675944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,512,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,6144,0.04199999901983473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,4096,0.02811288833618164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,4096,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,3584,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,2560,0.10756178034676446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,5120,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,1536,0.05548355645603604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,51200,0.07024888859854804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,1024,0.013660444153679742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,5120,0.0229751103454166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,64,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,8192,0.03973599937227037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,2048,0.0046160001721647055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,1024,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,6144,0.07773777511384752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,10240,0.04385422335730659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,6144,0.26558489269680446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,256,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,512,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,6144,0.01128533316983117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,2048,0.021965333157115515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,512,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,64,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,2048,0.02334844403796726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,5120,0.029024001624849107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,768,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,12288,0.04188444548183017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,51200,0.1390791071785821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,6144,0.10289955801433986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,3584,0.008200888832410177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,8192,0.023912888434198167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,12288,0.02923911147647434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,128,0.014931556251313953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,4096,0.09769511222839355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,512,0.003436444534195794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,512,0.14116711086697048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,2560,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,5120,0.04630400074852837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,8192,0.029063112205929224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,1024,0.04793511165512932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,5120,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,512,0.020414221617910597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,1536,0.27494221263461643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,3072,0.05052622159322103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,5120,0.00996800015370051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,1536,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,51200,1.3282204733954537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,6144,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,16384,0.016467556357383728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,2560,0.02458933326933119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,2048,0.029224889145957098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,12288,0.020605333977275424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,51200,0.07377066877153185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,1536,0.013375110924243927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,1024,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,10240,0.04342133469051785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,4096,0.013248888982666863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,8192,0.07343022028605144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,7168,0.0524213347170088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,2048,0.5083795653449165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,768,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,3584,0.24108444319831002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,8192,0.09696266386244033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,1536,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,1024,0.028568890359666612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,8192,0.03309422069125705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,4096,0.034296890099843345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,16384,0.225761784447564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,51200,13.267309400770399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,1024,0.03421422176890903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,7168,0.013986666997273764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,128,0.26936178737216526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,8192,0.13977955447302923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,32,0.003866666720973121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,768,0.014532445205582513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,3584,0.0662906699710422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,8192,2.5172790951199002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,768,0.004298666699065102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,1024,0.02234399980968899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,3072,0.014299554957283868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,10240,1.0888524585300021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,512,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,512,0.01538222200340695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,10240,0.02815022071202596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,4096,0.11581155988905166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,256,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,5120,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,6144,0.0673546658621894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,5120,0.042324443658192955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,2048,0.007806222471925948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,6144,0.03459199931886461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,7168,0.05057955450481839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,16384,0.7946782112121582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,1024,0.007671111159854465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,10240,0.2984071042802599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,65536,1.5240764617919922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,1024,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,51200,0.8657386567857531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,3584,0.45478132035997176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,2560,0.019549333386951022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,32,0.0028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,10240,0.02948533495267232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,3584,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,6144,0.010792888700962067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,6144,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,7168,2.9739636315239797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,4096,0.019705777366956074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,2560,0.0927199986245897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,256,0.029468443658616807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,8192,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,128,0.021737777524524268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,8192,0.012796444197495779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,6144,0.025088000628683303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,256,0.015244444211324057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,8192,0.05886844131681654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,51200,0.8894773059421115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,256,0.020957332518365648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,512,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,10240,0.06898488601048787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,51200,0.11994666523403591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,7168,0.023992889457278784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,12288,0.14725600348578557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,16384,0.02057422200838725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,2560,0.02270222206910451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,256,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,1024,0.014760888285107084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,65536,0.06334488921695285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,12288,0.35274222162034774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,4096,0.05836266941494412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,5120,0.020915556285116408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,256,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,3584,0.01646222174167633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,6144,0.09500000211927627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,65536,0.17800533771514893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,16384,0.016605334149466622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,3072,0.030925333499908447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,512,0.06693510876761542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,8192,0.024215110474162634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,5120,0.07281333208084106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,51200,4.943777720133464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,2560,0.04386933313475715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,256,0.0022719999154408774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,2048,0.05595733059777153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,5120,0.015781333049138386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,5120,0.018887110882335238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,7168,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,6144,0.0058142224119769196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,256,0.023805333508385554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,16384,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,10240,0.13789599471622044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,5120,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,3072,0.16471555497911242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,1536,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,2048,0.2771066559685601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,32,0.010793777803579966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,128,0.019627556204795837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,16384,0.06517066558202107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,128,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,2048,0.02091022166940901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,512,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,2560,0.023232888844278123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,8192,0.020305777589480083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,16384,0.3572559886508518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,65536,0.25859112209743923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,8192,0.06102222204208374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,256,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,1536,0.02611733310752445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,10240,0.05586222145292494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,1024,0.017885333961910672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,65536,0.08672710922029282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,512,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,1536,0.009385777844323052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,1024,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,8192,0.3467448817359076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,7168,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,64,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,8192,0.047304888566335045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,12288,0.41176711188422305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,64,0.0034799998005231223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,51200,0.052312887377209134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,32,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,6144,0.014870221416155497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,1024,0.018254222141371835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,4096,0.10265333122677273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,12288,0.1273039976755778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,32,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,10240,0.01885688967174954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,7168,0.7932586669921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,16384,0.3226871225568983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,51200,0.05700000127156576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,3584,0.02794755498568217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,3072,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,10240,1.156219588385688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,3584,0.01348444405529234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,51200,0.19276711675855848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,128,0.03820533222622342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,51200,0.3165217770470513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,1536,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,768,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,6144,0.08508977625105117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,5120,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,768,0.013330666555298699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,7168,0.04172711239920723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,8192,0.042504890097512134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,2560,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,2048,0.00423911131090588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,3072,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,768,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,8192,0.023077333966890972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,64,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,64,0.0019279999865425958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,65536,0.011600889265537262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,16384,0.1198542250527276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,32,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,1536,0.006765333314736684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,64,0.011316444310877057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,64,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,1536,0.007248889240953658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,2560,0.030786666605207656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,5120,0.015951111912727356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,2048,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,3584,0.011462222370836469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,2560,0.035001777940326266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,12288,0.01777777738041348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,8192,0.026695999834272597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,65536,0.20816977818806967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,512,0.04721066686842176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,6144,0.06897955470614962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,5120,0.041952000723944775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,4096,0.025133333272404138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,2048,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,65536,0.15977511141035292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,3072,0.026172444224357605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,256,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,1536,0.05583910809622871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,10240,0.11758666568332249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,8192,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,1024,0.05309333403905233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,16384,0.592084460788303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,512,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,51200,1.0561306211683485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,2048,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,51200,0.027661333481470745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,10240,0.027215111586782668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,32,0.05277599891026815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,16384,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,2048,0.6047671106126573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,3584,0.032010667853885226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,51200,20.529576619466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,10240,0.3647831016116672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,3072,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,3072,0.007798222204049428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,12288,0.024675556355052527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,1024,0.02472977836926778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,2560,0.019289778338538278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,256,0.01791822248035007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,4096,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,256,0.015798222687509324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,4096,0.010910222099887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,512,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,7168,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,5120,0.012791111237472959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,2560,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,512,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,7168,1.634195539686415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,128,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,16384,0.022016000416543748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,12288,0.034734222624037005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,7168,0.16056266095903185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,1024,0.010900444454616971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,3072,0.018174222773975797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,128,0.016223111086421542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,7168,0.019476445184813607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,10240,0.02842666705449422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,12288,0.028660443094041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,16384,0.4704337649875217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,768,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,1536,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,3584,0.22096889548831514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,51200,0.059232890605926514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,3584,0.039883553981781006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,5120,0.4768604702419705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,512,0.020727111233605277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,8192,0.044009778234693736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,1536,0.044481777482562594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,16384,0.27533777554829914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,2048,0.0092284447617001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,5120,0.05011111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,6144,0.36349066098531085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,8192,0.02880177895228068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,2048,0.020575111111005146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,1536,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,768,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,5120,0.01400888959566752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,1024,0.18660799662272134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,2560,0.004873777843183941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,128,0.02948088778389825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,4096,0.02369955513212416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,2048,0.012773333324326409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,7168,0.044518222411473594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,51200,0.04917600088649326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,65536,0.06743466854095459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,16384,0.07091466585795085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,6144,0.020795555578337777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,3584,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,51200,0.5416773160298666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,12288,0.018387556076049805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,2048,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,51200,0.10590222146775986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,2048,0.019316444794336956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,16384,0.043996443351109825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,16384,0.12495999866061741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,7168,0.018030222919252183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,64,0.005296000176005893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,256,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,16384,0.03477777706252204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,7168,0.13394399483998617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,65536,0.06819111108779907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,12288,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,32,0.007568000091446771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,128,0.020940444535679285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,256,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,6144,0.13738311661614314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,3072,0.045239110787709556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,256,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,1024,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,65536,0.14548533492618138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,12288,0.015043555034531487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,7168,0.027063111464182537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,2048,0.01516888870133294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,512,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,3584,0.01953866746690538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,768,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,768,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,8192,0.04455022348297966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,3584,0.04064711266093784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,512,0.013394667042626275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,3072,0.013437333206335703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,512,0.03291022115283542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,10240,0.04964888758129544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,32,0.0027884443600972495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,3072,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,1024,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,1024,0.005859555469618903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,1024,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,3584,0.07084088855319552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,51200,0.31548621919420033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,512,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,1536,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,32,0.0019164443429973389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,12288,0.4143839942084418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,3072,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,65536,0.15448533164130315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,12288,0.022486221459176805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,12288,0.014845333165592618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,512,0.024919110867712233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,2560,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,10240,0.014868444866604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,768,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,1024,0.019525332583321463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,32,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,512,0.14632266097598606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,1536,0.00775822252035141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,16384,0.14343555768330893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,1024,0.022191999687088862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,2560,0.010656889114114972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,3072,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,128,0.014946666028764514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,8192,0.024715556038750544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,4096,0.25604356659783256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,65536,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,2048,0.016759999924235873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,3584,0.19480266835954455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,65536,0.14982755978902182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,4096,0.06456889046563043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,2048,0.0236453331179089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,768,0.004640000147951974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,10240,0.06006044149398804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,5120,0.053416000472174756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,256,0.002516444358560774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,6144,0.023619555764728125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,2560,0.11104533407423232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,3584,0.04586311181386312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,1536,0.012646222280131446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,5120,0.5520248942905003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,1024,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,64,0.0041751112374994485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,64,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,128,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,2048,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,2048,0.043938666582107544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,12288,0.014670222997665405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,10240,0.022950222094853718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,256,0.08396355311075847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,6144,0.030710223648283217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,3072,0.03129333257675171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,8192,0.015968888998031616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,5120,0.384443547990587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,32,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,256,0.006467555546098285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,3072,0.033137778441111244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,16384,0.014916444818178812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,2560,0.007173333730962541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,3072,0.05627999703089396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,3584,0.01977599991692437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,3072,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,1024,0.3471173445383708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,1536,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,512,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,2048,0.14349955982632107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,51200,0.0510115557246738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,4096,0.026387555731667414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,3584,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,3072,0.022687110635969374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,4096,0.016294222739007738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,128,0.014193778236707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,6144,0.17422310511271158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,3584,0.08947644631067912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,12288,0.42149511973063153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,2560,0.10423022508621216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,512,0.014504000544548035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,7168,0.02875022093454997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,65536,0.6209288703070747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,12288,0.02165511084927453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,6144,0.14741599559783936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,16384,0.014961777461899651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,1536,0.0127368892232577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,3584,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,768,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,512,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,3584,0.043763554758495755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,65536,0.13115555710262722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,768,0.02163200080394745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,128,0.015178667174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,7168,0.1672391096750895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,5120,0.0136808885468377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,10240,0.05092089043723213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,8192,0.08733955356809829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,10240,0.028598222467634413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,128,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,32,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,3584,0.009536888864305284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,3584,0.021027555068333943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,65536,0.11616978380415176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,128,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,768,0.017889777819315594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,3072,0.0767457750108507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,7168,0.010348444183667501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,51200,0.05395466751522488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,12288,0.034684442811542086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,512,0.0681333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,1536,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,2048,0.008064000142945183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,5120,0.025137777129809063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,3584,0.11927555667029487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,4096,0.09309866693284775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,3584,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,512,0.16840710904863146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,128,0.0028008888992998335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,256,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,1024,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,1024,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,7168,0.043881777260038585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,512,0.007424888511498769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,3072,0.019164444671736825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,2560,0.007387555307812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,10240,0.15875200430552164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,10240,0.20685244931115043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,512,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,10240,0.025403555896547105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,512,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,512,0.047227554851108126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,5120,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,5120,0.13908178276485866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,7168,0.013952889376216464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,1024,0.016885333591037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,32,0.0023128889087173673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,8192,0.03583377930853102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,768,0.017071111334694754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,8192,0.1768266624874539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,512,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,51200,0.425839106241862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,256,0.0035751110149754416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,65536,0.06001600291993883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,256,0.013435555828942193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,51200,0.015970667203267414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,51200,0.09763999780019124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,3072,0.11488710509406196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,3584,1.0439821879069011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,6144,0.041511999236212835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,51200,0.10945510864257812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,2560,0.03751111030578613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,10240,0.035353776481416486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,16384,2.4641865624321833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,3584,0.02093422247303857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,2048,0.006734222173690796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,5120,0.05585600270165337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,12288,0.3583982255723741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,1536,0.31631377008226186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,5120,0.047094222572114736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,2560,0.036320888333850436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,7168,0.03149066699875726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,768,0.0031724443866146933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,2560,0.01513155632548862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,1024,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,128,0.0023066666391160754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,1536,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,256,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,32,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,1536,0.018944000204404194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,2048,0.011333333121405708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,2560,0.03757777810096741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,6144,1.7414995829264324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,1536,0.03313510947757297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,3584,0.06583022409015231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,5120,0.09584799740049575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,5120,0.016231111354298063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,256,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,3072,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,2048,0.011239999698268043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,51200,0.07319288783603244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,2560,0.04645955562591553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,8192,0.15601244237687853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,3072,0.008847999903890822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,32,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,5120,0.02757511039574941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,3072,0.5927546819051107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,6144,0.03787111242612203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,3072,0.07851466867658827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,65536,0.26354577806260854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,1024,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,3584,0.04302311274740431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,6144,0.8748835457695855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,2560,0.02802844511138068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,768,0.003834666477309333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,5120,0.024324445260895625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,2048,0.007828444242477417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,768,0.11157066292232937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,1024,0.016910221841600206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,16384,0.09464977847205268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,7168,0.017987555927700467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,16384,0.47683021757337785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,10240,1.4424880345662434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,16384,0.025455999705526564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,5120,0.46515289942423504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,3072,0.02311733365058899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,1536,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,7168,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,768,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,768,0.017176000608338248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,4096,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,2560,0.007044444481531779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,6144,0.042430222034454346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,3584,0.020947555700937908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,65536,0.11868622567918564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,3584,0.014734221829308404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,65536,0.528455999162462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,1024,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,4096,0.07078311178419325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,128,0.011692444483439127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,7168,0.06597777869966295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,6144,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,256,0.017557332913080852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,2560,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,2048,0.012323555847009024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,512,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,2560,0.08664800061119927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,12288,0.6987137794494629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,10240,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,16384,0.2983448770311144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,51200,0.02438666588730282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,10240,0.03059200114674038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,65536,0.19511022832658556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,3072,0.04765511221355862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,5120,0.03720000055101182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,12288,0.013737777868906656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,64,0.0041786668201287585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,7168,0.10063555505540635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,128,0.006421333385838403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,12288,0.033081776565975614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,12288,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,7168,0.01862844493654039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,128,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,12288,0.030033777157465618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,10240,0.030877331892649334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,7168,0.17025332980685762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,2048,0.010668444136778513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,65536,0.15791200266944036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,2048,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,6144,0.030007110701666936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,3072,0.019819556011093985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,1536,0.008744888835483128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,128,0.014905777242448596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,16384,0.02883022361331516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,3584,0.10428088903427124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,2048,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,10240,0.025742222865422566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,51200,0.10325600041283502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,5120,0.09202577670415242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,51200,0.4506186909145779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,512,0.01514488955338796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,16384,0.03843377696143256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,64,0.1517439948187934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,4096,0.020184889435768127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,2048,0.01236088905069563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,7168,0.01348444405529234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,1536,0.017398221625222098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,64,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,6144,0.04792355497678121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,10240,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,4096,0.6878106858995227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,3584,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,3584,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,2048,0.0808488859070672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,64,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,256,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,1536,0.019463999403847587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,256,0.010958222051461538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,768,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,3584,0.016269332832760282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,7168,0.0614195532268948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,2560,0.023660444551044043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,4096,0.1180773311191135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,6144,0.01963555481698778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,3584,0.012019555601808759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,8192,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,2048,0.01620533400111728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,128,0.02607822252644433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,4096,0.07754222551981609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,6144,0.00963199966483646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,256,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,4096,0.0042444442709287005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,7168,0.016571554872724745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,7168,0.023843554986847773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,128,0.015538665983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,5120,0.027256000373098586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,1536,0.008048888709810045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,4096,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,16384,0.02160444524553087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,51200,0.0495395561059316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,2048,0.025796444879637823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,1024,0.019135110908084445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,4096,0.018187556001875136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,4096,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,8192,0.02397511071628994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,1536,0.20469510555267334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,3584,0.014699555105633207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,768,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,7168,0.02346399923165639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,51200,0.17317510975731742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,512,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,8192,0.016318221886952717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,7168,0.01977777812216017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,16384,0.5508755577935113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,16384,0.028034667174021404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,1024,0.016089777151743572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,512,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,2560,0.033313777711656355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,8192,0.03972444600529141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,4096,0.09833421972062852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,65536,0.07689155472649468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,3072,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,12288,3.186936060587565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,768,0.08043822315004137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,16384,0.07992621925142077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,10240,0.013924444715181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,128,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,12288,0.029059555795457628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,256,0.02739288906256358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,16384,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,3072,0.03165244393878513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,1536,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,51200,0.0990648865699768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,8192,0.5225262112087673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,768,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,2048,0.034338666333092585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,5120,0.011788444386588203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,2560,0.03007733490731981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,51200,0.05168266759978401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,10240,0.028303109937243994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,4096,0.015798222687509324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,51200,0.08297955327563815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,4096,0.020733333296246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,6144,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,128,0.014071111877759298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,3584,0.027109333210521277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,5120,0.049399998452928334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,3584,0.07344977723227607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,6144,0.02038666605949402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,10240,0.053039110369152494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,256,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,1536,0.017269333203633625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,64,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,1536,0.37844710879855686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,256,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,51200,0.05095555716090732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,2048,0.03788266579310099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,8192,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,128,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,2048,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,3072,0.05766310956743029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,7168,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,3072,0.057405332724253334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,1024,0.03349688980314467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,2560,0.02938133478164673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,7168,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,5120,0.009905777871608734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,1536,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,10240,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,16384,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,512,0.00254133323000537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,256,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,3072,0.011476444701353708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,16384,0.9558213551839193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,12288,0.029799110359615747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,128,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,16384,0.042005333635542125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,768,0.01402844488620758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,3584,0.03459999958674113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,768,0.0174666659699546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,2048,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,256,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,3072,0.11529689364963108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,10240,0.008098666866620382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,4096,0.7021279864841037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,3584,0.01369866645998425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,256,0.019279999865425956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,8192,0.026428444517983332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,3584,0.023052444060643513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,2560,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,128,0.008946666287051307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,7168,0.06266844272613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,6144,0.1437902185651991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,5120,0.008721777962313758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,128,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,512,0.2123511102464464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,10240,0.033682667546802096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,6144,0.024458666642506916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,8192,0.02498044404718611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,3072,0.025086222423447505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,4096,0.014411555396185981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,51200,0.061907556321885854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,1024,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,512,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,6144,0.03247288862864176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,512,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,1536,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,32,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,4096,0.017676444517241586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,7168,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,768,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,128,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,1024,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,12288,0.09779555267757839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,6144,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,8192,0.24055287573072645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,64,0.010878221856223213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,7168,0.1689120001263089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,6144,0.033357332150141396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,65536,0.13206044832865396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,16384,0.19811199771033394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,6144,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,128,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,2048,0.02000799940692054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,10240,0.046022223101721875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,16384,0.3128711117638482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,256,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,512,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,128,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,51200,0.04933511217435201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,32,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,51200,0.06310311290952894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,3584,0.016519111063745286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,2560,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,8192,0.020502222908867728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,64,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,2560,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,2048,0.05045333173539904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,32,0.006200888918505774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,7168,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,65536,2.763429429796007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,51200,1.01218753390842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,12288,0.02739022175470988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,8192,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,2560,0.009152889251708984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,10240,0.04250311189227634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,2048,0.015443555182880826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,2560,0.08030488755967882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,65536,1.08078490363227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,6144,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,7168,0.2265991104973687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,10240,0.03176622258292304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,4096,0.009983110758993361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,6144,0.053256889184316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,256,0.007003555695215861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,10240,0.35450487666659886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,8192,0.016864889197879367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,16384,0.3363795545366075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,768,0.015132443772421943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,12288,0.07700444592369927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,7168,0.016238222519556682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,3072,0.00869333330127928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,10240,0.7614915635850695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,768,0.029751999510659113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,51200,0.052891555759641856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,10240,0.016851555969980028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,65536,0.4461858007642958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,3584,0.6210444238450792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,7168,0.041640000210867986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,3072,0.04415466719203525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,51200,0.5171768930223253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,65536,0.9712444941202799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,1536,0.06492888927459717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,2560,0.02163822286658817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,64,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,768,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,3584,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,512,0.008362666600280339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,3072,0.013103110922707452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,5120,0.29166489177280003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,65536,0.9003511004977756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,10240,0.1380231115553114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,7168,0.020628444022602506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,51200,0.35306310653686523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,4096,0.021931555536058214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,2560,0.01703466640578376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,4096,0.13283911016252306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,7168,0.021182222498787776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,512,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,5120,0.20440533426072863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,12288,0.01872977779971229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,16384,0.24695910347832573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,16384,0.05177422364552816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,6144,0.019780443774329293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,5120,0.02769511110252804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,32,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,768,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,16384,0.3620906670888265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,768,0.020967110991477966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,7168,0.5929457876417372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,256,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,7168,0.013855111267831592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,10240,0.04519199993875292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,256,0.01650399963061015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,64,0.0021031111892726687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,7168,0.06143822272618612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,1536,0.34749600622389054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,1024,0.029353777567545574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,5120,0.013478221992651621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,8192,0.02507200009293026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,5120,0.08445600006315444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,2560,0.029634667767418757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,768,0.003162666741344664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,65536,0.18210666709476045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,1024,0.018787556224399142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,65536,0.553326235877143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,1536,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,128,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,1536,0.02493422230084737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,1024,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,256,0.17034399509429932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,7168,0.033025778002209134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,8192,0.04160177707672119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,12288,0.057943112320370145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,768,0.01752000053723653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,256,0.012445333103338877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,1536,0.0511813329325782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,6144,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,8192,0.024254222710927326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,16384,0.018512000640233357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,51200,0.8450782034132215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,768,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,65536,0.1662666665183173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,512,0.14846755398644343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,32,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,51200,1.814616945054796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,65536,4.3690266079372835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,3072,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,6144,0.07095644209120008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,51200,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,1536,0.007081777685218387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,128,0.0019031111150979996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,4096,0.04537599947717455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,2048,0.004693333473470476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,10240,0.01793511046303643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,16384,0.08098666535483466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,6144,0.1016097797287835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,5120,0.34540356530083555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,2560,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,7168,0.11672622627682155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,768,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,32,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,7168,0.052008890443378024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,12288,5.009876251220703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,3072,0.1256844467586941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,65536,5.196806165907118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,51200,10.609331766764322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,8192,0.07014755407969157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,128,0.002565333412753211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,4096,0.030392888519499037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,768,0.0325217776828342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,1536,0.013677332964208392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,65536,0.09692177507612441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,6144,0.13274755742814806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,1536,0.047472887569003634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,4096,0.17636888557010225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,10240,0.09110755390591091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,16384,0.26831645435757107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,51200,0.04221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,1536,0.008519111408127679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,12288,0.02924799919128418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,128,0.006694222076071634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,768,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,5120,0.05714755588107639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,16384,0.5183262295193142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,1024,0.016970665918456185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,128,0.01888088881969452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,2560,0.01721333298418257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,10240,0.025156444973415796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,12288,0.02104622291194068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,2560,0.025443555580245122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,256,0.04071022073427836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,256,0.005656888915432825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,6144,0.01428533262676663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,2048,0.01960711181163788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,768,0.024703999360402424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,3072,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,256,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,51200,1.0532675849066842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,6144,0.13766222529941136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,65536,0.29674310154385036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,6144,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,5120,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,3584,0.04510577850871616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,256,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,512,0.013735999663670858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,12288,0.27170666058858234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,6144,0.08151288827260335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,32,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,65536,55.130842420789925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,512,0.006335111127959357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,256,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,64,0.019356444478034973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,2560,0.004119111018048393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,128,0.05579288800557455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,8192,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,768,0.07869955566194323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,768,0.029136889510684546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,1536,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,512,0.010081777969996134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,1024,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,10240,0.11510133743286133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,128,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,6144,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,8192,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,2560,0.025370667378107708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,768,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,512,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,2048,0.01717511150572035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,128,0.004942222187916438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,6144,0.016062221593326993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,2048,0.21807732847001818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,64,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,3072,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,1024,0.004135999828577042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,16384,0.03985155622164408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,2048,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,2048,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,7168,0.2325830989413791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,32,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,65536,0.45798312293158633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,3584,0.02055022286044227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,3584,0.08188977506425646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,256,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,3584,0.013250666360060373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,10240,0.027979556057188246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,5120,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,512,0.012785777449607849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,16384,0.026956443985303242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,3072,0.012675555215941535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,2048,0.02265600032276577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,1024,0.0765848888291253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,8192,0.13418755266401503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,10240,0.08346577485402425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,64,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,10240,0.016965332958433364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,768,0.029655999607510034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,7168,0.04563822348912557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,4096,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,4096,0.018244443668259513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,1024,0.0868897769186232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,10240,0.10772266652848984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,512,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,10240,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,7168,0.1377262274424235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,65536,1.178727149963379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,16384,0.770634651184082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,256,0.15603200594584146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,7168,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,2560,0.01366666704416275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,5120,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,5120,0.017735110388861764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,8192,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,2048,0.006792000184456508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,65536,3.243541293674045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,65536,0.14952088726891413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,5120,0.04184622234768338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,512,0.010764444039927589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,16384,0.023390221926901076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,4096,0.011158222125636207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,3584,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,65536,0.15358044041527644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,2560,0.010070222119490305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,4096,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,64,0.04081599911053976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,2048,0.007099555598364935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,8192,0.22267466121249727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,51200,0.010825778047243753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,2560,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,7168,0.024337778488794964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,128,0.04327199856440226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,256,0.005530666559934616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,16384,0.01683911184469859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,12288,0.27729066212972003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,51200,0.09920889139175415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,51200,0.06368711259629992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,51200,0.187553776635064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,3072,0.026686223016844854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,1536,0.03080444534619649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,256,0.010747555229398938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,7168,0.032539556423823036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,6144,0.011967110964987012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,16384,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,10240,0.04405600163671705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,10240,0.267922666337755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,5120,0.00906755526860555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,1536,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,10240,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,32,0.09391822417577107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,2048,0.004903111192915174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,3584,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,2048,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,1536,0.0373582210805681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,256,0.005377777748637729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,3072,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,12288,2.1411192152235246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,256,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,2560,0.02846933404604594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,10240,0.043219556411107383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,12288,0.11896533436245388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,51200,0.12782577673594156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,32,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,4096,0.00814933329820633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,512,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,3584,0.011508444117175208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,256,0.010079111489984725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,768,0.003989333493842019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,10240,0.08511377705468072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,768,0.06287288665771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,5120,0.7400133344862195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,1024,0.025677333275477093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,4096,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,512,0.0031386667655573953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,2560,0.00703022215101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,256,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,3584,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,256,0.006843555718660355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,12288,0.06753333409627278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,6144,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,3072,0.04014933440420363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,256,0.013468444347381592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,7168,0.02167111138502757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,2560,0.005190222213665645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,3072,0.06679377953211467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,10240,0.06995111041598849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,16384,0.023507555325826008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,65536,0.0867004460758633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,768,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,5120,0.10291022062301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,128,0.007096889118353526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,16384,0.0605564448568556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,2560,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,16384,0.4981413417392307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,1024,0.06002933449215359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,256,0.05128444565667046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,128,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,5120,0.01924088928434584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,5120,0.025794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,6144,0.029634667767418757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,1536,0.006372444331645966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,512,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,8192,0.06743022468354967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,7168,0.18059999412960479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,768,0.01497066683239407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,512,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,3072,0.007583111524581909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,16384,0.13365066051483154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,5120,0.013175999952687157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,2048,0.025172443853484258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,32,0.007054222126801808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,51200,0.7230230967203776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,5120,0.04548799991607666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,768,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,16384,0.01186044431395001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,51200,0.5859466658698188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,51200,0.1271804438696967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,32,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,512,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,16384,0.03276177909639146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,51200,0.10719111230638291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,6144,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,32,0.002581333327624533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,7168,0.03436355458365546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,5120,0.01835466590192583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,51200,3.351023144192166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,16384,0.039290666580200195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,32,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,32,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,16384,0.023750222391552393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,2048,0.068904889954461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,6144,0.11045866542392307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,768,0.004271999829345279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,51200,0.1166986624399821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,3072,0.02901155418819851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,512,0.00756444450881746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,4096,0.04415111078156365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,6144,0.014751111467679342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,768,0.019726221760114033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,1024,0.16911289427015516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,6144,0.055918223328060575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,1024,0.0032186667538351486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,768,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,512,0.028763555818133887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,3584,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,3072,0.027406222290462915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,256,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,12288,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,12288,0.09271022346284653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,4096,0.018626666731304593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,128,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,65536,2.9330372280544705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,10240,0.21089688936869302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,768,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,2560,0.030807998445298936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,256,0.015453333655993143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,512,0.2981911235385471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,3072,0.10429510805341934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,1536,0.007354666789372762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,2560,0.009775111244784461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,256,0.03445422318246629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,7168,0.04111466805140177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,5120,0.03497777713669671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,8192,0.025677333275477093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,256,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,2560,0.011400000088744693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,256,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,7168,0.021332444416152105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,512,0.02237777743074629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,3072,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,10240,0.05473777651786804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,8192,0.12278666761186387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,1536,0.005165333135260476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,4096,0.11065866549809773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,128,0.021912000245518152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,128,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,10240,0.18454489443037245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,5120,0.09106844663619995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,7168,0.10480444298850165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,768,0.02590488890806834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,256,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,2560,0.014908444550302295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,16384,0.021425777011447485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,51200,0.15615999698638916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,64,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,256,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,1536,0.004009777886999978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,512,0.003960888832807541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,7168,0.009977777798970541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,512,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,6144,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,32,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,2560,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,1024,0.0138586668504609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,8192,0.14330044057634142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,10240,0.027797334724002417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,256,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,256,0.06693689028422038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,2560,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,3072,0.015308444698651632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,16384,0.01759288873937395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,768,0.01791200041770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,10240,0.2510302331712511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,8192,0.020908445119857788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,10240,0.03831555445988973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,16384,0.021316443880399067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,12288,0.14069867134094238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,16384,0.06378755304548475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,6144,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,1024,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,128,0.031331555710898504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,51200,0.05053244365586174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,8192,0.052721778551737465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,65536,2.5862017737494574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,2048,0.20111821757422554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,768,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,4096,0.08624444405237834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,16384,0.019169777631759644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,10240,0.8891244464450412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,1024,0.2798408932156033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,256,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,1024,0.008951110972298516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,256,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,65536,0.6381013128492568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,3584,0.031326221095191106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,4096,0.05363022287686666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,2048,0.02091288897726271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,4096,0.024869332710901897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,3584,0.023297778434223596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,4096,0.034958223501841225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,2560,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,3072,0.0075795559419526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,8192,0.1646088891559177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,3072,0.02195377813445197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,10240,0.024706666668256123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,16384,0.28561509980095756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,8192,0.0807573331726922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,256,0.11578488349914551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,6144,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,10240,0.22065511014726427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,4096,0.018555555078718398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,65536,0.06477244695027669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,6144,0.019647111495335896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,768,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,128,0.010430222584141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,3072,0.034394668208228216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,3072,0.026547556122144062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,1024,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,768,0.003370666669474708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,65536,0.2811768849690755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,65536,0.5474150975545248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,1536,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,128,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,4096,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,768,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,10240,0.24987732039557564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,128,0.06343111064698961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,12288,0.07283555799060397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,768,0.01824977828396691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,12288,0.07435999976264106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,5120,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,256,0.005366222312053044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,3072,0.01256444470749961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,3584,0.01791822248035007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,8192,0.401325331793891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,3584,0.04297866755061679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,512,0.04288977715704176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,16384,0.11098221937815349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,128,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,2560,0.028769777880774602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,16384,0.204239116774665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,2048,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,256,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,128,0.016287111573749118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,2048,0.01275111072593265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,32,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,1024,0.23595733112759062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,8192,0.033527112669414945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,512,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,256,0.07150844732920329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,6144,0.06850933366351657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,2048,0.06644178099102445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,128,0.008264888491895463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,7168,0.31476354598999023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,3584,0.004931555440028508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,8192,0.08126044273376465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,2560,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,2048,0.031143999762005273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,10240,0.07110399670071073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,256,0.008683555656009251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,2560,0.06092088752322727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,128,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,3072,0.026733333865801494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,5120,0.05341866612434387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,1024,0.02311733365058899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,128,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,768,0.033519999848471746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,3072,0.009034666750166152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,256,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,3584,0.02214488883813222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,512,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,65536,0.6468995412190756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,256,0.026516444153255884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,7168,0.01180266671710544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,64,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,4096,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,12288,0.48687288496229386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,10240,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,12288,0.029527998632854883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,2048,0.05320800013012356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,1536,0.015187554889255099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,5120,0.025240000751283433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,7168,0.030453334252039593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,6144,0.02488533324665493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,4096,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,128,0.01718933383623759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,7168,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,768,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,7168,0.04388088981310526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,16384,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,16384,0.1281528870264689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,1536,0.011648889217111798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,3072,0.6195297771030003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,256,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,3072,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,16384,0.07606755362616645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,8192,0.9287137985229492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,65536,1.7552728652954102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,65536,0.8921200434366862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,8192,0.177839994430542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,6144,0.1630177762773302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,65536,0.06782666842142741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,1024,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,32,0.01882844501071506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,8192,0.018920888503392536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,512,0.15029599931504992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,12288,0.02260355485810174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,1536,0.045143110884560474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,65536,0.16582311524285212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,256,0.03458755546145969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,12288,0.4400453302595351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,12288,0.019541333119074505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,3072,0.015556444724400839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,1536,0.07487644751866658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,512,0.061499555905659996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,1024,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,51200,0.0688471131854587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,5120,0.01809599995613098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,128,0.0277137772904502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,2560,0.01533511115445031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,256,0.01953066719902886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,8192,0.0513315565056271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,51200,1.1674044926961262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,128,0.10168000062306722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,128,0.050422221422195435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,12288,0.22606311904059517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,2048,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,128,0.021249777740902368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,5120,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,768,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,2048,0.020404444800482858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,1024,0.030798223283555772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,10240,0.02276888820860121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,768,0.07496622535917494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,5120,0.07185333304935031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,51200,0.07816000117195977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,2048,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,10240,0.17046666145324707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,6144,0.04467911190456814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,4096,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,768,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,8192,0.018616000811258953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,2048,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,3072,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,16384,2.2990088992648654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,2048,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,16384,0.04225422276390923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,2560,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,12288,0.07260000043445163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,128,0.012019555601808759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,16384,0.07278311252593994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,5120,0.015544000599119397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,3072,0.025988444685935974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,1024,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,512,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,2560,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,8192,20.501497056749134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,12288,0.06295466423034668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,64,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,1024,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,3584,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,256,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,1536,0.01982133256064521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,3072,0.25640800264146596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,16384,0.04631466666857401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,768,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,3584,0.048545777797698975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,3584,0.06675644715627034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,16384,0.06784355640411377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,7168,0.1751928859286838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,51200,0.08390311400095622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,10240,0.01590133375591702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,128,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,1024,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,1024,0.26693688498602974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,5120,0.15914666652679443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,3072,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,768,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,2048,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,768,0.025957332717047796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,5120,0.02698488864633772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,8192,0.04514666729503208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,51200,0.12643378310733372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,51200,0.03977333174811469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,512,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,65536,1.8592427571614583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,7168,0.03467199868626065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,10240,0.05198311143451267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,512,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,6144,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,3072,0.008045333127180735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,16384,0.1773351165983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,2048,0.041195554865731135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,8192,0.7567937639024523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,1536,0.010890666809346942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,256,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,8192,0.19588088989257812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,1024,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,256,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,6144,0.008902221918106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,3584,0.43726666768391925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,2560,0.09485689136717056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,7168,0.293471097946167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,5120,0.028742220666673448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,51200,0.05393155415852865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,4096,0.028563555743959215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,51200,0.037823110818862915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,10240,0.5904666582743326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,1024,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,4096,0.04611555404133267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,128,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,1536,0.022679110368092854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,512,0.09036178059048122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,1024,0.03407377666897244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,256,0.005231110999981563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,512,0.003863999827040566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,51200,0.08107822471194796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,1024,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,7168,0.031162665949927434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,12288,0.03495999839570787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,7168,0.07067733340793185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,51200,0.05381955703099569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,3072,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,2560,0.008731555607583787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,2048,0.056906667020585805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,2560,0.008008889026112026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,4096,0.027279110418425664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,768,0.044554667340384595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,2048,0.020651555723614164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,1536,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,256,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,16384,0.852060423956977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,16384,0.05297155512703789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,7168,0.40861066182454425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,12288,0.0348364445898268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,7168,0.2067742215262519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,65536,2.0935236612955728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,16384,0.019759999381171334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,4096,0.11721244123246934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,16384,0.05122311247719658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,128,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,16384,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,3072,0.018998222218619455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,128,0.019303111566437613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,16384,2.7857822842068143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,512,0.02125866711139679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,8192,0.02302933401531643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,4096,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,1024,0.004312889029582341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,2560,0.06406133042441474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,768,0.02104266650146908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,512,0.002516444358560774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,4096,0.017875555488798354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,65536,0.03415288858943515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,1536,0.017470222380426195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,5120,0.009242666264375051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,3072,0.007288888924651676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,65536,2.1242533789740667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,51200,0.37603288226657444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,4096,0.025026667449209426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,6144,0.012111110819710625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,10240,0.03161688976817661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,3584,0.07625422212812635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,6144,0.007779555188284979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,256,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,4096,0.09065066443549262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,12288,0.043136000633239746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,512,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,65536,0.1504595544603136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,2048,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,5120,0.16270489162868923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,51200,0.04010044535001119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,2048,0.08245422442754109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,8192,0.03758044375313653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,65536,1.279985745747884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,2560,0.02591200007332696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,4096,0.020629333125220407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,5120,0.08387111292945014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,3584,0.09570044279098511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,128,0.0222382214334276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,65536,0.44584978951348203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,16384,0.0063271112740039825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,5120,0.06688621971342298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,256,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,128,0.002576888849337896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,6144,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,256,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,65536,0.9141048855251737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,1536,0.06128799915313721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,10240,0.011777777638700275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,1024,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,1024,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,4096,0.00851822230550978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,512,0.036357333262761436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,768,0.015419556034935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,512,0.02171733313136631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,65536,0.9911528693305122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,128,0.011510222322411008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,512,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,128,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,128,0.015673778123325773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,2560,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,8192,0.02196177840232849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,1024,0.04504266712400648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,1024,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,768,0.006080889039569431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,2048,0.032871110571755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,65536,0.1024071110619439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,256,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,1536,0.2435795466105143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,7168,0.019873778025309246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,8192,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,256,0.0022026666750510535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,12288,0.01773422294192844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,3072,0.02989333205752903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,10240,0.24901866912841797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,3072,0.03494488861825731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,512,0.013227555486891003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,1536,0.02158755560715993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,768,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,3584,0.009734222458468543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,256,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,768,0.08484177456961738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,2048,0.013011555704805585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,12288,0.7317822244432238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,128,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,5120,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,4096,0.034544888469907976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,51200,0.05231377813551161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,1536,0.15349422560797796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,4096,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,2560,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,1024,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,2048,0.025065778030289546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,8192,0.020630222227838304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,7168,0.0466542210843828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,3072,0.007361777954631382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,12288,0.04256177941958109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,3584,0.10169421964221531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,2560,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,5120,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,2560,0.357114659415351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,768,0.0028826666788922418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,10240,0.08236355251736112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,7168,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,4096,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,51200,0.15674755308363172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,65536,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,3584,0.017940445078743827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,4096,0.017587555779351127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,12288,0.14184355735778809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,8192,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,12288,0.017263111140992906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,6144,0.03212177753448486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,7168,0.06311200062433879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,5120,0.19177688492668998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,16384,0.04385422335730659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,64,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,4096,0.02316444449954563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,7168,0.02090222140153249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,1536,0.020638222495714825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,6144,0.07162311342027453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,7168,0.06566666894488864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,1024,0.015608888533380298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,3072,0.013682666752073499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,65536,0.07720088958740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,256,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,32,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,2048,0.0439955559041765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,5120,0.045591112640168935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,768,0.0459591117170122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,32,0.0016071111377742554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,5120,0.5379048983256022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,2560,0.013756444056828817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,4096,0.3779315683576796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,4096,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,65536,0.3551786740620931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,1536,0.0041751112374994485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,7168,0.16735022597842744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,5120,0.022151110900772944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,3072,0.013928888572586907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,51200,0.05141333407825894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,3072,0.024854222933451336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,128,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,65536,0.1652124457889133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,128,0.014935111006100973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,12288,0.011895111037625207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,1536,0.013175999952687157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,256,0.021635555558734473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,1536,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,7168,0.047694222794638745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,32,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,64,0.004914666629499859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,32,0.03921066721280416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,10240,0.014873777826627096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,8192,0.14055200417836508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,65536,0.3720133304595947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,512,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,3072,0.020755555894639757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,1536,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,7168,0.04165600074662103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,3584,0.007089777953094906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,512,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,12288,0.04712177647484673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,2560,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,12288,0.9013315836588541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,2048,0.02585244509908888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,256,0.11570844385359023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,512,0.011791999969217511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,2560,0.05781600210401747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,16384,0.04569244384765625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,3584,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,3072,0.01645777788427141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,256,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,32,0.0033359999457995095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,2560,0.004970666848950916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,32,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,1024,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,128,0.023337778117921617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,3072,0.05704444646835327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,4096,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,32,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,32,0.001790222194459703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,7168,0.015156444576051501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,768,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,3072,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,2560,0.04580710993872749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,2048,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,768,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,12288,0.3820986747741699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,3072,0.06805600060356988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,128,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,128,0.049548443820741445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,51200,0.05074933171272278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,1024,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,32,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,64,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,3584,0.020972443951500788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,51200,0.3549182150099013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,4096,0.022506665852334764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,1536,0.018978666928079393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,51200,0.04725066820780436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,2560,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,512,0.02533066603872511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,3584,0.0799271133210924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,2048,0.022435555855433147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,256,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,8192,0.020121778051058452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,2560,0.029778665966457788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,3072,0.013293333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,1536,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,768,0.006031111296680238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,65536,0.45270845625135636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,8192,0.12803644604153103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,51200,0.272037320666843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,65536,0.060645335250430636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,128,0.014783999986118741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,6144,0.017436444759368896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,4096,0.028260442945692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,512,0.013342222405804528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,3072,0.01854044364558326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,1536,0.021838222940762837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,768,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,51200,0.05172888769043816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,4096,0.028145776854621038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,128,0.005967999911970562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,512,0.025010666913456384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,5120,0.009594666461149851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,10240,0.03460799985461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,3072,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,16384,0.04930222365591261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,32,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,12288,0.012186666329701742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,12288,0.12848711013793945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,1024,0.015452444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,65536,0.0971484449174669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,10240,0.028359111812379625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,7168,0.009975999593734741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,5120,0.003930666794379552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,3584,0.32659822040134007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,5120,1.3086168501112196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,2560,0.019536000159051683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,8192,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,8192,0.021371554997232225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,12288,0.12975911299387613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,768,0.012453333371215396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,4096,0.12001155482398139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,51200,0.3797902266184489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,65536,0.3378711011674669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,32,0.0783733328183492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,1536,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,5120,0.0336737765206231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,128,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,2048,0.015632889337009855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,1024,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,1024,0.10371110836664836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,4096,0.008055999875068665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,512,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,2560,0.08660622437795003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,1024,0.061920887894100614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,51200,4.96720462375217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,2048,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,768,0.007429333196745978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,512,0.02273599969016181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,6144,0.014162666267818876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,2048,0.033851553996404014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,512,0.022595556245909795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,6144,0.2696177694532606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,16384,0.02625333269437154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,65536,0.0383857786655426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,10240,0.06418399678336249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,8192,0.024599111742443506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,10240,0.01053244454993142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,512,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,2048,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,768,0.058911111619737416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,768,0.020971554848882887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,1024,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,65536,0.3828302224477132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,32,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,6144,0.03680088784959581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,3584,0.03840177920129564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,1536,0.22029422389136422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,2048,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,51200,0.16869778103298613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,3072,0.5951315561930338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,10240,0.09212177991867065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,768,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,32,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,5120,0.14279199971093073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,1024,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,2048,0.024264888630972967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,4096,0.02994577752219306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,4096,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,3072,0.2404728995429145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,5120,0.08662222491370307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,4096,0.03262844350602891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,512,0.012110222544935016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,8192,0.09956533379024929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,512,0.027217777238951787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,3072,0.017487110363112558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,51200,0.012697777814335294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,10240,0.17370666397942436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,3072,0.06889777713351779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,8192,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,128,0.014877332581414117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,768,0.023707555400000677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,1536,0.018698667486508686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,8192,0.019546666079097323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,2048,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,7168,0.7140524652269152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,256,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,2048,0.015285332997639975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,12288,0.08563911252551609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,512,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,3584,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,2560,0.011307555768224927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,768,0.16323288281758627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,2048,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,256,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,10240,0.012103111379676394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,3584,0.09761066569222344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,2560,0.04508089025815328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,2048,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,256,0.005443555613358815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,16384,0.016198222835858662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,128,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,32,0.014716444744004143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,65536,0.05605600277582804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,1536,0.020921778347757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,768,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,512,0.004223111189073986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,256,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,5120,0.1489200062221951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,256,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,51200,4.915802849663629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,512,0.007330666813585494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,65536,0.9564826753404406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,7168,0.04622755448023478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,5120,0.05961689021852282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,1024,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,512,0.013855111267831592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,768,0.03045066528850132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,2560,0.06511733267042372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,51200,10.985318501790365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,6144,0.047008888589011304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,512,0.4484400219387478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,256,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,4096,0.012994666894276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,1024,0.00738666703303655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,51200,0.16673066880967882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,7168,0.03432622220781114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,3072,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,2048,0.15829777717590332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,3584,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,64,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,5120,0.047114666965272695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,768,0.029983109898037378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,65536,0.15939822461869982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,16384,0.0912160012457106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,256,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,2560,0.3070400026109484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,51200,0.60971466700236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,16384,0.04496888981925117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,4096,0.05124622252252367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,256,0.04555022385385302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,128,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,6144,0.2579760021633572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,51200,0.12388088968065049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,10240,0.16881867249806723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,256,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,512,0.012048000262843238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,256,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,6144,0.30206667052374947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,4096,0.04085066583421495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,32,0.014733332726690503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,2048,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,2560,0.047543111774656505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,3584,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,7168,0.1333448886871338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,1024,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,2048,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,3072,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,2560,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,5120,0.028749333487616643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,7168,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,3072,0.10057422187593247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,51200,0.12016977204216851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,512,0.025066667132907446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,768,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,8192,0.014516444669829475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,4096,0.01737422247727712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,8192,0.06729688909318712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,128,0.015536889433860779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,256,0.005517333332035277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,512,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,5120,0.10665333271026611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,65536,0.3820151223076715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,5120,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,51200,0.05453777644369337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,16384,0.5592968728807237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,12288,0.13106221622890896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,4096,0.08614933490753174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,6144,0.01308444473478529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,16384,0.01702844434314304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,65536,0.1709822283850776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,64,0.002459555657373534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,128,0.005231110999981563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,128,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,2560,0.02872177627351549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,3072,0.10519288645850287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,6144,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,2048,0.18645689222547743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,2048,0.7556542290581597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,32,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,64,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,512,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,10240,0.13434399498833552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,51200,0.07121600045098199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,1024,0.013423110875818463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,2560,0.3096302350362142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,6144,0.07835200097825792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,2560,0.3469742139180501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,256,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,65536,0.5452017784118652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,12288,0.22549690140618217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,2048,0.04246222310596042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,51200,0.12164266904195149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,1024,0.0028844444702068963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,5120,0.04893689023123848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,10240,0.016935111747847665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,3072,0.0372284452120463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,12288,0.03887910975350274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,128,0.020450666546821594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,64,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,8192,0.010647111468844943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,12288,0.04002844293912252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,3072,0.02910755409134759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,3072,0.045574221346113414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,6144,0.042465776205062866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,128,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,7168,0.28665955861409503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,65536,0.27331911192999947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,10240,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,3072,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,1536,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,256,0.02530044482813941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,12288,0.02328799996111128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,2560,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,1024,0.015218666858143277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,7168,0.2593173450893826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,5120,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,3072,0.007777777810891469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,3584,0.007818666597207388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,1536,0.02130488885773553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,8192,0.02425777746571435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,32,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,10240,0.012039110892348819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,1536,0.00702044450574451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,32,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,128,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,10240,0.03856089048915439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,128,0.08202311065461901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,3072,0.0070257774657673305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,8192,0.024308444725142583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,4096,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,128,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,512,0.23771378729078504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,3584,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,64,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,51200,0.09581066502465142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,6144,0.02106399999724494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,128,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,128,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,7168,0.009673777553770278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,2560,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,64,0.022357333037588332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,1024,0.01721600029203627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,256,0.030204445123672485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,8192,0.09553777509265476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,10240,0.020953777763578627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,7168,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,3584,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,128,0.015804444750150044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,6144,0.08710310856501262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,8192,0.0351582235760159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,10240,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,512,0.023237332701683044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,16384,0.43838755289713544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,3072,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,65536,0.7032133208380805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,32,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,256,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,7168,0.27360354529486763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,65536,0.34967999988132054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,2560,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,12288,0.5413306554158529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,768,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,10240,0.014988443917698331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,4096,0.005006222261322869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,7168,0.013036444783210754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,7168,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,1024,0.03691733214590285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,16384,0.02093422247303857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,4096,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,1024,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,3072,0.05714044306013319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,256,0.04705955584843954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,128,0.016572443975342643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,8192,0.3247182104322645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,768,0.017662222186724346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,6144,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,12288,0.02676888969209459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,32,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,4096,0.03999733262591892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,10240,0.12328088283538818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,768,0.018631110588709515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,2560,0.02666933337847392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,2048,0.03687999977005853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,12288,0.006699555450015598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,128,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,3072,0.008229333493444655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,7168,0.03757333424356248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,12288,0.060976889398362905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,6144,0.014214222629865011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,10240,0.4605289035373264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,12288,0.045349333021375865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,2048,0.04404800136884054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,8192,0.01331022216214074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,4096,0.07854488823148939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,6144,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,6144,0.02238133384121789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,65536,0.07377777496973674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,512,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,768,0.015295111470752291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,32,0.014170666535695394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,768,0.018244443668259513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,128,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,10240,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,1024,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,768,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,1024,0.026132444540659588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,32,0.042789333396487765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,1024,0.01701866587003072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,2560,0.035721778869628906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,4096,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,4096,0.1776640017827352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,8192,0.026153778036435444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,6144,0.02402399977048238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,7168,0.01440622243616316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,2560,0.031143999762005273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,51200,0.0626044472058614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,3072,0.037096887826919556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,6144,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,65536,0.40023467275831437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,2048,0.026604443788528442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,12288,0.03669688767857022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,51200,0.15288710594177246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,1024,0.03067022230890062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,7168,0.05144800080193413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,3072,0.007859555383523306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,7168,0.39328622817993164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,6144,0.02808533443344964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,8192,0.037426667080985174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,2048,0.07260000043445163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,4096,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,10240,0.09045422077178955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,10240,0.020119110743204754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,128,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,65536,0.13513599501715765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,12288,0.03568177753024631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,51200,0.05637066894107395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,4096,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,1536,0.01791466606987847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,3072,0.12822933991750082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,3072,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,8192,0.05195733242564731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,256,0.01647200021478865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,16384,0.03970844546953837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,2048,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,1536,0.10238755411571926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,3072,8.042733086480036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,1536,0.04163733455869887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,12288,0.026159110996458266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,512,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,5120,0.00870400004916721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,1024,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,4096,0.03581600056754218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,1024,0.004950222041871813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,64,0.013415111435784234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,32,0.011162666810883416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,1536,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,8192,0.011183111204041375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,6144,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,768,0.017123555143674214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,12288,0.026683555708991155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,3072,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,5120,0.027758222487237718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,3584,0.015086222026083203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,51200,0.15488799413045248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,8192,0.03494311041302151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,512,0.029380444023344252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,65536,0.19771288500891793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,5120,0.013560000393125745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,3072,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,64,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,4096,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,128,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,51200,0.0717173351181878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,3072,0.02564711040920681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,10240,0.09050133493211533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,3072,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,3072,0.5627591345045302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,512,0.005969777703285217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,64,0.019820445113711886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,1024,0.008893333375453949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,6144,0.06795110967424181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,12288,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,3072,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,65536,0.29996710353427464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,768,0.006127111199829314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,7168,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,12288,0.020788444413079154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,7168,0.05300355619854397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,256,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,64,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,1536,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,51200,0.41564355956183535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,512,0.032885332902272545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,16384,0.9986044565836588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,16384,0.023359111613697473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,2048,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,1536,0.10037422180175781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,3072,0.03495288888613383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,65536,0.04695466823048062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,256,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,3584,0.009448889229032729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,2048,0.029594666428036157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,2560,0.018570666511853535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,4096,0.06928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,8192,0.037984000311957464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,2048,0.0677751104036967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,512,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,256,0.002206222257680363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,3584,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,10240,0.0885724425315857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,65536,2.408224105834961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,4096,0.07929688692092896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,12288,0.08533422152201335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,3072,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,768,0.010988444089889526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,2560,0.4199022187127008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,1536,0.054752889606687755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,3584,0.034585777256223894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,6144,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,6144,0.13319110870361328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,128,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,16384,0.34801332155863446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,10240,1.522798220316569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,128,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,512,0.049357334772745766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,7168,0.021311110920376245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,2048,0.02534755567709605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,12288,0.27104798952738446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,51200,0.06242933538224962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,1536,0.009078222016493479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,512,0.003056888985964987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,128,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,512,0.027094221777386133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,51200,0.28476354810926646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,1536,0.3641680081685384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,1536,0.021349334054523047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,512,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,5120,0.017498667041460674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,12288,0.3301431073082818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,8192,0.010446222292052375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,7168,0.011328000161382886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,7168,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,128,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,65536,0.12477600574493408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,6144,0.15591643916236028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,8192,0.08164178000556098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,4096,0.047452443175845675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,768,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,256,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,128,0.011083555718262991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,2048,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,768,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,3584,0.09877244631449382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,12288,0.1304453346464369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,128,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,4096,0.07454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,128,0.014532445205582513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,2560,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,64,0.0018782222436534034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,1536,0.006350222147173352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,65536,0.6698542171054416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,3072,0.04349244303173489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,64,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,1536,0.01716444392999013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,1024,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,1536,0.028037332826190527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,8192,0.03752800159984165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,51200,0.2335360050201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,1536,0.008432889150248634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,256,0.004927111168702443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,7168,0.14586133427090114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,10240,0.11864977412753636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,5120,0.5789484447903103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,1536,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,8192,0.058915555477142334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,128,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,1024,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,1536,0.010306666294733683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,8192,0.03455288873778449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,512,0.022986667023764715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,1024,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,3584,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,8192,0.01833777791923947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,12288,0.018632888793945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,256,0.027832001447677612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,1024,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,2560,0.11766578091515435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,512,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,2560,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,1536,0.02916355596648322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,65536,0.15309422545962864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,3072,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,64,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,16384,0.03914311197068956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,16384,1.8821324242485895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,2048,0.021555556191338435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,768,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,8192,0.06201955344941881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,3584,0.02017511096265581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,128,0.036700444089041814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,512,0.0377786656220754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,16384,0.05714933077494303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,768,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,1024,0.05249244305822584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,1536,0.020291555258962844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,16384,1.2789875666300456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,16384,0.019319999549123976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,8192,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,16384,0.024334222078323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,3584,0.012963555753231049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,8192,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,768,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,2048,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,512,0.02856711215443081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,16384,0.01650311052799225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,1024,0.022706665926509436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,10240,0.01645066671901279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,7168,0.009473777479595607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,32,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,2048,0.026799110902680293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,51200,0.041273779339260526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,2560,0.28816178109910756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,16384,0.06427288717693753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,3584,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,32,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,12288,0.05005866620275709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,12288,0.03737511237462362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,51200,1.7307199902004664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,12288,0.016390222642156813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,3584,0.021594666772418555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,65536,0.5944266849093968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,3584,0.01849688920709822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,4096,0.010544000400437249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,65536,3.1326240963406033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,7168,0.032572444942262434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,65536,0.5789670944213867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,1536,0.02381244467364417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,2560,0.015388444066047668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,5120,0.031319111585617065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,32,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,51200,2.6496942308213973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,256,0.018745778335465323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,4096,0.011032000184059143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,512,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,10240,0.017521777086787753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,10240,0.034972445832358465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,8192,0.0421635541650984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,3072,0.05128533310360379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,6144,0.015220445063379077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,51200,0.06440444125069512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,512,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,32,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,768,0.00703288863102595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,51200,0.3290835486518012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,1024,0.007108444141017065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,768,0.00721866637468338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,7168,0.015561777684423657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,16384,0.016543111867374845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,4096,0.010094222095277574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,2048,0.1581102212270101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,1536,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,51200,0.11368978023529053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,3072,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,64,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,128,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,10240,0.04418489005830553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,10240,0.0713555547926161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,3072,0.08743377526601155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,12288,0.018412444326612685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,12288,0.27165510919359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,3072,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,5120,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,1024,0.01717511150572035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,2048,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,64,0.006066666709052191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,10240,3.085936016506619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,1536,0.010057777994208867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,10240,0.047268443637424044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,256,0.05928888585832384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,2560,1.075595537821452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,4096,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,4096,0.014646222194035849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,12288,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,51200,0.11190222369299994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,51200,0.24002223544650605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,7168,0.033089776833852134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,65536,0.08541244268417358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,2560,0.012131555212868584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,12288,0.2307751178741455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,16384,0.12879110707177058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,1536,0.031202667289310034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,51200,0.05552088883188036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,128,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,512,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,16384,3.1246871948242188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,7168,0.1934373378753662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,3584,0.009810667071077559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,768,0.026363554928037856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,32,0.006752000086837345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,16384,0.1487777762942844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,2560,1.4464293585883246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,2560,0.0232968893316057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,16384,0.06285244226455688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,51200,3.8638526068793406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,12288,0.029620445436901514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,32,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,2560,0.078329775068495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,3584,0.03096533483929104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,10240,0.057948443624708385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,1024,0.06694844696256849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,10240,0.04408444298638237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,3072,0.07038488652971056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,1536,0.22573423385620117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,65536,1.2732506857977974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,3584,0.06008533636728922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,3584,0.06931289037068684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,4096,0.026075555218590632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,2560,0.06307733058929443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,3584,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,6144,0.028029332558314007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,3584,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,2560,0.3332115544213189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,5120,0.009310222334331935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,12288,0.03659288750754462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,256,0.023362666368484497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,4096,0.016855110724767048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,1536,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,10240,0.22825333807203504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,768,0.0026613333159022858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,512,0.006011555592219035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,1536,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,8192,0.22637510299682617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,2048,0.007335111498832703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,8192,0.022687999738587275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,8192,0.03990844554371304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,12288,0.05173422230614556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,64,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,7168,0.04144800040456984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,512,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,512,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,8192,0.10541155603196885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,6144,0.010780444575680627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,8192,0.023737778266270954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,768,0.02606933315594991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,8192,0.01279022213485506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,8192,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,12288,0.10498666763305664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,12288,1.1260764863755968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,32,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,32,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,1024,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,256,0.006707555717892117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,2560,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,3072,0.010285333626800114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,3584,0.007725333174069722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,4096,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,256,0.0026506665680143568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,51200,20.143809000651043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,12288,0.06473955843183729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,768,0.003488000068399641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,256,0.18350844913058809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,7168,0.03791733251677619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,2048,0.033985777033699885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,10240,0.0336133340994517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,8192,0.43291023042466903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,1536,0.04750222298834059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,8192,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,1024,0.020620443754725985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,4096,0.03937066594759623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,1024,0.01147466649611791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,51200,0.8445413377549914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,4096,0.07768711116578844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,128,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,768,0.016152888536453247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,512,0.01091822236776352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,6144,0.035191112094455294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,4096,0.07227466503779094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,12288,0.029044446018007066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,51200,0.4486835797627767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,51200,0.6045680046081543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,12288,0.19081332948472765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,3584,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,2048,0.015944888194402058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,1536,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,3072,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,7168,0.05188622077306112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,3072,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,10240,0.12023644977145725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,2048,0.5718284183078343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,512,0.010027555955780877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,6144,0.04506311151716444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,1536,0.0038613333470291565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,6144,0.10236444738176133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,32,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,512,0.013234666652149625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,16384,1.069591098361545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,10240,0.31542934311760795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,10240,0.012418666647540199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,2560,0.029616001579496596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,2048,0.012817777693271637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,65536,0.024679111109839544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,768,0.011513777905040316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,4096,0.016908443636364408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,512,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,128,0.027277333868874445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,128,0.03551377852757772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,128,0.02016799979739719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,1536,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,8192,0.03196355700492859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,12288,0.020038222273190815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,10240,0.3699973424275716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,4096,0.0120319997270902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,32,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,2560,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,12288,0.058576888508266874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,2048,0.01920088960064782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,2560,0.013081777426931592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,1536,0.0824471116065979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,6144,0.0372613337304857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,3072,0.0070773329999711775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,3584,0.0437102218468984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,3072,0.018509333332379658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,2048,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,4096,0.10083288616604275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,4096,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,2560,0.05529600050714281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,2048,0.006739555547634761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,256,0.005405333307054307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,64,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,2560,0.4837946891784668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,16384,0.02123466630776723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,3072,0.0595102243953281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,3072,0.09069866604275173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,4096,0.5402355723910862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,1536,0.09899110926522149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,768,0.022405332989162866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,10240,0.06822399960623847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,16384,0.05341777867741055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,2560,0.005997333261701796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,12288,0.032299555010265775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,8192,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,7168,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,3584,0.5386800236172146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,128,0.022262222237057153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,3072,0.013148444394270578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,65536,0.24868265787760416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,12288,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,12288,0.04205333193143209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,128,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,4096,0.016805332567956712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,8192,0.022882666852739122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,7168,0.0753928886519538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,2560,0.017073776986863878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,768,0.0038248888320393036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,16384,0.016191111670600045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,2560,0.017889777819315594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,128,0.016584000653690763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,128,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,10240,0.1395804484685262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,2560,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,16384,0.12832266754574245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,256,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,51200,0.8619413375854492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,4096,0.010614222122563256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,8192,0.03920177618662516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,512,0.012170666621790992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,256,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,1024,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,512,0.027623999449941847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,1024,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,4096,0.2578168974982368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,1024,0.034522665871514216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,512,0.016794666647911072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,256,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,7168,0.06605866882536146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,10240,0.03354488809903463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,6144,0.008996444443861643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,2560,0.020759110649426777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,256,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,2048,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,1536,0.0038702223036024305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,2048,0.01292711082432005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,10240,0.14691555500030518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,7168,0.014408000641398959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,1024,0.025342222717073228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,65536,0.4139111042022705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,128,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,1536,0.027903109788894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,51200,1.0468595292833116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,512,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,768,0.08045511113272773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,256,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,64,0.0028497777465316984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,3584,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,3584,0.01796000036928389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,12288,1.103334214952257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,7168,0.19029600090450713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,8192,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,12288,0.10016000270843506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,3584,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,65536,3.225045310126411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,1024,0.013366222381591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,4096,0.32807376649644643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,128,0.02137511140770382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,32,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,3584,0.008811555802822113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,512,0.00676711110605134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,8192,1.1505858103434246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,16384,0.03890044490496317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,2048,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,8192,0.4312728775872125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,2048,0.05191466543409559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,4096,0.033767110771603055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,6144,0.06634044647216797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,8192,0.03442488776312934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,7168,0.013520888984203339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,2560,0.02510133385658264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,32,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,7168,0.21369422806633842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,10240,0.29982044961717397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,7168,0.032053334845436945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,2048,0.014205333259370593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,4096,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,16384,0.04417511158519321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,256,0.03563200102912055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,2560,0.015232000086042615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,512,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,768,0.017159110969967313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,1536,0.02571288910177019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,512,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,2048,0.009714666340086196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,10240,0.13686666223737928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,51200,0.5648088985019261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,2560,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,768,0.007438222567240397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,32,0.26707198884752065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,10240,0.12232800324757893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,1024,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,65536,0.15276355213589138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,3584,0.1729724407196045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,12288,0.016364443633291457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,2048,0.014714666538768344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,16384,0.03086044390996297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,128,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,5120,0.026783999469545152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,16384,0.021743999587164983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,2048,0.02539199921819899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,51200,0.05098488926887512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,51200,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,65536,0.06645688745710585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,256,0.005294222384691238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,64,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,6144,0.05386399891641405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,3072,0.10598133007685344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,12288,0.17465333143870035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,128,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,1536,0.003479111111826367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,16384,0.06396622127956815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,6144,0.06054400073157417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,6144,0.03199111090766059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,256,0.06857155428992377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,7168,0.024937777055634394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,128,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,4096,0.0886346697807312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,256,0.0022408889813555614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,256,0.008911111288600499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,512,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,512,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,2048,0.031996445523367986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,65536,0.0629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,16384,0.04024622175428603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,1536,0.007548444800906711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,32,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,512,0.05670844184027778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,16384,0.17240622308519152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,3072,0.026379555463790894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,768,0.0041697778635554844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,3072,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,65536,0.07011377811431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,8192,0.025568000144428674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,256,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,16384,0.1306648916668362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,4096,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,10240,0.04504444532924228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,2048,0.3770879904429118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,51200,0.16014311048719618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,128,0.05152088734838697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,12288,0.02199733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,5120,0.05626044670740763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,16384,0.09691466887791951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,8192,0.013030221892727746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,10240,0.016407110624843173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,64,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,32,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,5120,0.0837635530365838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,65536,0.1945564481947157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,65536,1.0695999993218315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,128,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,6144,0.015987555185953777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,256,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,65536,0.3262568844689263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,8192,0.01975733372900221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,2048,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,5120,0.017079111602571275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,128,0.015174221661355762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,256,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,51200,0.07318133115768433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,32,0.005787555542257097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,16384,0.04169599877463447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,512,0.002516444358560774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,768,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,3584,0.029179556502236262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,65536,0.7157546679178873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,51200,0.06406222449408637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,1536,0.01642400026321411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,65536,2.6616738637288413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,51200,0.6120888921949599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,32,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,6144,0.3042488892873128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,32,0.008371555142932469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,51200,0.11648800637986924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,4096,0.07937955856323242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,51200,2.4169938829210067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,51200,1.5694320466783311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,768,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,3584,0.02120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,2560,0.011729777687125735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,1536,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,51200,0.37863911522759336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,1024,0.006071110980378256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,256,0.03033866816096836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,51200,0.11990133921305339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,512,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,768,0.016304888659053378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,1024,0.01851733360025618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,512,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,16384,0.048447112242380776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,7168,0.031144890520307753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,2560,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,2048,0.025399999486075506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,7168,0.013246222502655454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,64,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,65536,0.07040444347593519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,32,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,12288,0.007546666595670912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,65536,0.06123289134767321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,32,0.034318221939934626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,128,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,5120,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,4096,0.026872888207435608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,2560,0.01752711170249515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,16384,0.126310216055976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,12288,0.03488622109095255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,128,0.012315555579132505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,2560,0.020232889387342665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,512,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,1024,0.021486222743988037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,64,0.008784000244405534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,6144,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,10240,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,7168,0.08965155813429092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,12288,0.01370844410525428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,64,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,64,0.07891111241446601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,512,0.010679999987284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,1024,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,12288,0.03251111176278856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,12288,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,12288,0.05190489027235243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,1536,0.020612445142534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,8192,0.019010666343900893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,1536,0.010092444717884064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,51200,0.04764266808827718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,7168,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,1024,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,16384,0.17562222480773926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,768,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,3072,0.09249155388938056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,6144,0.03428444597456191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,6144,0.10275110933515762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,4096,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,12288,0.03854488995340135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,51200,0.41577866342332626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,128,0.016099555624855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,2560,0.04313244422276815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,16384,0.2850026554531521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,512,0.029027554723951552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,3072,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,3072,0.03143289022975498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,10240,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,3584,0.06838133600023058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,32,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,768,0.011422221859296164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,2560,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,51200,1.2539520263671875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,1536,0.030472887886895075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,6144,0.02431022293037838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,1536,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,1536,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,512,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,512,0.006295111030340195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,256,0.008918222453859117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,4096,0.03560355636808608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,256,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,1024,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,128,0.010236444572607676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,1536,0.004640000147951974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,256,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,64,0.013193777865833707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,16384,0.04843022094832527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,64,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,1536,0.02330488959948222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,768,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,32,0.24633600976732042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,16384,0.05285955468813578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,16384,0.14332977930704752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,256,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,12288,0.06021333403057522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,768,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,1536,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,16384,0.3655066755082872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,8192,0.024307555622524683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,2560,0.02107911143038008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,1024,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,65536,0.06667821937137179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,768,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,3584,0.17461510499318442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,16384,0.019867555962668527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,12288,0.018544000056054857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,4096,0.04165244433614943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,1024,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,5120,0.02086755633354187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,1536,0.011420444481902652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,1024,0.006008000009589725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,768,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,2048,0.13977600468529597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,1024,0.02901244494650099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,7168,0.02050844497150845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,4096,0.1943324406941732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,10240,0.014128888646761576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,1024,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,256,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,3072,0.04212711254755656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,4096,0.14104178216722277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,5120,0.026546667019526165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,1024,0.00979555563794242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,3584,0.18299733267890084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,4096,0.03955199983384874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,1536,0.11382310920291477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,2560,0.012346666720178394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,6144,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,3072,0.13076444466908774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,256,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,2048,0.00907466643386417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,12288,0.029176887538697984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,7168,0.11397955152723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,2048,0.025702221526039973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,8192,0.03948355714480082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,3584,0.02590311070283254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,1024,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,12288,0.18954843944973418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,8192,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,768,0.01847111185391744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,3072,0.0411155554983351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,12288,0.017835555805100333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,512,0.030809776650534734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,65536,1.1870098114013672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,4096,0.03195022212134467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,8192,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,1024,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,7168,0.33594311608208555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,16384,0.02903466754489475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,1024,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,32,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,256,0.15967644585503474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,51200,0.830984009636773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,768,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,1024,0.04814044303364224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,10240,0.12986932860480413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,4096,0.024300444457266066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,1536,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,12288,0.049588445160124034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,1536,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,6144,0.1695173316531711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,10240,0.13437155882517496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,512,0.02120622164673275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,6144,0.04579822222391764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,2560,0.022279999322361414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,2048,0.02964977754486932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,6144,0.043215112553702466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,768,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,7168,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,3584,0.02626222206486596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,3584,0.010115555591053432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,2048,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,65536,0.8964133792453342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,256,0.15692888365851507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,3072,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,51200,0.07045511404673259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,1024,0.026322666141721938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,6144,0.20719733503129747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,6144,0.07607022258970472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,1536,0.14878400166829428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,4096,0.11813867092132568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,512,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,768,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,7168,0.02196177840232849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,8192,0.3312026659647624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,7168,0.02181333303451538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,2048,0.030000001192092896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,4096,0.010419555836253695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,512,0.005535999933878581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,3584,0.027497778336207073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,2048,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,12288,0.0829440024163988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,128,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,10240,0.10814755492740208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,128,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,10240,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,65536,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,7168,0.031153778235117596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,4096,0.020462221569485135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,64,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,16384,0.06208622455596924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,16384,0.04842577709092034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,128,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,5120,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,2048,0.017328888177871704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,12288,0.03968266646067301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,51200,0.7530710962083605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,7168,0.05993511279424032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,4096,0.048822220828798085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,1536,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,65536,0.070942223072052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,8192,0.09822489155663384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,10240,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,1024,0.02218222286966112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,768,0.010409778190983666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,768,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,64,0.004211555752489301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,51200,1.0404684278700087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,8192,0.11837066544426812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,1536,0.03268444538116455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,64,0.0063697778516345555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,65536,1.4800853729248047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,512,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,1536,0.1909644471274482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,768,0.03784711162249247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,6144,0.12828177875942653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,768,0.025154666768179998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,768,0.03575377662976583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,64,0.010314666562610203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,128,0.022641777992248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,4096,0.41257421175638836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,3072,0.011695110963450538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,1536,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,7168,0.008857777549160851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,512,0.09217333131366306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,51200,0.32295110490587026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,12288,0.29302578502231175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,64,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,1024,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,10240,0.0209840006298489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,128,0.019627556204795837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,64,0.01315111087428199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,2560,0.023261333505312603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,5120,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,51200,0.06268622477849324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,2560,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,10240,0.031552000178231135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,6144,0.010951110886202918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,10240,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,256,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,10240,0.15625422530704075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,768,0.018760888112915885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,64,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,1024,0.013125333521101209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,8192,0.022984888818528917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,32,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,5120,0.48296356201171875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,2048,0.022278222772810195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,16384,0.01497511068979899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,768,0.024604444702466328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,51200,0.9436897701687283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,8192,0.013968888256284924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,5120,0.07171911001205444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,6144,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,4096,0.06683022446102567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,6144,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,512,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,2048,0.024715556038750544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,7168,0.028166221247778997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,2560,0.012624888784355588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,2560,0.023896889554129705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,4096,0.03391644358634949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,12288,0.013088888592190213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,3072,0.016474665866957772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,12288,0.018589332699775696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,1536,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,2560,0.09200711382759942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,1536,0.012438221938080259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,256,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,10240,0.051719112528694995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,6144,0.03577866819169786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,65536,0.03266933229234483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,256,0.018332444959216647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,3072,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,51200,0.11596622731950547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,4096,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,4096,0.023668444818920557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,6144,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,3072,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,32,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,8192,0.012895111408498553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,12288,0.0372977786593967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,32,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,3072,0.21526400248209634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,768,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,6144,0.011408000356621213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,2560,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,12288,0.07489599784215291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,65536,0.08831822209888035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,3584,0.02837688724199931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,16384,0.021135111649831135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,2560,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,4096,0.018388445178667705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,65536,0.14942310916052923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,12288,0.028840889533360798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,16384,0.03685244586732652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,3072,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,256,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,256,0.05470310979419284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,10240,0.10061955451965332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,1024,0.042048000627093844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,65536,0.07420177592171563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,12288,0.7584408654106988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,128,0.031219555271996394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,7168,0.02014400064945221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,10240,0.01329955541425281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,16384,0.3878862327999539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,51200,1.0133049223158095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,1024,0.015676443775494892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,16384,0.005394666559166378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,128,0.022286223040686712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,1024,0.0063662222690052455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,1536,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,3072,0.023934221929974023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,5120,0.014083556003040738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,512,0.01819644371668498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,7168,0.10939555697970921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,6144,0.10477599832746719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,2560,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,10240,0.06064089139302572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,16384,0.060192002190483936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,12288,0.25841599040561253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,512,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,1536,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,51200,0.13908444510565862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,3072,0.01112711098459032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,1024,0.003911111089918348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,4096,0.23417954974704316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,1024,0.011710222396585675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,2560,0.03416711091995239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,65536,0.08354577753278945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,32,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,16384,0.0669120020336575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,1024,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,12288,0.016957332690556843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,256,0.04593511091338264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,12288,0.048900445302327476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,51200,0.05472177929348416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,32,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,16384,0.07907288604312472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,7168,0.033688889609442815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,8192,1.0615208943684895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,65536,0.22415288289388022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,3072,0.024496888120969135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,5120,0.013008000122176277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,5120,0.010213333699438307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,51200,0.05254044466548496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,12288,0.17704089482625326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,5120,0.01533333294921451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,2560,0.042592889732784696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,768,0.020678222179412842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,16384,0.031848000155554876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,2560,0.022593778040673997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,12288,0.2324319945441352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,2048,0.010667555862002902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,65536,0.9764933056301541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,1536,0.008936888641781276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,2048,0.00416711096962293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,1024,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,64,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,12288,0.02184444500340356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,51200,0.014088888963063559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,1024,0.03188622328970168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,768,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,3072,0.7314737637837728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,2560,0.029847999413808186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,8192,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,51200,0.05179288983345032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,3584,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,3072,0.06300444735421075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,16384,0.02400622268517812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,4096,0.030469334787792627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,7168,0.00959822204377916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,12288,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,7168,0.03587199913130866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,256,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,65536,0.06099377738104927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,6144,0.023525334066814844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,4096,0.009391999906963771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,16384,0.01995111174053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,51200,0.11840621630350749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,5120,0.026322666141721938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,5120,0.04012088974316915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,3072,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,8192,1.5903537538316515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,3584,0.14289600319332546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,768,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,6144,0.017280000779363845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,2048,0.03836888737148709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,1024,0.021313778228229944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,1024,0.027102222045262653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,65536,1.0532880359225802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,3584,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,128,0.019318222999572754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,1536,0.005049777941571342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,2560,0.007366221812036302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,3584,0.10779466893937852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,51200,0.12937511338127985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,16384,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,1024,0.016380444169044495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,512,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,16384,0.03531733486387465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,16384,0.06525422467125787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,5120,0.016735111673672993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,32,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,3072,0.005456888841258154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,16384,0.02051911089155409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,3584,0.007601777712504069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,64,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,51200,5.7242185804579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,6144,0.025524444050259058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,128,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,51200,0.055083556307686694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,10240,0.1831484503216214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,5120,0.012267555627557965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,768,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,128,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,51200,0.4464133580525716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,1024,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,5120,0.030710223648283217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,16384,0.03309155503908793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,128,0.018423111902342904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,512,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,1024,0.02313066687848833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,16384,0.30607912275526256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,64,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,5120,0.017045333981513977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,10240,0.04145155681504144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,51200,0.5085653199089898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,10240,0.330859555138482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,16384,0.022427555587556627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,768,0.02035999960369534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,5120,0.012886222038004132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,5120,0.12633955478668213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,2048,0.03493422269821167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,2560,0.027217777238951787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,10240,0.0120319997270902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,8192,0.01331022216214074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,768,0.002540444541308615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,10240,0.04543911086188423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,2048,0.025055110454559326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,51200,0.3295102119445801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,512,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,16384,0.17313867145114473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,128,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,768,0.0032142222755485107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,8192,0.10477688577440049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,65536,0.6447591251797146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,51200,0.6723795466952853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,16384,0.04033689035309686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,1536,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,128,0.12638489405314127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,10240,0.172508438428243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,10240,0.026745777991082933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,12288,0.01965688831276364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,8192,0.03148266673088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,2048,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,5120,0.00943199959066179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,3072,0.09808533059226142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,768,0.03856711255179511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,1536,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,8192,0.026597332623269822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,4096,0.3241928948296441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,65536,0.8967404895358615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,65536,0.8447413444519043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,4096,0.010279999838935005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,64,0.007575111256705389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,128,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,4096,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,2048,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,128,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,128,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,5120,0.008616888688670265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,5120,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,64,0.007373332977294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,5120,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,8192,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,8192,0.004949333353175057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,51200,0.058218664593166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,2048,0.023383999864260357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,10240,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,2048,0.13111555576324463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,512,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,3584,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,16384,0.020592000749376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,4096,0.0474639998541938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,5120,0.01926399932967292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,6144,0.011373333632946014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,2048,0.012903110848532783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,768,0.07187910874684651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,128,0.019544000426928203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,5120,0.024101333485709295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,512,0.021075555019908484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,1024,0.013536888692114087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,768,0.01923644377125634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,7168,0.06609244479073419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,3072,0.015198222464985318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,128,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,12288,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,12288,0.23952889442443848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,4096,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,768,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,5120,0.13292355007595485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,2048,0.03683911098374261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,768,0.027870221270455256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,512,0.011857777833938599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,256,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,51200,0.03803999887572394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,16384,0.02163999941613939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,7168,0.010669333239396414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,65536,0.06784355640411377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,2560,0.16057333681318495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,256,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,51200,0.05432711044947306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,2560,0.029153777493370905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,768,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,16384,0.280603567759196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,2048,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,2560,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,2048,0.01755911111831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,5120,0.008722667064931657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,2560,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,4096,0.019699555304315355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,3584,0.09103822045856053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,1536,0.08977422449323867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,10240,0.0277164445983039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,16384,3.5842488606770835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,1536,0.05061066812939114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,6144,0.4507466422186957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,2560,0.007165333463086023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,1536,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,8192,0.046161777443355985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,7168,0.05286666750907898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,10240,0.41922932200961643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,768,0.02459288967980279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,512,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,10240,0.04976355698373583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,256,0.005848888721730974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,3584,0.025573333104451496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,8192,0.02296000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,3072,0.039108445247014366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,2560,0.01702222228050232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,512,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,65536,0.025269332859251235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,16384,0.19448178344302705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,7168,0.029057777590221826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,5120,0.07816977633370294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,768,0.03980799847178989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,1536,0.0036862223512596558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,6144,0.012442666623327466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,7168,0.019895111521085102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,32,0.07814044422573514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,4096,0.012114666402339935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,12288,0.04720088839530945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,1024,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,3584,0.014408888088332282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,51200,0.058802664279937744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,128,0.0022044444663657085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,768,0.01054044481780794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,65536,0.12384888860914443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,3072,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,7168,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,8192,0.007012444237867991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,768,0.014125333891974555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,768,0.009553777674833933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,7168,0.016075554821226332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,1024,0.0032604444358083936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,10240,0.012153777811262341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,4096,0.029772443903817072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,3584,0.026250667042202417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,1024,0.02585244509908888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,3072,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,65536,1.051020410325792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,65536,0.2854391203986274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,2560,0.009226666556464301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,64,0.01163200040658315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,1024,0.0064693333374129395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,32,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,2048,0.02644266684850057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,16384,0.17226933108435738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,2048,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,8192,0.020785777105225455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,16384,0.07571111122767131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,51200,0.037512001064088606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,3584,0.025620443953408137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,65536,0.0665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,16384,0.04977866676118639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,1024,0.017163554827372234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,8192,0.05770844221115112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,51200,0.36711910035875106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,3072,0.013528889252079858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,128,0.02442755632930332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,6144,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,256,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,10240,0.07622577746709187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,3584,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,3584,0.016391999191708036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,32,0.011724444727102915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,32,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,32,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,2560,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,10240,0.8169715669420031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,4096,0.023050665855407715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,2048,0.03224088748296102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,256,0.04664088951216804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,7168,0.09091111024220784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,1536,0.014109333356221518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,3072,0.0478097763326433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,51200,2.0471252865261502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,6144,0.03989955451753404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,2048,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,3072,0.020308444897333782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,64,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,512,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,3584,0.08814933564927842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,3584,0.06221155325571696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,64,0.006045333213276333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,2048,0.04211733407444424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,512,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,64,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,10240,0.05348177750905355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,768,0.002879111096262932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,1536,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,51200,0.03572088811132643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,6144,0.04204622242185804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,2560,0.007183999651008182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,5120,0.02264533274703556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,65536,0.6125466558668349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,2560,0.0214711113108529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,65536,0.5022586716545953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,2560,0.04533066683345371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,3072,0.007473777565691207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,51200,0.08682488732867771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,6144,0.10803466373019749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,51200,0.18673511346181235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,128,0.027292443646325007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,10240,0.04233955674701267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,3072,0.018031999468803406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,2048,0.12338399887084961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,1024,0.015213333898120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,1024,0.013535999589496188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,12288,0.016551110479566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,51200,0.03799022237459818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,3584,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,65536,0.022591999835438196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,65536,0.08558399809731378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,5120,0.024022221565246582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,128,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,65536,0.6332071092393663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,16384,0.01700622174474928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,10240,0.012846222354306115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,256,0.007393777370452881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,768,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,2048,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,128,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,128,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,2048,0.08044444190131293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,512,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,6144,0.014653333359294467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,1024,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,128,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,3584,0.0188400000333786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,7168,0.010321777727868822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,2560,0.03233155608177185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,3584,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,1024,0.021145777569876775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,51200,0.06655822197596233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,2560,0.007686222592989604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,128,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,3072,0.06119110849168566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,1024,0.011372444530328115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,5120,0.0343831115298801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,256,0.00620266670982043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,6144,0.010026666853162978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,2048,0.02973955538537767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,5120,0.03252977795071072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,2048,0.02923733327123854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,6144,0.08668888939751519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,64,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,1024,0.024717332588301763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,2560,0.012871999707486896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,768,0.014918223023414612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,51200,0.029838220940695867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,2560,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,65536,0.15090221828884548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,3072,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,128,0.02179466684659322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,51200,0.06157333321041531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,1536,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,8192,0.02458933326933119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,6144,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,8192,0.03003466791576809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,32,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,2560,0.025433777107132807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,51200,0.051790220869912036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,3584,0.007346666521496243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,768,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,3072,0.13076444466908774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,512,0.01147466649611791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,1024,0.012158222496509552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,51200,9.026969061957464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,128,0.0038480001191298165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,2560,0.031727999448776245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,5120,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,10240,0.014829332629839579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,4096,0.010621333287821876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,128,0.10780888795852661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,5120,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,10240,0.021132444341977436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,2048,0.01919911139541202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,12288,0.04084711273511251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,32,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,6144,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,7168,0.037442667616738215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,2560,0.010308444499969482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,51200,0.548525333404541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,128,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,65536,2.4110018412272134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,10240,0.05348177750905355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,128,0.01296622223324246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,65536,0.17306221856011283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,768,0.017134222719404433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,1024,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,5120,0.06383466720581055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,32,0.010407999985747868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,64,0.014744889405038623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,6144,0.031888888941870794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,10240,0.028441776831944782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,51200,3.638755374484592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,128,0.015585778488053216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,768,0.032872888776991106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,16384,0.05384622348679436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,51200,0.6245786878797743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,128,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,10240,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,64,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,10240,0.03940533267127143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,2048,0.022258667482270136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,256,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,6144,0.010185778141021729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,7168,0.12372977203792995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,12288,0.01423377792040507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,5120,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,256,0.006875555548402999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,7168,0.018210666047202218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,16384,0.4182017909155951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,1024,0.019699555304315355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,65536,0.08248177501890394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,2560,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,1536,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,12288,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,10240,0.08686755763159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,5120,0.02405511173937056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,7168,0.07587377892600165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,5120,0.027715555495686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,6144,0.0110880004035102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,3072,0.01628800067636702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,3072,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,5120,0.029488888051774766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,12288,0.04689866635534498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,256,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,6144,0.040693332751592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,3072,0.5606506665547689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,64,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,768,0.013887999786270989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,128,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,10240,0.01349066694577535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,16384,0.035960889524883695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,65536,0.1370826694700453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,10240,5.0906524658203125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,6144,0.1911546654171414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,1536,0.5336728625827366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,65536,0.18187556001875135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,6144,0.026874666412671406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,512,0.29711198806762695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,8192,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,51200,0.47153155008951825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,32,0.0033200000309281876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,2048,0.02455288834042019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,3584,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,2048,0.006776888751321369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,65536,0.11828088760375977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,16384,2.895571602715386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,5120,0.040904889504114784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,16384,0.10800711313883464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,512,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,2048,0.17664710680643717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,10240,1.2729804780748155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,768,0.008434666527642144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,512,0.003871110992299186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,4096,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,7168,0.07193777958552043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,1536,0.009776888622177972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,5120,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,51200,2.708823098076714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,3584,0.07338400019539727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,65536,0.12686310874091253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,2560,0.018011555075645447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,8192,0.048972444401847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,32,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,256,0.01720533271630605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,2560,0.00702400008837382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,6144,0.024538666009902954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,3072,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,16384,0.012873777912722694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,8192,0.06320622232225206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,1024,0.2821306652492947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,7168,0.03809244434038798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,12288,0.05854755640029907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,12288,0.03894222113821242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,2048,0.006738666858938005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,3072,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,64,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,128,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,8192,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,51200,0.048659556441836886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,16384,0.06473600202136569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,64,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,5120,0.028361777464548748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,3072,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,6144,0.014559111661381192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,32,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,4096,0.08238489098019071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,1536,0.020636444290479023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,8192,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,256,0.003842666745185852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,12288,0.04848533206515842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,16384,0.19083821773529053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,768,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,256,0.014561777313550314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,512,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,4096,0.05084888802634346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,5120,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,2560,0.007869333028793335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,1024,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,7168,0.03611822260750665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,16384,0.04572800132963392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,12288,0.21198577351040312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,768,0.005721777677536011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,3584,0.22312710020277235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,128,0.001981333312061098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,512,0.0169493340783649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,16384,0.05145688851674398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,3072,0.011909333368142446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,3584,0.04843022094832527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,2560,0.01681155628628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,10240,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,51200,0.056051552295684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,512,0.04852711160977682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,1024,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,3584,0.012996444271670448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,16384,0.16893333858913848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,7168,0.023968888653649226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,3072,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,3584,0.03291199935807122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,3584,0.015104000767072042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,65536,0.22814223501417372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,512,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,7168,0.15789243910047743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,3072,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,2048,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,3584,0.03050400151146783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,768,0.03559199968973795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,3072,0.01626755628320906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,256,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,51200,0.09887199931674534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,16384,0.020971554848882887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,768,0.008655110994974772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,256,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,8192,0.04142133394877116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,16384,0.5927555296156142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,2560,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,4096,0.020303999384244282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,128,0.007054222126801808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,7168,0.21017421616448295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,1536,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,256,0.019873778025309246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,2048,0.0653653343518575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,128,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,256,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,1024,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,1536,0.0041742221348815495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,2560,0.011045333411958484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,7168,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,32,0.004254222330119875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,3072,0.05380622214741177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,65536,0.10129955742094253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,65536,0.8882071177164713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,12288,0.024400000770886738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,256,0.020263110597928364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,3584,0.014076444837782117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,16384,0.10414844751358032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,2048,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,16384,0.5673342280917698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,2048,0.07433066765467326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,8192,0.04140533341301812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,512,0.009880888793203566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,65536,0.5643635325961643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,2560,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,3072,0.02159822152720557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,6144,0.04281066523657905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,51200,0.08775111039479573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,1536,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,6144,0.04499199986457825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,10240,0.40919110510084367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,768,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,16384,0.04861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,1024,0.006887110984987683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,51200,0.06530400117238362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,51200,0.512481795416938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,8192,0.09832977586322361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,3584,0.006483555667930179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,3072,0.06276444594065349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,1024,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,64,0.2008275588353475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,32,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,3584,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,12288,0.0981280008951823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,128,0.02111377815405528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,2048,0.028823110792371962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,1536,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,128,0.0270115551021364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,4096,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,256,0.015842666228612263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,768,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,2048,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,5120,0.01590044465329912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,2048,0.03641866644223531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,3072,0.01111822244193819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,32,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,12288,0.05312533179918925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,65536,0.07797333266999987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,1536,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,512,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,3584,0.0203795548942354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,768,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,128,0.003481777591837777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,16384,0.05571999814775255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,128,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,3584,0.028011557128694322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,1536,0.031200889084074233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,2560,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,16384,0.051319112380345665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,2048,0.1173235575358073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,6144,0.08268800046708849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,6144,0.21508177121480307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,2048,0.07685599724451701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,10240,0.028528001573350694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,2048,0.028006222512986925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,51200,0.12535999880896673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,256,0.006322666588756774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,8192,0.04469777809249031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,1024,0.0188737776544359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,2560,0.014736889137162102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,256,0.016591999265882704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,4096,0.05581955777274238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,1024,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,7168,0.22516356574164498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,2560,0.13907289505004883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,256,0.017495999733606975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,2048,0.0351200004418691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,51200,0.7952115800645617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,2048,0.023005333211686876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,12288,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,2560,0.06431821982065837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,10240,0.03251288996802436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,4096,0.010407110883129967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,32,0.01937599976857503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,512,0.02401688860522376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,512,0.020392000675201416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,12288,0.06337422132492065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,51200,0.06898311111662123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,7168,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,512,0.018070222602950204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,65536,0.0744782222641839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,1024,0.047285334931479565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,3584,0.06365066766738892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,768,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,65536,0.14783377117580837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,512,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,256,0.029324445459577773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,5120,0.004512888689835866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,7168,0.010863111250930362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,128,0.02584444483121236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,1024,0.008809777597586313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,128,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,16384,0.02075111038155026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,1536,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,64,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,65536,0.6312959988911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,64,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,51200,0.1851466629240248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,10240,0.06776355372534858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,3072,0.020290666156344943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,12288,0.12995110617743597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,6144,0.018970666660202872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,4096,0.4817262225680881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,12288,0.22171289390987822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,32,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,512,0.003061333257291052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,64,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,768,0.022358222140206233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,16384,0.05611644188563029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,10240,0.028213332096735638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,512,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,768,0.017514665921529133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,256,0.013261333107948303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,128,0.014109333356221518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,65536,0.035516444179746844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,10240,0.11117866304185654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,65536,0.0236835562520557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,3584,0.01035288886891471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,10240,0.46833154890272355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,64,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,64,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,256,0.09174844291475083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,10240,0.01792622274822659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,4096,0.010823110739390055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,65536,0.558486196729872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,3072,0.05872888697518242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,65536,0.17424355612860787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,64,0.017631111873520743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,4096,3.2666134304470487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,2560,0.11436443858676487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,3584,0.06405778063668145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,1536,0.017674666312005784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,10240,0.17905244562360975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,12288,1.1403466330634224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,1536,0.047263112333085805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,512,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,6144,0.025047111842367385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,3584,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,1024,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,4096,0.04974488748444451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,3072,0.00741244438621733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,16384,0.04213066564665901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,1536,0.028054220808876887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,6144,0.09870933161841498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,10240,0.01677600045998891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,32,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,768,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,512,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,1024,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,6144,0.03678844372431437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,3584,0.06325422392951117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,1024,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,16384,0.12014488379160564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,2048,0.01756888959142897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,512,0.01921599937809838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,128,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,6144,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,128,0.0017893332988023758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,64,0.02602133320437537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,2048,0.014649778604507446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,5120,0.004593777573770947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,32,0.026668444275856018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,10240,0.12600355678134495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,16384,0.6795786751641167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,3584,0.27588621775309247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,1024,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,128,0.02126933303144243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,2560,0.026484444737434387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,51200,0.016941333810488384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,64,0.02402577797571818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,8192,0.011508444117175208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,4096,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,16384,0.11317333910200332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,4096,0.028688000308142766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,32,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,8192,0.011210666762457954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,2560,0.12400266859266494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,3584,0.027064000566800434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,768,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,768,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,7168,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,16384,0.697455088297526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,64,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,12288,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,51200,0.06924711333380805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,128,0.08396889103783502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,1536,0.03193333413865831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,10240,0.04989688926272922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,5120,0.170470224486457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,12288,0.04689155684577095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,64,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,65536,0.026313778426912095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,10240,0.0129439996348487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,3584,0.02588444451491038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,12288,0.019075555933846366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,128,0.0019199999256266488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,12288,0.9349501927693685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,16384,0.045863999260796436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,5120,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,3584,0.008174222376611497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,128,0.016878222425778706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,3584,0.021589333812395733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,7168,0.023048889305856492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,512,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,768,0.010402667025725046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,7168,0.044405334525638156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,1024,0.0872337751918369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,768,0.016851555969980028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,16384,0.09947733084360759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,51200,0.1148480044470893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,8192,0.7002195782131619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,3072,0.05036711030536228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,5120,0.021707556313938562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,3072,0.08050044377644856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,5120,0.11018844445546468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,7168,0.012911111116409302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,64,0.00618488879667388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,1536,0.019300444258583915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,1536,0.006672888994216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,3584,0.02479466630352868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,4096,0.3933600054846869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,2560,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,12288,0.07552888658311632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,51200,0.013559111290507846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,1024,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,65536,0.8709724214341906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,10240,0.16295377413431802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,3584,0.023021333747439917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,64,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,8192,0.14149688349829778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,12288,0.06266044245825873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,5120,0.02294933299223582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,6144,0.021906667285495337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,1536,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,5120,0.05795200003517998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,5120,0.021839999490314062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,10240,0.0407591097884708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,5120,0.04733599887953865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,7168,0.10944533348083496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,128,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,2560,0.025578666064474318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,10240,0.02233244478702545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,16384,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,2048,0.44562398062811953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,16384,0.2923911147647434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,128,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,1536,0.020280000236299302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,512,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,51200,61.8606431749132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,65536,0.16261333889431423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,1536,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,1536,0.03509511219130622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,1024,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,8192,0.02293777796957228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,768,0.02734222180313534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,2560,0.013807999591032663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,51200,0.35996354950798887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,1536,0.033471110794279314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,5120,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,8192,0.07672622468736437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,4096,0.14502755800882974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,3072,0.017779555585649278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,1024,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,512,0.05434222353829277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,128,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,8192,0.012124444047609964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,10240,0.01219733307758967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,256,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,512,0.013461333182122974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,2048,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,5120,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,12288,0.014167111780908374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,3072,0.09821510977215236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,3584,0.02258666687541538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,64,0.012711111042234631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,2560,0.028536889288160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,768,0.06178222099939982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,51200,0.6006239785088433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,7168,0.0321066677570343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,64,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,10240,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,10240,0.023007111416922674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,6144,0.057245333989461265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,32,0.0019386667344305252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,51200,0.6086728837754991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,4096,0.030255112383100722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,7168,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,10240,0.025621333056026038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,5120,0.017479999197853934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,3584,0.005643555687533484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,128,0.07091110944747925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,128,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,5120,0.01612444387541877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,256,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,4096,0.05925599733988444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,5120,0.008430221842394935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,10240,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,768,0.01884533299340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,32,0.006207110981146495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,128,0.022671999202834234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,3072,0.13799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,16384,0.03143466512362162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,6144,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,128,0.018160000443458557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,65536,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,3584,0.00815822184085846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,512,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,2560,0.01239111108912362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,16384,4.6657663981119795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,8192,0.04642933275964525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,4096,0.0565155545870463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,512,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,4096,0.016167999969588388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,2560,0.017290666699409485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,65536,0.029327998558680218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,1024,0.028015110227796767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,2560,0.16243910789489746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,2048,0.0871608919567532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,3072,0.008319999608728621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,5120,0.0770897799068027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,256,0.016138666205936007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,2560,0.024706666668256123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,7168,0.011868444581826528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,10240,0.11478932698567708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,65536,0.33163287904527455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,2048,0.014848000473446317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,512,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,65536,1.518351131015354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,16384,0.23566222190856934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,512,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,3072,0.02664533257484436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,51200,0.17178755336337617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,2048,0.02974577744801839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,65536,0.6333608627319336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,2560,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,768,0.026950221922662523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,8192,0.030215111043718126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,2560,0.013665777941544851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,4096,0.017687999539905123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,16384,0.01962222158908844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,2560,0.07856888903511895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,4096,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,2560,0.012707555459605323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,8192,0.01277599980433782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,128,0.010382222632567087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,2560,0.2667164537641737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,8192,0.0424542228380839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,2048,0.01406222250726488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,1024,0.04598044355710348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,7168,0.219903998904758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,256,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,6144,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,3072,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,3584,1.1522195604112413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,2048,0.01828533411026001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,2560,0.007214222517278459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,128,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,8192,0.05672444237603081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,10240,0.032279110617107816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,2560,0.05271733469433255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,16384,0.0295448899269104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,65536,0.13954577181074354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,4096,0.1436986658308241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,16384,0.020993777447276648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,4096,0.010703110860453712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,16384,0.04763911167780558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,64,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,10240,0.08186133040322198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,5120,0.017647110753589206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,2560,0.006670222100284364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,4096,0.013947554760509066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,6144,0.008754666480753157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,768,0.00906577789121204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,1536,0.0392871101697286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,4096,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,1024,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,64,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,256,0.017501332693629794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,6144,0.018138666947682697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,12288,0.018090666996108163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,1024,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,256,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,2560,0.03932444585694207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,6144,0.0828568869166904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,3584,0.043064001533720225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,16384,0.02642044425010681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,1536,0.01646933290693495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,256,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,1536,0.004061333421203825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,2560,0.013680000272062091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,1024,0.008719111482302347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,768,0.24039467175801596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,6144,0.04234222239918179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,768,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,65536,0.0596275528271993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,65536,1.3484942118326824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,2048,0.02239911092652215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,512,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,3072,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,1024,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,4096,0.009331555830107795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,128,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,4096,0.008202667037645975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,768,0.011676444775528379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,2048,0.010298666854699453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,3584,0.09470133648978339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,2048,0.025613332788149517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,512,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,6144,0.03233511249224345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,4096,0.08209333154890272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,12288,0.042527112695905894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,512,0.03257777624660068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,16384,0.7411715719434949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,32,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,4096,1.5416249169243708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,5120,0.24249421225653756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,768,0.019463999403847587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,8192,0.0686640010939704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,4096,0.018216888109842937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,8192,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,4096,0.12159910466935898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,3072,0.03770133190684848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,5120,0.05927111042870415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,8192,0.16693688763512504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,2560,0.06801955567465888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,3072,0.027104000250498455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,12288,0.03727022144529555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,5120,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,768,0.042040000359217324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,1024,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,4096,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,51200,0.05515377720197042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,2048,0.2298142115275065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,32,0.006739555547634761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,1024,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,128,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,256,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,8192,0.02966133422321743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,256,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,65536,0.272053321202596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,10240,0.09291999869876438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,512,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,10240,0.6910470856560601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,5120,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,6144,0.0284862220287323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,768,0.007104888558387756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,1024,0.016806221670574613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,3072,0.05960888995064629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,768,0.037069333924187556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,10240,0.06163466638988919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,3072,0.019311999281247456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,128,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,128,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,2048,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,65536,0.13479199674394396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,3072,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,65536,0.12245244450039333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,1024,0.14869956175486246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,1024,0.026734222968419392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,128,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,6144,0.04058488872316148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,7168,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,65536,0.29649776882595485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,768,0.1149795585208469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,256,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,512,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,51200,0.33075910144382054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,7168,0.013275555438465543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,4096,0.01817599932352702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,128,0.01959377858373854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,8192,0.1419688860575358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,256,0.01754133403301239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,1536,0.024327110913064744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,8192,0.15482489267985025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,51200,0.4569813410441081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,2560,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,1024,0.02273066673013899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,7168,0.10624621974097358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,2560,0.0998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,6144,0.009425777528021071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,2560,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,128,0.09123822053273518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,4096,0.03274577856063843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,7168,0.04709866642951965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,256,0.008061333662933773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,3584,0.009893333746327294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,2560,0.07865777942869398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,64,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,7168,0.10881600115034316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,512,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,5120,0.03762133253945245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,1024,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,65536,0.05662577682071262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,128,0.008729777402347988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,1024,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,10240,0.03345422281159295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,2560,0.007033777733643849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,8192,0.01827555563714769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,2560,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,128,0.010793777803579966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,512,0.03273955649799771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,7168,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,3072,3.698632982042101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,10240,2.0011634826660156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,768,0.13749066988627115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,256,0.0035031110876136353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,3584,0.06298400296105279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,2560,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,7168,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,2048,0.018593778212865193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,2048,0.007744000189834171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,7168,0.030917333232031927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,3072,0.024061333802011278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,16384,0.26614220937093097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,5120,0.009523555636405945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,256,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,10240,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,3072,0.0075288886825243635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,2560,0.007163555257850223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,768,0.04381510946485731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,12288,0.07945333586798774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,3584,0.008018666671382057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,12288,0.4624017609490289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,2048,0.33845867051018613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,16384,0.037984887758890785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,64,0.016720889343155753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,16384,40.11892361111111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,10240,0.027129777603679236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,5120,0.024660444921917383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,16384,0.26807822121514213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,10240,0.01683999929163191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,8192,0.022123555342356365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,256,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,5120,0.010925333533022137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,1536,0.006992888947327931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,10240,0.04523644513554043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,128,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,16384,0.026201777988009985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,256,0.020240000552601285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,256,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,4096,0.012387555506494312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,2560,0.23303643862406412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,16384,0.0185199992524253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,7168,0.03682488865322537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,2048,0.07370133532418145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,10240,0.01625955601533254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,10240,0.012375999655988483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,6144,0.11788800027635361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,1024,0.06371821959813435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,768,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,512,0.008160889148712158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,5120,0.015058666467666626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,4096,0.17868622144063315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,10240,1.381566259596083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,8192,0.014070222775141398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,3584,0.07593244314193726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,7168,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,1024,0.021314667330847845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,65536,0.13805066214667425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,768,0.04754755563206143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,768,0.004936000125275718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,65536,0.5339182217915853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,8192,0.06250666909747653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,4096,0.012712000144852532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,128,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,6144,0.0914728906419542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,2048,0.015537778536478678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,128,0.013119999733236102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,2560,0.012469333079126147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,3584,0.03144888745413886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,51200,0.048856000105539955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,5120,0.009465778039561378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,10240,0.033253331979115806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,65536,0.16878222094641793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,16384,0.15491910775502524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,768,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,2560,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,51200,0.0579342246055603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,256,0.042188445727030434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,32,0.01016533292002148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,5120,0.02369155486424764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,768,0.0033280000918441345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,3072,0.02127733329931895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,10240,0.084733333852556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,3072,0.060452441374460854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,768,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,51200,0.21534755494859484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,16384,0.016548444827397663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,512,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,256,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,51200,0.13514667087131077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,65536,0.07944889201058282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,65536,1.649971538119846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,5120,0.00942844483587477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,1536,0.00817599975400501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,1024,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,2048,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,8192,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,65536,0.07791644334793091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,1024,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,5120,0.3359235657585992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,10240,0.016890666551060148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,32,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,6144,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,5120,0.03997155692842271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,3584,0.03591466612286038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,4096,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,64,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,5120,0.01574222246805827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,10240,0.015056888262430826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,256,0.031129777431488037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,5120,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,65536,2.508098602294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,16384,0.04426133301522997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,3584,0.3057839870452881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,10240,0.1208382182651096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,51200,0.04756888747215271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,512,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,3584,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,32,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,1024,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,3584,0.011360000405046674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,3584,0.01889866590499878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,8192,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,128,0.04223733478122287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,256,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,3584,0.0640044477250841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,51200,0.02840177880393134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,12288,0.1406266689300537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,6144,0.07591733005311754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,10240,0.00872622181971868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,65536,0.3479315439860026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,12288,0.0540924436516232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,3584,0.007416889071464539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,256,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,65536,0.20452800061967638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,2560,0.011227555572986603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,16384,0.0199306673473782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,2560,0.015614221493403116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,2560,0.0046160001721647055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,6144,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,12288,0.031183110343085393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,12288,0.200835559103224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,2048,0.024282667371961806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,128,0.016302223006884258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,16384,0.012440889245933957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,8192,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,10240,1.6936729219224718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,8192,0.039445334010654025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,2048,0.01757600075668759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,2048,0.00453333349691497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,2560,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,8192,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,2048,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,128,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,256,0.03163999981350369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,512,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,2560,0.025124443901909724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,16384,0.3511333465576172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,2560,0.013410666750537025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,65536,0.14397154914008245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,6144,0.028767999675538804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,3072,0.03187911046875848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,16384,0.046521776252322726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,3584,0.011200000014570026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,1536,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,5120,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,12288,0.05186133252249824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,12288,0.18319643868340385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,128,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,64,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,64,0.00702311098575592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,4096,0.017398221625222098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,4096,0.0719004472096761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,768,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,7168,0.014926221635606555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,7168,0.03479289015134176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,2560,0.03547644284036424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,16384,0.026100445124838088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,51200,0.6609777874416775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,2048,0.026688888669013977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,12288,0.01824711097611321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,64,0.0017768889665603638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,1536,0.01714933415253957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,10240,0.013777777552604675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,12288,0.1319928831524319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,1024,0.07754488786061604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,12288,0.10337333546744452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,10240,0.11095822519726223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,4096,0.0258906665775511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,16384,0.021254221598307293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,3072,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,3584,0.09627999862035115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,2048,0.2333564493391249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,4096,0.02956355611483256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,7168,0.02184533410602146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,65536,0.025182222326596577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,2560,0.02020355562369029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,4096,0.009763555394278632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,2048,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,65536,0.4463626543680827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,51200,0.1187351147333781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,16384,0.030400888787375555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,12288,0.024451555477248296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,5120,0.0119982221060329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,10240,0.015065777632925244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,1536,0.02055377761522929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,256,0.005342222336265776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,1536,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,12288,0.028764443265067205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,6144,0.04200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,2560,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,2048,0.02535466684235467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,768,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,4096,0.012090666426552666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,12288,3.6229697333441844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,8192,0.13350133101145426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,8192,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,7168,0.02739022175470988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,8192,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,3072,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,5120,0.0302817788388994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,768,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,8192,0.013289777768982781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,1024,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,7168,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,10240,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,128,0.006768888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,6144,0.017788444956143696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,1536,0.13200178411271837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,51200,0.05858933263354831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,256,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,10240,0.05302933189604017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,2560,0.054414222637812294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,128,0.03938311007287767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,1536,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,3072,0.09275644355350071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,1024,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,1024,0.015412444869677225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,12288,0.06736799743440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,5120,0.02887111239963108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,1024,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,5120,0.11356443829006618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,512,0.014783110883500842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,512,0.005454222361246745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,128,0.14369955327775744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,64,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,8192,0.11076088746388753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,768,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,4096,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,51200,0.054839111036724515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,256,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,6144,0.06057600180308024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,3584,0.3995795514848497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,65536,0.09364000293943618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,10240,0.48970757590399844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,3072,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,2048,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,3584,0.020561777883105807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,16384,0.21262399355570474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,12288,0.07635821898778279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,2048,0.043041778935326465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,1536,0.039015110996034406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,3584,0.014621333943472968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,64,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,65536,1.9151200188530817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,1536,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,8192,0.036995556619432234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,1536,0.052480889691246875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,7168,0.05834311246871948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,10240,0.029395557112163965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,512,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,7168,0.03541244400872125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,2048,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,6144,0.016576889488432143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,8192,0.027952889601389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,768,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,3584,0.008943110704421997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,2048,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,2560,0.1555804411570231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,7168,0.009444444543785518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,51200,0.2037982278399997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,12288,0.029688000679016113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,5120,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,12288,0.08600266774495442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,256,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,10240,0.03525244527392917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,8192,0.011496889094511667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,768,0.1405831045574612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,10240,1.1403830846150715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,16384,0.303422212600708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,2048,0.010208889014191097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,768,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,2048,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,1536,0.009835555321640438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,2560,0.02875111169285244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,4096,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,2560,0.027077333794699773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,8192,0.1993795500861274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,128,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,3072,0.029284444120195176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,256,0.012249777714411417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,1536,0.26451200909084743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,2048,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,16384,0.030290666553709242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,1024,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,64,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,2560,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,768,0.0031475555151700974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,1024,0.013191111385822296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,256,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,6144,0.012447111308574677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,3584,1.686489741007487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,2560,0.025086222423447505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,3072,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,2048,0.058789332707722984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,64,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,3584,0.023694222172101338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,51200,0.04638400011592441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,256,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,7168,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,65536,0.06327644321653578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,5120,0.4752168655395508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,5120,0.1792595518959893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,51200,0.055436445607079395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,8192,1.128893322414822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,768,0.02437422176202138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,65536,2.7625910441080728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,2560,0.11456888251834446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,8192,0.040707555082109236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,51200,0.18240622679392496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,16384,0.023984889189402264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,3584,0.013937777943081327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,32,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,12288,0.0943582190407647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,10240,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,10240,0.1293706629011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,3584,0.014506666196717156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,512,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,512,0.006002666635645761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,16384,0.800980461968316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,7168,0.3624160024854872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,256,0.013540444274743399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,3072,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,2048,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,1024,0.002895111011134254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,1536,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,4096,0.007370666497283512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,8192,0.10510577758153279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,2048,0.026942221654786006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,512,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,768,0.017596445149845548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,3072,0.01640533407529195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,32,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,128,0.015206222732861837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,4096,0.06351377566655476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,3072,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,1024,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,5120,0.016228444046444364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,51200,0.3910924328698053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,3584,0.03443200058407254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,2560,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,6144,0.03152889013290405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,4096,0.02492711113558875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,2560,0.01829600003030565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,65536,0.5911111301845974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,1536,0.020336000455750358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,12288,0.13939289251963297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,2048,0.03699377841419644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,7168,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,10240,0.020630222227838304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,512,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,65536,0.13358755906422934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,4096,0.01851377718978458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,512,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,5120,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,10240,0.10684266355302598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,16384,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,12288,0.1117582188712226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,128,0.010102222363154093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,768,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,3072,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,1024,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,6144,0.02236355510022905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,7168,0.01996888882584042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,1536,0.0074311114019817775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,512,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,768,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,3072,0.007558222446176741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,3072,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,512,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,16384,0.28734487957424587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,2560,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,32,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,2560,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,3584,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,128,0.04107644491725498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,7168,0.05455466773774889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,12288,0.01481066644191742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,1024,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,8192,0.04053777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,3072,0.021902221772405837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,51200,0.04905688762664795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,256,0.04092799954944187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,4096,0.036502222220102944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,3072,0.024280000064108107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,2560,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,1536,0.003960000144110786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,5120,0.09857955243852402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,2048,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,8192,0.04930044545067681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,1536,0.02683911058637831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,256,0.010699555277824402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,3584,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,256,0.01516266663869222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,65536,0.6415448718600804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,256,0.015765332513385348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,512,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,3584,0.018160000443458557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,8192,0.029836446046829224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,768,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,12288,0.18434666262732613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,12288,0.03533511029349433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,512,0.06578489144643147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,3584,0.020625778370433383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,2048,0.03610488772392273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,3072,0.009967111051082611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,256,0.017703110973040264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,51200,0.04379733403523763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,64,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,768,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,128,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,3584,0.09595021936628555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,2048,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,256,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,8192,0.03366400135887994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,3072,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,2048,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,128,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,2560,0.06589777602089776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,2048,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,64,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,8192,0.011438222395049201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,2048,0.005043555465009478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,1024,0.029534220695495605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,51200,0.06183377901713053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,7168,0.02869422237078349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,64,0.00666222224632899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,12288,0.21710756089952257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,6144,5.991288079155816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,12288,0.2362142139010959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,12288,0.05394044518470764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,12288,0.023027555810080633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,65536,2.4549734327528214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,1536,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,1024,0.017335111896197002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,16384,0.05715555614895291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,32,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,3584,0.04591022266281975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,2048,0.011919111013412476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,1536,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,51200,0.05208622084723579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,3584,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,1024,0.32262221972147626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,256,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,16384,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,12288,1.0271182590060763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,3072,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,5120,0.027395556370417278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,8192,0.3697982099321153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,8192,0.014567111929257711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,65536,0.06810400221082899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,6144,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,512,0.01368711143732071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,4096,0.032855110036002264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,6144,0.03532000051604377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,5120,0.01789155602455139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,5120,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,768,0.02982577681541443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,10240,0.11291200584835476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,7168,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,10240,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,10240,0.017887999614079792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,3072,0.26923465728759766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,2560,0.7638950877719455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,2048,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,768,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,10240,0.03551822238498264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,7168,0.22792977756924102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,51200,0.06562578015857272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,6144,0.11284266577826606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,128,0.015896888242827523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,256,0.02818044357829624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,768,0.023307555251651343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,8192,0.06600799825456408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,16384,0.02142933342191908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,12288,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,1024,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,65536,0.049654222197002836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,2560,0.01850755512714386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,1024,0.05264622304174635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,3584,0.08033421966764662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,65536,0.0950853361023797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,8192,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,2560,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,1536,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,6144,0.00703288863102595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,12288,0.026352889007992212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,5120,0.009480000370078618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,64,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,5120,0.1446417835023668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,5120,0.2939528889126248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,10240,0.027612444427278306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,2048,0.026895999908447266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,65536,0.12679289446936712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,2048,0.00646577775478363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,512,0.02269422180122799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,8192,0.20152621799045137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,64,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,2048,0.0372977786593967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,768,0.026279111703236897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,10240,0.1974488894144694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,51200,0.19701600074768066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,7168,0.2718942165374756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,1024,0.0255422227912479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,65536,0.3754275639851888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,8192,0.029474665721257527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,512,0.010026666853162978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,128,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,51200,0.18448621696896025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,12288,0.01961955593691932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,256,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,4096,0.0239964442120658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,65536,0.6605475743611654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,3584,0.013388444152143268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,7168,0.4054711129930284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,51200,0.10883288913302952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,3584,0.034318221939934626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3072,4096,0.12243199348449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,32,0.0036115555299652945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,7168,0.7185511059231229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,65536,0.06523822413550483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,512,0.04115644428465102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,4096,0.04480266571044922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,2560,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,8192,0.8526879946390787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,6144,0.07130933470196195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,6144,0.036954667833116316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,2560,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,8192,0.02387999991575877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,512,0.018191110756662156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,12288,0.021199999584092036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,10240,0.6638302273220485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,512,0.006737777756320105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,10240,1.3398044374254015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,12288,0.4525369008382161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,7168,0.02834933333926731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,2048,0.023931556277804907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,512,0.012914666699038612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,65536,0.024304888314670984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,6144,0.03758666581577725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,1536,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,3584,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,1536,0.025060445070266724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,16384,0.2183075613445706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,3584,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,4096,0.018570666511853535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,1024,0.006217777729034424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,16384,0.052899556027518384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,5120,1.809893290201823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,768,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,128,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,768,0.003559111307064692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,1536,0.07120444377263387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,768,0.02351911034848955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,1536,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,768,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,128,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,2560,0.01051111105415556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,5120,0.02630222174856398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,1024,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,256,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,512,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,512,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,8192,0.025928000609079998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,16384,0.07941510942247179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,12288,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,4096,0.025348444779713947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,6144,0.024803555674023096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,10240,0.1690195533964369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,256,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,4096,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,5120,0.023572444915771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,6144,0.029968000120586816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,8192,0.01569422251648373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,2560,0.019395555059115093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,65536,0.15360177887810603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,6144,0.013636444177892474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,6144,0.01864177816443973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,8192,0.02472177810139126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,12288,0.39075999789767796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,51200,10.314013163248697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,256,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,16384,0.38111289342244464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,65536,0.0851057767868042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,64,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,2048,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,5120,0.07309866613811918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,65536,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,512,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,12288,0.2251537905799018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,65536,0.1270888911353217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,12288,0.02697511182890998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,6144,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,3072,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,512,0.01221866657336553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,5120,0.062112000253465444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,7168,0.01421511173248291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,4096,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,2048,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,7168,0.019146665930747986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,8192,0.04221422142452664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,1024,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,1536,0.2987324396769206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,512,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,256,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,8192,0.015863999724388123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,2560,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,1536,0.01604711181587643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,6144,0.012655999925401477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,8192,0.025544888443417017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,5120,1.5824124018351238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,256,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,65536,0.9037058088514539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,51200,0.09000000026490952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,2560,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,5120,0.014924445086055331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,512,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,5120,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,5120,0.09648444255193074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,10240,0.013531555732091268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,12288,0.005504000104135937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,256,0.0159288893143336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,512,0.036955555280049644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,65536,0.07653955618540446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,128,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,5120,0.00959911114639706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,128,0.009115555220180089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,16384,0.3062240017784966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,256,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,512,0.020218667056825426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,768,0.012900444368521372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,16384,0.04447022080421448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,5120,0.03218844532966614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,2048,0.00452888881166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,256,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,51200,0.12042044268714057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,51200,0.20361244678497314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,10240,0.021374222305085924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,32,0.013623110949993134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,2560,0.10160799821217854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,4096,0.01809511085351308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,64,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,8192,0.026763555076387193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,128,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,12288,0.016515556308958266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,3072,0.03164800008138021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,128,0.0041573333243529005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,10240,0.04970399869812859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,1024,0.0058542220956749385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,16384,0.05208711160553826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,16384,0.1417751047346327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,1536,0.054923554261525474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,768,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,64,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,7168,0.03587822119394938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,12288,0.017844445175594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,4096,0.016767111089494493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,32,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,2048,0.020961778031455144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,1536,0.015192889504962497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,16384,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,1536,0.04035733474625482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,3072,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,10240,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,3072,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,3584,0.02676533328162299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,3072,0.028182221783532038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,51200,0.2556684547000461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,2048,0.00909511082702213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,8192,0.12701243824428982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,2560,0.026438222991095647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,64,0.0017848888205157386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,6144,0.021146666672494676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,16384,0.03839022252294753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,768,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,16384,0.044065776798460216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,1024,0.0038213332494099936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,3584,0.07319644424650404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,6144,0.023973332511054143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,16384,0.040934221612082586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,128,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,65536,1.54131285349528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,128,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,3072,0.023760888311598036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,256,0.02843377656406826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,7168,0.01942755613062117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,1024,0.060416890515221484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,1536,0.01000088867213991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,64,0.04172799984614054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,32,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,1024,0.05572888586256239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,5120,0.0594080024295383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,16384,0.07773422532611422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,1024,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,12288,0.013463999662134381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,7168,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,3072,0.1867448886235555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,7168,0.030697776211632624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,51200,0.045356445842319064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,3072,0.0069848886794514125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,256,0.03594844539960226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,16384,0.12504355112711588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,512,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,8192,0.13724622461530897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,4096,0.09706488582823013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,1536,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,1024,0.021360889077186584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,6144,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,6144,0.02110133402877384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,12288,0.02985066506597731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,512,0.006364444477690591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,16384,0.0325857765144772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,8192,0.011436444189813403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,32,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,3584,0.028712001111772325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,128,0.03196622265709771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,3584,0.03126577867401971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,5120,0.017853332890404593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,10240,0.06827821996476915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,8192,0.06115200122197469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,51200,1.115696907043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,3584,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,16384,0.016069332758585613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,7168,0.009722666607962714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,256,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,2560,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,5120,0.03442310955789354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,51200,0.04229155513975355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,2560,0.045927110645506114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,256,0.06258844667010836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,16384,0.23994578255547416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,10240,0.0425946679380205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,4096,0.013972444666756524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,12288,0.03875911235809326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,1024,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,12288,0.22023910946316191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,768,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,1024,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,32,0.0020791112134853997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,5120,0.027089777919981215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,51200,0.11430399947696263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,1024,0.0037919998996787598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,512,0.12936177518632677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,2560,0.007831999825106727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,16384,0.6187653541564941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,512,0.05135644475618998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,51200,0.09838310877482097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,2560,0.01347644461525811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,768,0.01202222208182017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,10240,0.2610399987962511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,128,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,6144,0.0188737776544359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,512,0.008583111067612966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,4096,0.015569777952300178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,2560,0.004873777843183941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,768,0.004640000147951974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,1024,0.003428444473279847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,256,0.07583289013968574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,6144,0.021762667430771723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,8192,0.029670221938027277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,7168,0.16906310452355278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,3072,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,128,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,51200,0.1575982173283895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,32,0.20340355237325033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,12288,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,8192,0.01976711054642995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,256,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,1024,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,3072,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,3584,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,128,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,8192,0.018331555856598746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,3584,0.017180444465743173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,2048,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,6144,0.028885334730148315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,65536,6.083980560302734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,8192,0.023936889237827722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,7168,0.01298577752378252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,16384,0.021260443660948012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,65536,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,512,0.008360889222886827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,768,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,1024,0.01678222252262963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,16384,0.021125333176718816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,65536,0.2196728918287489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,7168,0.02369511127471924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,4096,0.028297778632905748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,1024,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,16384,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,7168,0.06959555546442668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,7168,0.015637333194414776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,12288,0.06947910785675049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,5120,0.01532799998919169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,128,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,768,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,8192,0.21572089195251465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,2048,0.0070053330726093715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,5120,0.03320444292492337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,3584,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,128,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,32,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,5120,0.009172444542249044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,3072,0.008024888734022776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,64,0.01552088889810774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,768,0.010292444792058732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,65536,0.24772087732950845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,128,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,2560,0.21024089389377168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,16384,0.03590755661328634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,128,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,4096,0.028189334604475234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,10240,0.02160800000031789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,7168,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,51200,0.060002664724985756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,6144,0.01368799971209632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,8192,0.04134666588571336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,65536,0.08285955588022868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,12288,4.696709526909722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,256,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,6144,0.16123999489678278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,65536,0.13267466757032606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,10240,0.05968622366587321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,128,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,7168,0.1298675537109375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,6144,0.14141333103179932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,3584,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,51200,0.42533244027031797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,1536,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,10240,0.07401511404249403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,12288,0.16227732764350042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,3584,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,65536,0.3579715622795953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,256,0.01734044485621982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,6144,0.025827555192841425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,2048,0.5363546477423774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,768,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,65536,0.15748622682359484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,64,0.013783999615245394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,2560,0.16315200593736437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,256,0.015465777781274585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,32,0.06392355759938557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,128,0.019366222951147292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,256,0.008236444658703273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,10240,0.02478666603565216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,3584,0.013387555877367655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,51200,0.27773067686292857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,6144,0.023689778314696416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,6144,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,2048,0.35793156094021267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,256,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,51200,0.42069511943393284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,128,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,768,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,1024,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,7168,0.026774222652117412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,16384,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,8192,0.04398222102059258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,1536,0.0046977777447965415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,1536,0.017846221725145977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,3584,0.3395688798692491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,1536,0.0064604443808396654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,2048,0.03132799930042691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,16384,0.17392355865902373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,2048,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,768,0.022830223043759663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,32,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,3584,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,12288,0.016507556041081745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,512,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,512,0.06968533330493502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,4096,0.007863110966152614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,2560,0.006727111008432176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,64,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,4096,0.017331555485725403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,768,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,8192,0.03449777762095133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,768,0.054415110084745616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,2048,0.01997511088848114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,256,0.015523556205961438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,12288,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,65536,0.3149493270450168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,256,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,7168,0.012818666795889536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,12288,0.4097591241200765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,1024,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,768,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,3584,0.011047999891969891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,768,0.023235556152131822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,1536,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,2048,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,6144,0.07201955715815227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,512,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,12288,0.20291288693745932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,64,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,5120,0.01387555566098955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,51200,0.03987377882003784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,10240,0.012109333442317115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,3584,0.03969511058595445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,32,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,4096,0.09791022539138794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,6144,0.009538667069541084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,16384,0.02795555525355869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,3072,0.11381955941518147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,3072,0.008185777399275038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,8192,0.034966223769717746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,128,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,65536,0.050276445017920605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,4096,0.088136891523997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,1024,0.010956444674068026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,1536,0.017656000124083627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,12288,0.04781422350141737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,10240,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,256,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,4096,0.009631111390060848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,128,0.014831110835075378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,64,0.005191111316283544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,1024,0.03376177615589566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,256,0.005383111122581694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,3072,0.03642844491534763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,512,0.16401599513159856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,1536,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,2048,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,1024,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,4096,0.014376888672510782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,51200,11.41569349500868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,768,0.01661155621210734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,51200,0.06923200024498834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,64,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,2048,0.004543111142185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,32,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,7168,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,768,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,64,0.09923466708925034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,5120,0.013888888888888888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,5120,0.011206222077210745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,4096,0.02511466708448198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,1024,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,32,0.004537777768241035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,8192,0.0378479990694258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,2560,0.05726222197214762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,64,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,8192,0.0358951124880049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,12288,0.21909688578711617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,2048,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,2560,0.005192000004980299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,4096,0.008072888685597314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,65536,0.19780177540249297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,6144,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,4096,0.03422044383154975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,5120,0.052758223480648465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,32,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,7168,0.28521156311035156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,3072,0.012881778180599213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,51200,0.12241155571407741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,256,0.0020480000724395118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2048,2048,0.0069919998447100324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,1024,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,8192,0.060545775625440806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,51200,0.07189333438873291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,10240,0.015847111741701763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,2560,0.02847466535038418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,51200,0.07947822411855061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,4096,0.023296000228987798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,1536,0.018569777409235638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,4096,0.04380977816051907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,1536,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,16384,0.04279111160172356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,2560,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,7168,0.06115644507937961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,1024,0.006066666709052191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,512,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,2048,0.07623822159237333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,7168,0.021295999487241108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,65536,0.3979359997643365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,12288,0.030601776308483545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,256,0.006681777950790193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,256,0.006890666567616993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,2560,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,2048,0.009069333473841349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,7168,0.03176000052028232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,1024,0.035924444595972695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,12288,0.03771022293302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,2048,0.018615111708641052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,64,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,768,0.03843377696143256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,65536,0.06510400109820895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,16384,0.025920000341203477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,10240,0.7827999856736926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,2048,0.024950222836600408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,2048,0.036855111519495644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,6144,0.06020977762010363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,768,0.0038311113086011675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,6144,0.03036710951063368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,51200,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,256,0.03545777665244208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,7168,0.01129333343770769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,1024,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,12288,0.018590221802393597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,128,0.0041848888827694785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,5120,0.03705688979890611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,16384,0.03564622335963779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,3584,0.6231306923760308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,7168,0.02697333362367418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,1024,0.0068897778789202375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,2048,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,65536,0.5402284198337132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,6144,0.018428444862365723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,6144,0.014575999644067554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,768,0.01089422239197625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,1024,0.03407288922203912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,128,0.11182133356730144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,128,0.08075111442142062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,5120,0.010448888772063784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,64,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,6144,0.03468711177508036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,3072,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,2048,0.01960533360640208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,32,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,51200,0.04036533170276218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,51200,7.111987643771702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,768,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,5120,0.04931555522812737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,12288,0.021864000293943617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,128,0.00566133318675889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,512,0.07434844308429293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,65536,0.018952889574898612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,3072,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,3584,0.02456444501876831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,1024,0.024093333217832778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,2560,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,10240,0.30916889508565265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,5120,0.06775555345747206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,3072,0.009147555463843876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,12288,0.05191377798716227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,1536,0.02696977721320258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,51200,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,128,0.005407111098368962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,2048,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,1536,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,3584,0.08393066459231907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,512,0.005380444642570283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,12288,1.8847423129611547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,1024,0.023230221536424425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,16384,0.1580373313691881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,2048,0.004855999930037392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,512,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,4096,0.014871110518773397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,1536,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,256,0.019483556350072224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,256,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,5120,0.09482489029566447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,3584,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,3584,0.029261334074868098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,3584,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,32,0.11281600263383652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,2560,0.024455999334653217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,1536,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,51200,0.5167662302652994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,4096,0.029331554969151814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,512,0.10952088567945693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,3072,0.1443084478378296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,51200,2.7722515530056424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,768,0.008481777376598781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,10240,0.07106310791439481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,10240,0.032891554964913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,51200,0.13024622864193386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,12288,0.22657333479987252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,12288,0.2781919903225369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,16384,0.020796444680955675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,4096,0.09466222259733413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,3072,0.009940444595283931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,256,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,256,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,1536,0.0035582222044467926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,65536,0.056516442033979625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,512,0.017972444494565327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,10240,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,65536,9.643528408474392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,65536,5.147988637288411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,2560,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,64,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,1536,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,4096,0.03717511230044895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,4096,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,7168,0.03119644522666931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,1024,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,768,0.036398222049077354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,64,0.010676444404655032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,128,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,4096,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,512,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,3584,0.04773244592878553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,16384,0.03416444526778327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,5120,0.04831288920508491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,8192,0.031705776850382485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,1536,0.004086222085687849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,768,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,51200,1.1250133514404297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,12288,0.10486400127410889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,8192,0.03309155503908793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,768,0.05669155385759142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,12288,0.01626755628320906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,6144,0.009954666925801171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,12288,0.035635554128222995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,32,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,12288,0.03117777903874715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,1536,0.053782221343782216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,4096,0.011775111158688864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,6144,0.016319110989570618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,1536,0.003552000141806073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,10240,0.040327111879984535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,1536,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,1024,0.019519110520680744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,3072,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,2560,0.04677777820163303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,12288,0.02462933295302921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,512,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,256,0.006712889091836081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,51200,0.46393420961168075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,7168,0.04569155640072293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,2048,0.007160888777838812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,32,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,1024,0.004514666481150521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,6144,0.009977777798970541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,10240,0.011755555868148804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,512,0.004461333155632019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,32,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,1024,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,1024,0.007495111061467065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,768,0.028245333168241713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,8192,0.058693336115943066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,1536,0.029808001385794744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,2048,0.016936888297398884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,8192,0.04723555511898465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,3072,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,256,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,1024,0.007232888705200619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,8192,0.04424622323777941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,6144,0.016715556383132935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,12288,0.027218666341569688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,4096,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,64,0.012644444074895648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,10240,0.03657955593532986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,4096,0.011874666644467248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,4096,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,3072,0.05004177822007073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,5120,0.03202222122086419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,512,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,2048,0.03185155656602647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,65536,0.18991466363271078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,128,0.001863111017478837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,8192,0.04679288797908359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,2048,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,128,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,3072,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,3584,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,1536,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,4096,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,128,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,5120,0.0682871143023173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,1536,0.017204445269372728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,5120,0.024675556355052527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,51200,0.09606133566962348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,512,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,1536,0.019235556324323017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,3584,0.024650666448805068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,1024,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,65536,1.215703116522895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,8192,0.025660443637106154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,7168,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,3584,0.11540355947282578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,8192,0.5047982003953722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,1024,0.02840177880393134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,512,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,5120,0.011034666664070554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,1024,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,512,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,16384,0.021420444051424663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,3072,0.01166044423977534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,6144,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,16384,0.12221154901716445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,2560,0.09901866647932266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,51200,0.16675910684797499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,64,0.046759112013710864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,2048,0.006717333363162146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,32,0.008579555484983657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,3584,0.027442667219373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,8192,0.04747910963164436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,8192,0.010770666930410596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,512,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,10240,0.014819555812411837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,2560,0.007336888876226213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,256,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,768,0.045259555180867515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,8192,0.07529244158003065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,1536,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,16384,0.022612444228596155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,1024,0.0324924455748664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,256,0.010762666662534079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,4096,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,256,0.0053075556125905775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,256,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,5120,0.00869777798652649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,51200,0.01924622224436866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,2560,0.005640888793600931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,16384,1.1820702022976346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,3584,0.20648533768124053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,51200,0.0529093345006307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,2560,0.07778044541676839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,64,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,7168,0.010325333310498131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,256,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,3584,0.016833777228991192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,5120,0.030418667528364394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,65536,0.13950488302442762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,768,0.0203128887547387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,64,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,3072,0.022289777795473736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,8192,0.019366222951147292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,256,0.1871884398990207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,2048,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,1536,0.0561351113849216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,1024,0.014855111638704935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,32,0.011714666253990598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,4096,0.007756444315115611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,7168,0.03682311044798957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,2560,0.045864890019098915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,4096,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,2560,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,51200,0.3514880074395074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,6144,0.04723733332422045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,32,0.03386222322781881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,2048,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,6144,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,1024,0.06782044304741754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,2048,0.02863466739654541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,2048,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,5120,0.014933332800865173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,128,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,65536,1.539903958638509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,1536,0.028704000843895808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,2560,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,12288,0.030215998490651447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,6144,0.045722666713926524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,12288,0.015517334143320719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,1024,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,12288,0.134716444545322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,16384,0.01015911085738076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,6144,1.4327075746324327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,7168,0.13640444808536106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,12288,0.017009778155220878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,6144,0.012279999752839407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,64,0.003830222205983268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,128,0.0032640000184377036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,65536,0.14507467216915554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,5120,0.04493066668510437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,3584,0.0075599998235702515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,2048,0.017504000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,51200,0.062936888800727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,3584,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,5120,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,4096,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,128,0.03768800033463372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,4096,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,8192,0.014148443937301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,1024,0.008181333541870117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,4096,0.04344088832537333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,1536,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,3584,0.009192888935407003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,2560,0.8702221976386176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,3584,0.040694223509894475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,51200,0.03676177726851569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,256,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,65536,0.8925191031561958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,65536,0.2914222346411811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,32,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,3072,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,64,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,7168,0.03527466787232293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,7168,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,2048,0.004501333253251182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,256,0.016977777083714802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,768,0.05083733465936449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,256,0.02938666608598497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,7168,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,8192,0.08697333600785996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,7168,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,128,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,2560,0.017041777571042378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,4096,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,3072,0.023027555810080633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,1536,0.030836443106333416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,768,0.008940444224410588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,128,0.02162222233083513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,1536,0.020398222737842135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,64,0.005176888985766305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,768,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,1024,0.15718932946523032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,1536,0.22958578003777397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,12288,0.017714665995703805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,768,0.015954666667514376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,768,0.015967999895413715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,3072,0.03221866819593642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,6144,0.08955199850930108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,2048,0.013789333403110504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,65536,0.06772355238596599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,2048,0.03991644581158956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,4096,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,8192,0.004318222403526306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,65536,3.4943919711642795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,768,0.024311999479929607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,4096,0.008034666379292807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,5120,0.015068444940778943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,768,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,1024,0.03376977642377218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,10240,0.03628355595800612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,1536,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,8192,0.022456889351209004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,65536,0.046783109505971275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,12288,0.01924711134698656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,4096,0.07232977946599324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,3584,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,12288,0.5756062401665581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,768,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,5120,0.008937777744399177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,3584,0.23945421642727324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,3584,0.02572622232966953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,768,0.019746666153271992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,2560,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,10240,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,51200,0.02842666705449422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,768,0.03407911128467984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,7168,0.08867644601398045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,768,0.010395555860466426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,1536,0.013412444127930535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,768,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,512,0.0025457777082920074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,12288,0.11015822490056355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,12288,0.13804533746507433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,2560,0.03607911202642653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,65536,0.09754133224487305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,8192,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,6144,0.8077982266743978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,3072,0.007368000017272101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,12288,0.13151555591159397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,512,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,4096,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,256,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,256,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,3584,0.09119200044208103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,7168,0.04928177926275465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,16384,0.04075288772583008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,65536,0.5113849110073513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,16384,0.19807644685109457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,2048,0.006526222245560751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,5120,0.01129244433508979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,4096,0.010619555910428366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,1024,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,2560,0.008143111235565608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,6144,0.03088266650835673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,2048,0.2276408937242296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,256,0.018255111243989732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,3584,0.01538044379817115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,3584,0.018153778380817838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,512,0.010618666807810465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,3072,0.07770311170154147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,12288,0.028902222712834675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,51200,0.3494035667843289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,16384,0.02256177696916792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,1536,0.02118844456142849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,10240,0.0284551117155287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,1024,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,4096,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,8192,0.023697778582572937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,128,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,7168,0.040752000278896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,1536,0.0038604442444112566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,16384,6.362181345621745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,8192,0.05396177702479892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,6144,0.13253777556949192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,6144,0.027041777968406677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,65536,0.5122622383965386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,16384,0.06672710842556424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,64,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,51200,0.8004444440205892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,4096,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,16384,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,2560,0.00981244444847107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,16384,0.02237244447072347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,128,0.01885777711868286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,1536,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,16384,0.0210844443904029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,2560,0.00814577771557702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,8192,0.037296000454160906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,2048,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,6144,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,32,0.04597599969969856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,64,0.025032889511850145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,128,0.015589333242840238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,512,0.007053333024183909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,7168,0.364104880226983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,51200,0.09792621930440266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,3584,0.140608893500434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,5120,0.0349333319399092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,4096,0.01869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,6144,0.03289422061708238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,2560,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,5120,0.08550488948822021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,2560,0.006183111005359226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,8192,0.10177422232098049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,32,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,16384,0.019863110449579027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,10240,0.016502221425374348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,768,0.014892444014549255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,3584,0.0306640002462599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,65536,1.0894506242540147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,256,0.024680000212457445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,1536,0.0038933331767717996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,32,0.03011911114056905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,32,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,2560,0.28919021288553876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,10240,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,2048,0.008625778059164682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,1024,0.06563644276724921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,6144,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,3584,0.12045421865251328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,2048,0.021934222843911912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,1536,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,3072,0.036455111371146307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,6144,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,1024,0.022872888379626807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,3584,0.05671999851862589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,3584,0.044046223163604736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,2560,0.0302782224284278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,6144,0.02535733415020837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,16384,0.12145422564612494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,7168,0.061596446567111544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,4096,0.008452444440788692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,768,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,1024,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,1536,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,6144,0.019271110494931538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,10240,0.023007111416922674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,2048,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,8192,0.043791999419530235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,256,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,1024,0.021350221501456365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,7168,1.1776871151394315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,2048,0.024895111719767254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,1536,0.09062488873799641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,768,0.03150666753451029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,256,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,1536,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,3072,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,51200,0.05162400007247925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,2560,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,32,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,64,0.012776888906955719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,768,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,1536,0.025850666893853083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,5120,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,7168,0.019935111204783123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,8192,0.033244444264305964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,256,0.007711110843552484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,3072,1.2464408874511719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,1024,0.00999466660949919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,65536,0.44582398732503253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,10240,0.21657421853807238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,16384,0.04002755549218919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,12288,0.017657778329319425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,3584,0.014889778362380134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,128,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,51200,0.13360267215304905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,3584,0.08986133337020874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,6144,0.016582222448454965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,2048,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,3072,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,10240,0.017218665944205392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,256,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,2560,0.02608266638384925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,512,0.01260444439119763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,16384,1.4093902375962999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,256,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,768,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,7168,0.01497866710027059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,65536,0.20020088884565565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,1536,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,12288,0.01707466608948178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,64,0.011633777783976661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,2560,0.049775110350714795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,5120,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,1536,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,128,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,1536,0.003663111064169142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,768,0.03105688757366604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,128,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,8192,0.17689243952433267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,3072,0.029185778564876978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,6144,0.12399200598398845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,2048,0.009293333523803288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,1536,0.019448000523779128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,3584,0.3157422277662489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,65536,0.8855448828803169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,8192,0.025768000218603346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,3072,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,8192,0.01089777797460556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,5120,0.013270222478442721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,51200,0.08989422188864814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,12288,0.03462400039037069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,3584,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,512,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,256,0.003903111235962974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,10240,0.1588862207200792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,512,0.015297777122921415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,7168,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,10240,0.04368266794416639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,3072,0.024023110667864483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,5120,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,51200,0.05710311068428887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,5120,0.02306755549377865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,512,0.04012444284227159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,128,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,3072,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,3584,0.05921600262324015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,512,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,2048,0.004528000122971005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,1536,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,1024,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,10240,0.09143377674950494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,2560,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,5120,0.1283173296186659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,16384,0.015000000596046448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,2048,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,128,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,10240,0.2706640031602648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,3584,0.007920000288221572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,1536,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,10240,0.508337762620714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,8192,0.030551999807357788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,3072,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,7168,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,256,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,6144,0.03492000036769443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,16384,0.05153155657980177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,768,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,7168,0.144923554526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,2560,0.046874665551715426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,4096,0.09250577953126694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,5120,0.00904266701804267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,4096,0.009091555244392818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,1536,0.08221244149737887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,65536,0.23192177878485787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,128,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,4096,0.10737244288126628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,16384,0.02128000060717265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,5120,0.03471555643611484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,4096,0.024877332978778418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,4096,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,512,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,32,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,65536,1.4002319971720378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,51200,0.09166133403778076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,128,0.011013333168294696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,7168,0.020669332808918424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,12288,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,256,0.02238133384121789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,256,0.10032000144322713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,2048,0.026122666067547266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,2048,0.036342223485310875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,1536,0.0067528887755341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,2048,0.06425955560472277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,12288,0.0244195560614268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,3584,0.013004444539546967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,2048,0.08243378003438313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,1024,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,5120,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,12288,0.013639999760521783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,256,0.01460800071557363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,3584,0.01443377799457974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,256,0.0021057778762446507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,1536,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,65536,0.2481537659962972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,51200,0.28422310617234975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,16384,2.0904835595024958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,3584,0.011370666325092316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,1024,0.1464666657977634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,3584,0.02427644365363651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,2048,0.020092444287406076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,5120,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,65536,0.047951999637815684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,16384,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,3072,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,768,0.028581334484948054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,256,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,5120,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,1536,0.0124195557501581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,51200,0.6106426450941298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,51200,0.6141662067837185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,51200,0.05367111166318258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,768,0.012424888710180918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,16384,0.02375999920898014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,7168,0.05817777580685086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,7168,0.014481777946154276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,10240,0.03886577818128798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,3584,0.02475999957985348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,1536,0.008040000167157915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,2048,0.007081777685218387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,51200,0.07491822375191583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,6144,0.1276808844672309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,6144,0.009187555147541894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,5120,0.045904888047112354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,2048,0.012364444633324942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,256,0.12148266368442112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,3072,0.052780446079042226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,2048,0.04324977927737766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,51200,0.11604266696506077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,7168,0.024375999967257183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,5120,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,1536,0.01955111159218682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,256,0.2302622265285916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,3072,0.016175111134847004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,3584,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,65536,3.2102595435248484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,768,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,256,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,4096,0.014654222461912366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,12288,0.12801867061191136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,3584,0.007214222517278459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,2560,0.027699554959932964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,128,0.010641777680979835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,6144,0.03304444419013129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,4096,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,4096,0.011884444289737277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,51200,4.654193878173828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,16384,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,4096,0.009473777479595607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,64,0.023016888234350417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,3584,0.010952889091438718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,10240,0.02295733326011234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,16384,0.06901866859859891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,4096,0.007070221834712558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,51200,0.5242346657647027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,51200,1.7174933751424153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,3072,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,16384,0.024899555577172175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,768,0.013731555806265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,3584,0.05055555701255798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,512,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,8192,0.015104888214005364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,8192,0.008266666697131263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,128,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,512,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,3072,0.018554665976100497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,16384,0.1458186705907186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,8192,0.03381777803103129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,1536,0.11399732695685492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,7168,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,256,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,512,0.014276444911956787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,512,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,16384,0.2788311110602485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,256,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,51200,0.1664888858795166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,10240,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,3072,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,512,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,8192,0.019588443968031142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,2560,0.03415200114250183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,16384,0.08535466591517131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,128,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,512,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,1024,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,6144,0.09827555550469293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,1024,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,128,0.010992888775136737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,768,0.03548622131347656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,51200,0.04650755392180549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,4096,0.010633777413103314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,6144,0.1234053373336792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,2560,0.038482666015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,128,0.016120889120631747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,10240,0.015660444895426433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,16384,0.026886221435334947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,128,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,512,0.030827555391523574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,10240,0.047600888543658786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,1536,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,512,0.00700444479783376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,1536,0.004671111288997862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,10240,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,10240,0.8086791038513184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,1536,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,6144,0.011429333024554782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,512,0.02903377678659227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,8192,0.013263999587959714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,64,0.04061955544683669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,128,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,7168,0.016559999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,65536,0.06250222524007161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,5120,0.0629626644982232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,65536,0.17549778355492485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,65536,0.11558755238850911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,1024,0.007559110720952352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,256,0.015100444356600443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,8192,0.050048887729644775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,6144,0.12057333522372776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,16384,0.01737777723206414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,32,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,2048,0.00814222213294771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,5120,0.015508444772826301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,2048,0.0486195551024543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,16384,1.6720150841606989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,1536,0.02141688929663764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,128,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,16384,0.09673155678643121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,512,0.014178666803571912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,768,0.03877777854601542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,3584,0.011972444752852121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,12288,0.022687999738587275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,3072,0.012808889150619507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,51200,0.09472088681326972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,12288,0.13209777408176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,8192,0.31211643748813206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,1536,0.04429777794414096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,8192,0.012784000072214337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,12288,0.038867556386523776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,32,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,6144,0.028004444307751123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,768,0.01648888819747501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,6144,1.107192039489746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,8192,0.5177857610914443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,7168,0.027152889304690894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,768,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,512,0.013738666971524557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,256,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,65536,0.1302515533235338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,16384,0.020911999874644812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,128,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,5120,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,3584,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,64,0.0017768889665603638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,10240,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,12288,0.8813235494825574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,8192,0.015020444989204407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,2560,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,12288,0.020302222834693063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,3584,0.0785048868921068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,65536,0.1577377716700236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,12288,0.050400889582104154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,1024,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,128,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,16384,0.08159199688169691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,512,0.0073928890956772705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,4096,0.01720888912677765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,1536,0.009879111415810054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,51200,0.1724720001220703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,2560,0.020563556088341605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,7168,0.01182044463025199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,3072,0.032567110326555036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,16384,0.06444799900054932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,768,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,256,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,2048,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,2048,0.09371911154852973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,6144,0.053483555714289345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,512,0.010302222437328762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,16384,0.01752800080511305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,128,0.012313777373896705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,768,0.0038408889538711975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,12288,0.07803022199206881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,256,0.013643555343151093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,51200,0.19727555910746256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,16384,0.021760889225535925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,7168,0.019871110717455547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,7168,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,512,0.029993779129452173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,5120,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,10240,0.012890666723251343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,5120,0.029615110821194116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,768,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,256,0.004530666602982415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,3584,0.10196799702114528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,2560,0.020244444410006206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,10240,0.01605422298113505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,3072,0.012892444100644855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,65536,0.15812444686889648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,16384,0.0514977773030599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,4096,0.09462400277455647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,256,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,256,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,3584,0.01370844410525428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,4096,0.015442666080262927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,1536,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,128,0.03226222263442145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,3584,0.025632888078689575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,64,0.0038782221575578055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,256,0.020727111233605277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,65536,0.027462222509913977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,16384,0.29961689313252765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,512,0.18883821699354383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,512,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,7168,0.021181333396169875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,768,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,12288,0.10423377487394546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,128,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,256,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,7168,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,16384,0.1891982158025106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,10240,0.05278933379385206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,51200,0.3919546604156494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,4096,0.03628799981541104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,5120,0.01622577839427524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,1024,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,512,0.03646666804949442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,256,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,5120,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,3584,0.05061422122849358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,5120,0.019276445110638935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,5120,0.34465689129299587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,5120,0.0385537776682112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,2560,0.19087555673387316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,51200,0.132240891456604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,4096,0.022670222653283015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,256,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,7168,0.18348266018761528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,65536,0.06685333119498359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,768,0.025978666212823655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,3072,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,512,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,2560,0.06887377632988824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,8192,0.7483466466267904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,10240,0.06506399975882636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,3072,0.019183999962276883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,10240,0.020216888851589628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,2560,0.0221733334991667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,5120,0.025048888391918604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,256,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,5120,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,65536,0.09250133567386204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,4096,4.096982320149739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,2048,0.024783111280865137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,8192,0.024263999528355066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,8192,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,5120,0.022082666556040447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,51200,0.13471199406517878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,6144,0.009280888570679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,2560,0.026126222478018865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,64,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,51200,0.09535200066036648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,1024,0.0672995580567254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,3584,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,128,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,10240,0.04602844516436259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,12288,0.06780178017086454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,2560,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,512,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,5120,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,512,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,3072,0.02645333276854621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,32,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,4096,0.009087111386987898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,2048,0.019299555155966017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,256,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,6144,0.0229404436217414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,51200,0.10330933332443237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,5120,0.06945421960618761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,512,0.012777778009573618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,64,0.01420088940196567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,3584,0.00777422222826216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,768,0.09096533060073853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,51200,0.22144888506995308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,4096,0.022945778237448797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,51200,0.05130399929152595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,10240,0.04617866542604235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,128,0.010055111514197456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,64,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,2048,0.01369244439734353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,768,0.007430222299363878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,128,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,256,0.012717333104875354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,51200,1.8354596032036676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,1536,0.06509421931372748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,1536,0.009257777697510189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,65536,0.07950222492218018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,256,0.029453333881166246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,512,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,768,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,16384,0.19829421573215059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,12288,0.5182204246520996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,8192,0.1768204503589206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,256,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,12288,0.10630310906304254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,51200,0.3672364552815755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,65536,0.4522906409369574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,128,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,768,0.016366221838527255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,65536,0.0841475526491801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,3072,0.07440088854895698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,3584,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,3072,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,768,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,65536,1.355484432644314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,768,0.007762666377756331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,1024,0.005970666805903117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,8192,0.05355911122428047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,65536,0.07206044594446818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,4096,0.08613422181871201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,512,0.015569777952300178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,12288,0.21156977282630074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,768,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,5120,0.046803557210498385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,3584,0.052375111314985484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,768,0.012855110896958245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,12288,0.021944888763957556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,256,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,2048,0.01830044388771057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,3072,0.012713777522246042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,8192,0.09070755375756158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,2560,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,1024,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,65536,0.18557600180308023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,5120,0.030663109487957422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,65536,0.14976444509294298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,7168,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,8192,0.013080000049538083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,6144,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,6144,0.061087999078962535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,10240,0.032115555471844144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,1536,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,10240,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,6144,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,8192,0.1798106696870592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,65536,0.9695155885484484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,16384,0.28734755516052246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,65536,2.5446879069010415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,4096,0.07813866933186848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,65536,0.1758373313479953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,5120,0.023360000716315374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,65536,0.5854133499993218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,8192,0.05797688828574287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,51200,0.06449866957134671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,7168,0.04629688792758518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,512,0.01958666741847992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,3584,0.01851466629240248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,256,0.01724088854259915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,1536,0.07969866858588324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,1024,0.006735999965005451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,12288,0.14692621760898167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,512,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,16384,0.065738664733039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,1536,0.01872977779971229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,10240,0.02946399980121189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,4096,0.13693777720133463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,3584,0.021701332595613267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,512,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,8192,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,3072,0.09632355637020534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,12288,0.024312888582547505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,3072,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,6144,0.010055999788973067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,5120,1.5844489203559027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,10240,0.08630666467878555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,1024,0.045379555887646146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,256,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,256,0.03423822257253859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,6144,0.17744089497460258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,51200,0.15509066316816542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,4096,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,2560,0.012506666282812754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,51200,0.03897155655754937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,16384,0.12540622552235922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,65536,0.11405955420600043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,5120,0.20624177985721162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,4096,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,4096,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,128,0.009562667045328353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,1536,0.02343911098109351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,6144,0.02737866673204634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,64,0.014574221438831754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,2048,0.02884711159600152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,12288,0.11807200643751357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,3584,0.006800000038411882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,256,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,5120,0.009720889230569204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,1024,0.02081155611409081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,3584,0.02757155564096239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,128,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,1024,0.022667555345429316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,5120,0.009392889009581672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,4096,0.023403555154800415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,51200,0.06052355633841621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,1536,0.17692355314890543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,128,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,2048,0.031607998741997614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,768,0.1920106675889757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,1536,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,12288,0.04024089044994778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,6144,0.1127857764561971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,7168,0.02309155629740821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,7168,0.05648710992601183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,16384,6.224848853217231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,7168,0.5048711034986708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,6144,1.0348506503634982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,65536,0.17046222421858045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,12288,1.2630799611409504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,1536,0.07730133003658719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,10240,0.4278417693244086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,3584,0.008232889076073965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,2560,0.15281689167022705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,51200,0.9010622236463758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,12288,0.04743644595146179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,3584,0.0422417786386278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,51200,0.6084071265326606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,12288,0.03010666701528761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,1536,0.02387733260790507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,256,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,5120,0.046690666013293795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,3584,0.0075795559419526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,768,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,6144,0.046987556748920016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,1024,0.03484977616204156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,7168,0.09331644243664211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,51200,0.07980622185601129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,768,0.036421332094404436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,128,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,5120,0.0728728903664483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,51200,0.055580443806118436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,16384,0.09568622377183701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,4096,0.010493333141009012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,64,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,1024,0.03590755661328634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,2048,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,128,0.017610667480362784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,4096,0.008756444685988957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,4096,0.05834488736258613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,65536,1.748151143391927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,8192,0.04279999931653341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,10240,0.01871466636657715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,12288,0.27898044056362575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,4096,0.008429333567619324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,51200,0.0710080001089308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,3072,0.04756177796257866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,128,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,1024,0.01587466730011834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,1024,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,5120,0.011401777466138205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,10240,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,2048,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,1024,0.025067556235525344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,16384,0.29418222109476727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,2560,0.02630666726165348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,10240,0.009718221922715506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,3584,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,12288,0.03247288862864176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,51200,0.12532622284359402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,10240,0.02123466630776723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,768,0.023170666562186346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,7168,0.027748444014125403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,3072,0.017623111605644226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,16384,0.8136800130208334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,768,0.24976534313625762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,8192,0.016583111551072862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,2560,0.008721777962313758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,12288,0.024520888924598694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,3072,0.015784889459609985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,12288,0.07048710849550036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,4096,0.041434665520985924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,256,0.00998488896422916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,768,0.025008888708220586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,16384,0.2060204479429457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,65536,0.06394755840301514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,16384,0.392955568101671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,32,0.0019040000107553268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,1536,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,12288,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,1024,0.00600355573826366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,768,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,3584,0.012905778156386482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,3072,0.0431235565079583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,128,0.014497778481907315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,768,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,1024,0.007886221839321984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,2560,0.039572444227006696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,6144,0.032045334577560425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,2560,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,1024,0.009741333623727163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,2048,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,64,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,4096,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,3072,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,2560,0.033067554235458374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,3584,0.02296000056796604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,5120,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,2560,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,16384,0.030790223015679255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,10240,0.030407110850016277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,4096,0.012717333104875354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,10240,0.10257689158121745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,4096,0.09907733069525824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,7168,0.01959288948112064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,64,0.011935110721323224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,128,0.0462026662296719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,6144,0.08896266751819187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,8192,0.15435732735527888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,3072,0.1126284466849433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,65536,0.06576622194714017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,6144,0.01759022143152025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,768,0.015558222929636637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,7168,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,1536,0.006800888727108638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,256,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,7168,0.2925022178226047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,1536,0.028379556205537584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,64,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,64,0.02366399930583106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,65536,0.34676265716552734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,768,0.023896000451511804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,768,2.593329747517904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,16384,0.17325866222381592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,2560,0.05446933375464546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,7168,0.05084177851676941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,7168,0.010055999788973067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,4096,0.03260177705023024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,256,0.27236888143751353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,1024,0.0032195556494924757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,51200,0.17520266109042695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,512,0.013761777844693927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,32,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,768,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,5120,0.010233778092596266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,256,0.011295110815101199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,12288,0.03926577832963731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,128,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,12288,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,4096,0.013507555756303998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,8192,0.010295111272070143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,51200,0.4250640074412028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,3584,0.046006222565968834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,10240,0.023552000522613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,256,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,8192,0.022462222311231825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,128,0.001869333287080129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,51200,0.4359537760416667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,512,0.011544889046086205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,51200,0.022270222504933674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,768,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,51200,0.5209369129604763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,6144,0.04136711027887132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,512,0.009854222337404886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,8192,0.011075555450386472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,10240,0.09811911318037246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,12288,0.05750399827957153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,7168,0.039621333281199135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,512,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,4096,0.03100444542037116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,5120,0.11266756057739258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,10240,0.043087111579047314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,768,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,1536,0.018704889549149405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,12288,0.048673778772354126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,768,0.006383111079533895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,3072,0.010592888626787396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,3072,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,512,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,16384,0.19590222835540771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,128,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,1024,0.01240711079703437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,2048,0.005124444348944558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,1024,0.024734222226672705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,12288,0.04795555604828728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,5120,0.18830488787757027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,5120,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,2048,0.3656933307647705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,256,0.08299999766879611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,2560,0.00869511150651508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,16384,0.06432266367806329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,128,0.009444444543785518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,5120,0.023141332798533972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,256,0.021195555726687115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,1536,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,128,0.03323200013902452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,65536,0.12370488378736709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,32,0.0021111110432280437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,8192,3.5121430291069875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,128,0.03241599930657281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,64,0.03216977914174398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,6144,0.15356711546579996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,2048,0.007115555306275685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,10240,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,32,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,12288,0.04049066702524821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,1536,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,512,0.02925866511132982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,3072,0.05653866794374254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,65536,2.7099342346191406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,65536,0.08200622267193265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,512,0.13623021708594427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,1536,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,768,0.0041911109454101985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,768,0.004467555632193883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,64,0.06319822205437554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,65536,2.8249715169270835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,51200,0.12167644500732422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,3072,0.05359377794795566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,3584,0.020949333906173706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,6144,0.06651466422610812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,7168,0.01261333293384976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,2560,0.039055999782350324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,1024,0.015822221835454304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,2048,0.0070053330726093715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,16384,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,8192,0.01149511088927587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,65536,0.11922578016916911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,6144,0.11065600315729777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,128,0.014179555906189812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,1536,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,3584,0.022812444302770827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,65536,0.13518844710456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,12288,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,64,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,256,0.005524444497293896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,10240,0.1503431134753757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,1536,0.05653511153327095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,4096,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,5120,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,10240,0.17114667097727457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,1536,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,64,0.018986667195955913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,2048,0.125453339682685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,8192,0.022796443767017786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,1024,0.016395555602179635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,512,0.016872000363137987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,256,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,3584,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,1024,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,1024,0.0032764443506797156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,256,0.05376266770892673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,512,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,128,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,64,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,7168,0.04430133435461256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,1024,0.01608444419172075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,512,0.0599857767422994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,65536,0.06116977665159437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,128,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,6144,0.03921155465973748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,4096,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,256,0.022666666242811415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,8192,0.05955466959211561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,128,0.021634666456116572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,5120,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,10240,0.03560444381501939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,6144,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,5120,0.029892444610595703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,4096,0.019558222757445443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,16384,0.05408711234728495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,12288,0.2790613439348009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,1536,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,7168,0.020120888948440552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,1024,0.0034977777136696708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,1536,0.01351999988158544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,32,0.047954665289984814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,256,0.002190222135848469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,1024,0.009768889182143742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,768,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,3072,0.12343377537197536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,51200,0.39205333921644425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,10240,0.029211554262373183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,8192,0.08620444271299575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,2560,0.8203795750935873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,512,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,2560,0.037736889388826154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,512,0.03354310989379883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,6144,0.6914284494188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,1536,0.014562666416168213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,2048,0.11078577571445042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,768,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,1536,0.04851822058359782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,65536,0.07256622446907891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,3072,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,64,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,512,0.04667199982537163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,3584,0.04917777909172905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,3584,0.6124337514241537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,256,0.0022213332768943575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,12288,0.014179555906189812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,1024,0.020948444803555805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,64,0.011935999823941125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,5120,0.07307110892401801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,32,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,4096,0.02903911140229967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,768,0.01696888936890496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,768,0.00812533332241906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,16384,0.018619555566045973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,7168,0.22913776503668892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,10240,0.704528013865153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,4096,0.08305155568652682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,5120,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,3072,0.05611911084916857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,10240,0.9459671444363065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,12288,0.06725155644946627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,6144,0.7181173430548774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,65536,0.18096444341871473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,6144,0.013088888592190213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,1024,0.023064000738991633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,128,0.002516444358560774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,768,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,512,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,16384,0.02172800070709652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,768,0.021272889441914026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,12288,0.013809777796268463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,65536,0.057557331191168896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,5120,0.05818577607472738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,16384,0.04090844591458639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,6144,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,128,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,10240,0.04519733455446032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,1024,0.007337777978844113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,1024,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,8192,0.21274133523305258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,12288,0.013756444056828817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,5120,0.0041928887367248535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,256,0.05675644344753689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,51200,0.05469510952631632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,64,0.002473777780930201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,1536,0.09862666659884983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,3584,0.04478311207559374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,3072,0.035641776190863714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,51200,13.38890160454644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,2048,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,16384,0.26266490088568795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,2560,0.02148355543613434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,768,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,1024,0.007879111501905654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,10240,0.016559110747443307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,3584,0.14641422695583767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,512,0.0024560000747442245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,6144,0.030066665675905015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,2048,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,16384,0.5886364512973362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,128,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,4096,0.020222221811612446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,6144,0.06678044133716159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,2048,0.021596444977654353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,32,0.01145244472556644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,5120,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,1536,0.027837332752015855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,768,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,4096,0.07147644625769721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,6144,0.06611466407775879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,1536,0.003910222401221593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,1536,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,6144,0.02660711109638214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,1024,0.11251644293467204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,256,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,1536,0.02035466664367252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,256,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,8192,0.10417155424753825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,32,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,128,0.02295911146534814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,16384,0.30943200323316783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,256,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,51200,65536,0.9547866185506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,6144,0.04247822364171346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,12288,0.051691555314593844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,2048,0.0292871097723643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,32,0.005295111073387994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,16384,0.02537422213289473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,3072,0.015750222735934787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,512,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,51200,0.07087110810809664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,2048,0.009633777870072259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,51200,0.05680977635913425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,6144,0.15091644393073186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,3072,0.016172443826993305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,32,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,3072,0.01555022266176012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,7168,1.04311646355523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,16384,0.13982044325934515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,4096,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,4096,0.04620888829231262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,1536,0.0047004446387290955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,16384,0.053061332967546254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,512,0.044181333647833936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,3584,0.0075395554304122925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,128,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,16384,3.5442551506890188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,1024,0.00599644457300504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,1024,0.029393778906928167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,512,0.019561777512232464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,5120,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,51200,0.5699715614318848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,768,0.021987555755509272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,768,0.004516444272465176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,10240,0.02546399997340308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,768,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,64,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,3072,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,1536,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,7168,0.011984888878133563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,8192,0.14228355884552002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,7168,0.04453955425156487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,8192,0.2548302279578315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,2560,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,4096,0.051797333690855235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,8192,0.02018311123053233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,64,0.0038240001433425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,8192,0.01573333309756385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,1024,0.0063582224150498705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,12288,0.025274667474958632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,256,0.024970667229758367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,1024,0.0029031110720502008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,5120,0.03176000052028232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,10240,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,16384,0.02072266737620036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,4096,0.029496888319651287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,768,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,16384,0.057555556297302246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,1536,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,4096,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,256,0.020979555116759405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,12288,0.06874577866660224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,512,0.01348888874053955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,1536,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,2560,0.07485689057244195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,3584,5.420764499240452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,12288,0.023906666371557448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,5120,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,6144,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,3584,0.02991466720898946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,10240,0.01796888808409373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,64,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,1536,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,1024,0.03113422128889296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,1024,0.01610488858487871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,2560,0.430623107486301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,2048,0.09851733181211685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,10240,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,3584,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,8192,0.015014222926563688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,2048,0.17108621862199572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,128,0.008595555192894405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,32,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,2048,0.01980533368057675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,51200,1.132388432820638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,256,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,4096,0.02537511123551263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,512,0.004059555629889171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,3072,0.00963288876745436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,128,0.14975999461279974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,2048,0.05004000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,32,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,2048,0.014718222949239941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,4096,0.03091288937462701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,32,0.07600000169542101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,512,0.008385777473449707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,1536,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,65536,0.17662489414215088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,65536,0.13164710998535156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,2560,0.015637333194414776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,3072,0.03440711233350966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,2560,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,2048,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,7168,0.027294221851560805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,2048,0.007678222325113084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,5120,0.553244431813558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,512,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,4096,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,8192,0.03930666711595323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,16384,0.10493333472145928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,10240,0.1511075496673584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,16384,0.03664799862437778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,12288,0.04922755559285482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,8192,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,12288,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,2048,0.020225778222084045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,65536,0.07068800263934664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,10240,0.0583404435051812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,256,0.0042071110672420925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,12288,0.016889777448442247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,128,0.024173332585228816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,256,0.023027555810080633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,7168,0.06295288933648004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,128,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,1024,0.15147822433047825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,8192,0.03463110989994473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,7168,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,2048,0.05199733376502991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,3072,0.031123555368847314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,1024,0.005635555419656966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,5120,0.03442400031619602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,512,0.005433777968088786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,256,0.01687466601530711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,5120,0.03316088848643833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,32,0.0018648888087934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,12288,0.12446755833095974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,512,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,51200,1.1651715172661674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,65536,0.21931644280751547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,1024,0.019581332802772522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,4096,0.0726426641146342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,16384,2.77819463941786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,3072,0.28013955222235787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,128,0.020224000016848247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,16384,5.373327891031901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,12288,0.013487999637921652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,16384,0.13845955000983343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,6144,0.015011555618709989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,6144,0.00903911143541336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,10240,0.054304000404145986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,10240,0.03985155622164408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,16384,0.23305598894755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,128,0.025792888469166223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,64,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,12288,0.05935644441180759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,65536,2.49869516160753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,1536,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,2048,0.04661688870853848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,256,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,16384,0.31541599167717826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,2560,0.4046364360385471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,768,0.0032115555885765287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,3072,0.06858844227261014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,16384,0.0372284452120463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,5120,0.10585511393017238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,16384,0.3814693291982015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,3584,0.010864888628323873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,32,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,1536,0.011729777687125735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,3584,0.018538667096032035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,512,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,16384,0.15717155403561062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,12288,0.15932710965474448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,3072,0.13860177993774414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,256,0.020729778541458976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,32,0.014077333940400017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,1024,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,16384,0.06528888808356391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,256,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,65536,0.05825777848561605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,5120,2.003040101793077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,2048,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,32,0.04014399978849623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,10240,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,51200,5.668800777859158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,10240,0.14977422025468615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,7168,0.15945600138770208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,768,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,2560,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,5120,0.010999999940395355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,256,0.04691466689109802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,65536,0.06618666648864746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,10240,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,32,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,3584,0.1732515494028727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,7168,0.029836446046829224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,12288,0.21182666884528267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,4096,0.025220443805058796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,2048,0.03807288739416335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,3072,0.0713662240240309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,7168,0.02369333306948344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,65536,0.05724977784686618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,32,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,10240,0.023216888308525085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,7168,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,51200,28.141526963975693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,2560,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,10240,0.011917333636018964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,1536,0.020215999748971727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,3584,0.02276977731121911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,51200,0.10099199745390151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,51200,10.891837226019966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,3584,0.0203457772731781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,512,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,3072,0.022450667288568284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,3072,0.05330311258633932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,128,0.07698844538794623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,1536,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,768,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,12288,0.04431110951635572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,128,0.014841778410805596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,51200,0.28027555677625865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,512,0.006118222243256039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,10240,0.13917156060536703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,3584,0.06551733281877306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,32,0.006060444646411472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,5120,0.01331999980741077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,65536,0.0873475538359748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,10240,0.08435289065043132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,16384,1.0081937577989366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,1024,0.031200889084074233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,64,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,4096,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,16384,0.3906702200571696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,256,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,7168,0.015463110473420886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,16384,0.048058665461010404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,2048,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,128,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,2560,0.0074542222751511475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,51200,0.2225511074066162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,5120,0.04923911227120293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,6144,0.08420888582865398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,2560,0.03145510951677958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,128,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,5120,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,128,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,5120,0.008896888958083259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,8192,0.032448000378078885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,65536,0.11183378431532119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,768,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,768,0.030935999419954088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,8192,0.5252701971266005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,512,0.020560888780487906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,51200,0.07195733653174506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,4096,0.008358221915033128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,7168,0.01132266637351778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,2048,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,7168,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,1024,0.04607377780808342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,64,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,512,0.05864622195561727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,2048,0.006559999866618051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,128,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,4096,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,8192,0.05487199955516391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,16384,1.1535537507798936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,1536,0.012755555411179861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,7168,0.02709688908523983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,3072,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,10240,3.3412231869167752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,7168,0.019705777366956074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,1536,0.02485688858562046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,8192,0.0818835563129849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,3072,0.007786666353543599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,3072,0.025498666697078284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,256,0.033245331711239286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,2560,0.02106222179200914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,2560,0.01942755613062117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,2048,0.1168737808863322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,768,0.034512887398401894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,6144,0.0376835564772288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,2560,0.010168000227875179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,1536,0.007419555551475949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,512,0.02959377898110284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,5120,0.008887111312813228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,7168,0.040928890307744346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,6144,0.05296088920699226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,4096,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,1024,0.02593244446648492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,7168,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,256,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,6144,0.1557991107304891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,12288,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,12288,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,2560,0.6354879803127712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,12288,0.06980088684293959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,8192,0.04251199960708618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,2560,0.0844915575451321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,10240,0.043190220991770424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,4096,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,5120,0.015557333827018738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,1024,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,768,0.0038453332251972626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,2048,0.01884622209601932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,65536,0.05943466557396782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,256,0.020647111866209242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,256,0.0075777777367168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,64,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,1024,0.01332711097266939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,3584,0.016179554992251925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,2560,0.017165333032608032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,2560,0.015249777171346875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,51200,0.4126799901326497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,10240,0.15098578400082058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,32,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,3072,0.04840711090299818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,256,0.02016088863213857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,1024,0.03801777627733018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,5120,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,65536,1.9220248328314886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,12288,0.20973777770996094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,512,0.004549333204825719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,1536,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,1024,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,2048,0.030368000268936157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,768,0.01628266606065962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,5120,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,51200,0.16411733627319336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,16384,0.14912711249457464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,2048,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,768,0.18772355715433756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,64,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,3072,0.025662221842341956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,3584,0.03267466690805223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,6144,0.9140844345092773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,1024,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,32,0.0028044444819291434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,1536,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,6144,0.11458310816023086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,128,0.003160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,8192,2.307596418592665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,4096,0.08728977706697251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,51200,0.5472488933139378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,256,0.02093244426780277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,512,0.020310221446885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,8192,0.03074044320318434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,256,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,10240,0.01942222151491377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,2048,0.03874666823281182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,1024,0.03873422079616123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,6144,0.06445511182149251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,51200,0.010065777434243096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,2560,0.06788711415396796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,2560,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,256,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,12288,4.025338490804036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,256,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,16384,0.6971724298265246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,32,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,5120,0.13945244418250188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,12288,0.019464888506465487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,512,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,128,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,2560,0.12851733631557888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,256,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,10240,0.061174220508999295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,10240,0.09128355317645603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,32,0.011393778026103973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,6144,0.0617555578549703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,3072,0.07511911127302381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,4096,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,51200,0.06963022549947102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,8192,0.02344888945420583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,2560,0.006849777781301075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,512,0.010055999788973067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,65536,0.0722400016254849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,3072,0.004846222284767362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,1024,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,1024,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,5120,0.04494755466779073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,1024,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,3584,0.014554666148291694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,2560,0.01812799937195248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,3584,0.00778311077091429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,3584,0.11831822660234238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,6144,0.015626667274369132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,512,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,5120,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,1024,0.004510222209824456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,1536,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,5120,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,128,0.01923822197649214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,6144,0.05203022228346931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,10240,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,128,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,1536,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,3584,0.004859555512666702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,7168,0.3362826771206326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,7168,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,4096,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,16384,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,8192,0.02402488887310028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,12288,0.015560888581805758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,2560,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,768,0.007536888950400882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,7168,0.3057404359181722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,4096,0.015994666351212394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,128,0.01942933268017239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,8192,0.05303288830651177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,768,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,3072,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,256,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,768,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,51200,0.07650399870342679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,7168,0.024530667397711012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,12288,0.12625243928697374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,2560,0.0634488860766093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,51200,0.12017422252231175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,128,0.021176889538764954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3584,6144,0.051613334152433604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,7168,0.07974755764007568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,512,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,51200,0.050019555621676974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,65536,0.24805866347418892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,512,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,10240,0.013536888692114087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,1536,0.032744000355402626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,6144,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,128,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,16384,0.5196453200446235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,10240,0.01587022178702884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,256,0.0069022224181228215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,10240,0.025690666503376428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,8192,0.11564977963765462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,51200,1.5241751141018336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,10240,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,51200,0.05632088581720988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,1024,0.03245688809288873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,2560,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,6144,0.020434666011068556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,512,0.019318222999572754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,128,0.014977777997652689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,128,0.14034844769371882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,65536,0.07318311267428927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,512,0.0027955555253558685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,51200,0.06042666567696465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,2560,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,7168,0.020655999581019085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,7168,0.12206133206685384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,10240,1.449626710679796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,4096,0.055495109823015004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,65536,0.6017244656880697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,7168,0.03453333179155985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,65536,0.18256889449225533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,1536,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,512,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,6144,0.055142223834991455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,12288,0.032371554109785296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,5120,0.009778666827413771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,5120,0.08284978071848552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,5120,0.04436444573932224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,2048,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,6144,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,3584,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,64,0.003539555602603488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,7168,0.11712533897823757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,1024,0.01927288870016734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,65536,0.0653182201915317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,512,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,51200,0.06145155429840088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,768,0.008561777571837107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,256,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,4096,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,8192,0.0366195539633433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,4096,0.025817778375413682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,51200,0.14578933186001247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,7168,0.11152977413601345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,4096,0.1582231124242147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,65536,0.05248444279034933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,1024,0.00869599978129069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,128,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,65536,0.08630844619539048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,8192,0.044063998593224414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,16384,0.04192622171507942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,51200,0.07606933514277141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,1024,0.0031537777847713898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,128,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,1536,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,65536,0.06136888927883572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,3072,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,3584,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,4096,0.09436621930864121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,512,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,1024,0.04578488734033373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,2560,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,2560,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,3584,0.10087288750542535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,1024,0.03701511025428772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,3072,0.023687111006842718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,16384,0.37746310234069824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,8192,0.024995555480321247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,8192,0.13007556067572698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,65536,0.23757243156433105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,2048,0.06785955693986681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,6144,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,3584,0.04381066560745239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,8192,0.024921778175565932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,128,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,3584,0.008219555848174626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,51200,0.3511386712392171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,3072,0.48276175392998594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,256,0.01705333259370592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,768,0.02411199940575494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,32,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,2560,0.02772088845570882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,7168,0.03751644492149353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,3584,0.10480799939897324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,12288,0.6122639973958334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,1024,0.006405333264006509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,4096,0.018052443861961365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,1024,0.026135111848513286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,4096,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,7168,0.09180355734295315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,128,0.013450666434235044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,5120,0.12778577539655897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,256,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,256,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,768,0.02548444436656104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,16384,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,3584,0.010507555471526252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,5120,0.010995555255148145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,256,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,5120,0.07677421967188518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,128,0.13922488689422607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,65536,0.1368968884150187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,12288,0.22500533527798125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,3584,0.022993778189023335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,512,0.06307999955283271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,3072,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,5120,1.7456604639689128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,7168,0.020273778173658583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,4096,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,128,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,8192,0.011310222248236338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,12288,0.17622221840752494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,4096,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,8192,0.05409333440992567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,1024,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,8192,0.023349333140585158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,3584,0.03494222296608819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,65536,0.06989422109391955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,7168,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,128,0.005432888865470886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,7168,0.08983733256657918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,4096,0.07322844531801012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,10240,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,51200,0.06006311045752632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,5120,0.039686222871144615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,128,0.009368889033794403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,64,0.005631111148330901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,51200,96.5036349826389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,128,0.08361955483754475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,65536,0.4687022103203668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,2048,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,2560,0.007840889195601145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,5120,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,2048,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,32,0.005865777946180767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,65536,0.06194400125079685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,2048,0.010961777634090848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,65536,0.16559555795457628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,256,0.011353777514563667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,2560,0.0218248897128635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,4096,0.07026400168736775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,256,0.015994666351212394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,1024,0.020451555649439495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,16384,0.020482665962643094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,512,0.005524444497293896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,3072,0.00814133303032981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,5120,0.08895022339291042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,12288,0.060368888907962374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,8192,0.05867999792098999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,1536,0.0075342224703894714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,16384,0.696904871198866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,64,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,1536,0.03252622154023912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,3072,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,1024,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,2048,0.009494221872753566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,51200,0.8065155347188314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,5120,0.11715822749667698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,4096,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,65536,0.11264888445536296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,7168,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,3584,0.01847644481394026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,512,0.03996977872318692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,51200,0.13909955819447836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,1024,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,128,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,1024,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,12288,0.016417778200573392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,1536,0.4332711166805691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,3072,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,2048,0.05897421969307793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,12288,0.09377777576446533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,1024,0.042266666889190674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,2560,0.04000711109903123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,1024,0.08521066771613227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,2048,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,8192,0.3730897903442383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,128,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,7168,0.0085235552655326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,3584,0.07141688797208998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,4096,0.0801102254125807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,3072,0.018185777796639335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,32,0.0017884444031450485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,4096,0.007374222079912822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,5120,0.12788355350494385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,51200,0.05782044596142239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,8192,0.06734489070044623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,1536,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,768,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,6144,0.00573955559068256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,65536,0.23820355203416613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,256,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,16384,2.036407046847873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,256,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,3072,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,1024,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,16384,0.25659288300408256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,3072,0.04238311118549771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,2560,0.12362310621473525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,1536,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,3072,0.08070666922463311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,3072,0.010582221878899468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,1024,0.012732444538010491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,12288,0.0236488895283805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,8192,0.04384444488419426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,5120,0.008572444319725037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,128,0.020409777760505676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,128,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,3072,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,65536,0.6602915657891167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,1536,0.014279999666743808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,2560,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,5120,0.07803999715381198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,10240,0.015983111328548856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,51200,0.024238222175174292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,512,0.015524443652894763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,65536,0.1694942182964749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,512,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,2560,0.05667733483844333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,65536,0.06735377841525607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,2560,0.021147555775112573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,64,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,16384,0.09147110912534927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,5120,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,5120,0.020471110939979553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,4096,0.015322667029168872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,16384,0.03234044379658169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,65536,0.09106489022572835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,512,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,8192,0.04491466614935133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,1536,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,16384,0.7050426801045736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,2560,0.0203128887547387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,3072,0.02555911077393426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,768,0.004839111119508743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,2048,0.14941510889265272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,256,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,10240,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,3584,0.011765333513418833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,768,0.048763556612862476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,2560,0.013524444566832649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,1024,0.016153777639071148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,3584,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,65536,7168,0.0814817746480306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,65536,45.53667534722222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,5120,0.12643910778893366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,4096,0.009248889154858058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,64,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,3072,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,2560,0.004876444323195351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,12288,0.026516444153255884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,16384,0.15585066212548151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,6144,0.027976890405019123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,3584,0.03235733177926805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,3072,0.015751111838552687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,768,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,2048,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,4096,0.9493306477864584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,51200,0.1412044498655531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,12288,0.04133866561783685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,64,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,768,0.027074666486846075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,512,0.009762666291660732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,5120,0.01309333327743742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,512,0.13451377550760904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,10240,0.03636622097757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,8192,0.101547552479638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,512,0.01181422256761127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,64,0.013630222115251752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,512,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,768,0.0046862223082118565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,1536,0.0495413343111674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,2048,0.013083555632167392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,512,0.01164533363448249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,768,0.021951999929216173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,32,0.12108800146314834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,1536,0.021893334057595994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,12288,0.17983467049068877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,8192,0.1600666708416409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,64,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,3584,0.01462044484085507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,2048,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,2560,0.010494222243626913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,1536,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,2048,0.21654844284057617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,51200,0.050182223320007324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,51200,0.15812622176276311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,10240,0.0882435573471917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,512,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,12288,0.16411111089918348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,4096,0.3320044411553277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,51200,0.2038471168941922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,5120,0.010176888770527309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,8192,0.02347999976740943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,16384,2.6454249487982855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,4096,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,65536,0.12387111451890732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,3072,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,3072,0.02162933349609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,1536,0.009864889085292816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,4096,0.031214220656288996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,6144,0.03972266780005561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,128,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,1024,0.026310222016440496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,10240,0.19591466585795084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,64,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,3584,0.052879111634360425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,32,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,256,0.01424533294306861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,65536,0.06495110856162177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,1024,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,7168,0.07323466406928168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,5120,0.011763555308183035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,768,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,256,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,1536,0.03443911009364658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,8192,0.14969066778818765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,768,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,4096,0.03234222200181749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,8192,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,12288,0.015094222293959724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,4096,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,12288,0.04463288850254483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,10240,0.7011217541164823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,6144,0.17728177706400552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,32,0.002456888970401552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,128,0.03454577922821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,768,0.13522756099700928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,128,0.021627555290857952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,65536,0.053393777873780995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,16384,0.04051999913321601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,2048,0.004192000048028098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,7168,0.039896888865364924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,8192,0.40371110704210067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,32,0.07866666714350383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,3584,0.03346844514211019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,1024,0.0047306666771570844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,512,0.01203555530971951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,10240,0.036128888527552284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,32,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,2048,0.29931733343336314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,1536,0.022861333356963262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,1024,0.02590311070283254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,10240,0.25966133011711967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,2048,0.04034311241573758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,1024,0.016650666793187458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,2048,0.008854221966531541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,3072,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,8192,0.10351644621955024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,128,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,1536,0.006434666613737742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,5120,0.08995555506812201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,3584,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,128,0.021073778470357258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,512,0.07315555546018812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,1536,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,3584,0.04648355642954508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,1024,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,12288,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,65536,0.4885875384012858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,16384,0.032058666149775185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,32,0.011121778024567498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,1536,0.0038764443662431505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,256,0.015456888410780164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,5120,0.04007555709944831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,1536,0.0296382208665212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,1024,0.010994666980372535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,1536,0.3481546772850885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,1536,0.02622577713595496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,10240,0.03730310996373495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,5120,0.020632000433074105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,10240,0.013976888524161445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,512,0.05760177638795641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,16384,0.31161512268914116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,12288,0.1312168836593628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,512,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,1536,0.05464355481995476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,256,0.020224889119466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,64,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,10240,0.026597332623269822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,3072,0.07640710804197523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,256,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,3584,0.008063111040327284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,1536,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,2560,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,51200,0.11519732740190293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,512,0.01090666651725769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,12288,0.0640631119410197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,2560,0.007072889142566257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,3072,0.020969778299331665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,65536,0.0950782232814365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,5120,0.04973777797487047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,1536,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,2048,1.1744231118096244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,65536,0.3962239954206679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,65536,0.9506319893731011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,8192,0.013694222602579327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,3584,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,128,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,10240,0.01994311147265964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,4096,0.015440889530711703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,2048,0.006124444305896759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,768,0.033914665381113686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,128,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,1024,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,16384,0.025461332665549383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,12288,0.05923644701639811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,6144,0.009719111025333405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,10240,0.07433778047561646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,51200,0.21062400605943468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,512,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,16384,0.016122667325867545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,2560,0.027907556957668726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,768,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,1536,0.019069333871205647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,2048,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,12288,0.016602666841612924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,2048,0.020096888144810993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,5120,0.045307556788126625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,16384,6.381964365641276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,1024,0.013748444616794586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,8192,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,64,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,3584,0.13120710849761963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,2048,0.02774311105410258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,1536,0.04407555527157254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,128,0.006526222245560751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,16384,0.014287999934620328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,10240,0.03202044301562839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,8192,0.011414222419261932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,12288,0.01368533323208491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,128,0.01851377718978458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,1536,0.009962666365835402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,3072,0.01551644504070282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,4096,0.02808533443344964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,7168,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,128,0.011383111278216044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,4096,0.015432000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,512,0.0040693332751592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,768,0.013190222283204397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,256,0.024784889486100938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,512,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,8192,0.1484017769495646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,128,0.11141244570414226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,256,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,7168,0.020950223008791607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,7168,0.041674666934543185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,256,0.0070159998204973005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,256,0.00221422231859631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,10240,0.019143111175960965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,512,0.006049777898523543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,65536,0.1360657744937473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,51200,0.10319644212722778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,65536,0.059375113911098905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,3584,0.03013333347108629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,7168,0.033647110064824425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,128,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,4096,0.03608266512552897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,5120,0.04211466511090597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,128,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,2560,0.01166311071978675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,16384,2.45231204562717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,768,0.01997777819633484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,7168,0.02197955548763275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,16384,0.013271999855836233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,8192,0.015056888262430826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,2048,0.49735556708441836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,128,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,512,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,128,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,16384,0.016844444804721408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,64,0.00360000009338061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,2048,0.1872640053431193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,7168,0.03511111272705926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,512,0.8199173079596626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,4096,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,4096,0.04808177881770664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,1536,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,512,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,512,0.05741066402859158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,65536,0.18398666381835938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,2048,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,1536,0.02123911182085673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,1536,0.008512000242869059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,5120,0.0236791107389662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,2560,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,10240,0.0896008875634935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,10240,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,768,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,1024,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,7168,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,4096,0.009111999637550777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,1536,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,8192,2.007287131415473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,7168,0.016400888562202454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,2048,0.009882666998439366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,128,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,128,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,65536,27.56643846299913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,6144,0.04746133420202467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,3072,0.029513776302337646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,7168,0.009912889036867354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,51200,0.6643173429701064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,10240,0.06450755728615655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,10240,1.343163596259223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,3584,0.01637066735161675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,128,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,128,0.16590578026241726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,2048,0.3293173313140869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,4096,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,8192,0.04428799947102865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,4096,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,3072,0.02789333462715149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,10240,0.12674933009677464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,2048,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,64,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,12288,65536,0.8071680068969727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,12288,0.03276000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,3072,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,6144,0.058528886901007764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,65536,1.1651919682820637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,1536,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,16384,0.07915021975835164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,3584,0.03703022334310744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,3072,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,3072,0.014934221903483072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,256,0.007777777810891469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,3072,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,65536,0.07300355699327257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,10240,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,2048,0.01977599991692437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,12288,0.02016533414522807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,1024,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,3072,0.01863466699918111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,3584,0.1340222226248847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,128,0.015480889214409722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,7168,9.033635457356771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,12288,0.18792888853285047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,512,0.006752000086837345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,4096,0.018167111608717177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,768,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,2560,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,65536,0.17889155281914604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,5120,0.04851999878883362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,12288,0.04814577764934964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,6144,0.014630221658282809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,1536,0.016234666109085083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,512,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,12288,0.7041093508402506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,16384,0.10934222406811184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,2560,0.10576177967919244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,512,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,4096,0.022077333596017625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,512,0.024821332759327356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,1024,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,8192,0.23074576589796278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,4096,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,32,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,7168,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,1024,0.03427822060055203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,1024,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,2560,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,512,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,51200,0.23901600307888457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,128,0.0027955555253558685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,1536,0.1003191139962938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,16384,0.09257511297861735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,2048,0.03240888979699876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,3072,0.006504888749784893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,256,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,10240,0.03894311189651489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,768,0.003055111194650332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,10240,0.1757946676678128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,128,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,32,0.0017893332988023758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,256,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,128,0.010399111443095738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,4096,0.07011555300818549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,3072,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,768,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,768,0.01888799998495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,768,0.017613333132531907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,64,0.021019554800457422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,3584,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,2560,0.16984711752997506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,12288,0.00452622233165635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,256,0.002248888835310936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,8192,0.02597155504756504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,10240,0.0165955556763543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,1536,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,3072,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,2560,0.029785778787400987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,768,0.01573244399494595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,3072,0.019832000136375427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,2560,0.028777778148651123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,5120,0.04349244303173489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,2048,0.020260444945759244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,768,0.005452444569932089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,65536,0.023896889554129705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,12288,0.18744266033172607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,51200,1.1480933295355904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,3584,0.03973866502443949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,1536,0.015513777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,2560,0.019938665959570143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,512,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,6144,0.018576888574494254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,768,0.02035555574629042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,6144,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,7168,0.08182844188478258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,12288,0.6357377900017632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,32,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,3072,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,3584,0.2667342291937934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,5120,2.1718016730414496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,8192,0.01494488947921329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,128,0.01752088963985443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,16384,0.10659466849433051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,6144,0.01278222186697854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,6144,1.6560977300008137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,4096,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,4096,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,512,0.01625866691271464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,256,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,32,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,512,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,8192,0.09533600012461345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,3072,0.1233724488152398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,8192,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,1536,0.019333332777023315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,256,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,512,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,5120,0.062076442771487765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,7168,0.010444444086816577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,7168,0.027122666438420612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,3584,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,32,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,3584,0.06115288866890801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,5120,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,51200,0.07089777787526448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,512,0.005071111023426056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,3584,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,16384,0.16854222615559897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,6144,0.054711110062069364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,1536,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,3072,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,128,0.0018782222436534034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,3584,0.021738666627142165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,768,0.12715021769205728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,512,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,4096,0.15057867103152806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,7168,0.03696444299485948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,8192,0.01941777765750885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,3072,0.06229333082834879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,128,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,1536,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,6144,0.02253688871860504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,51200,0.19819733831617567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,2048,0.004886222382386525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,512,0.002537777854336633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,6144,0.04934044347869026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,256,0.04908000098334419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,3072,0.008601778083377415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,768,0.017665778597195942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,3584,0.020720889170964558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,2560,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,51200,0.6928639941745334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,2560,0.013256889250543384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,1536,0.02530844509601593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,3072,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,10240,0.7455484602186414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,32,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,6144,0.04745511213938395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,1024,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,6144,0.15055199464162192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,5120,0.10819022523032294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,3584,0.016227554943826463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,512,0.014095111025704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,2560,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,2048,0.8662248717414008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,768,0.022222222553359136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,5120,0.05788977940877279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,16384,0.044179555442598134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,1536,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,7168,0.7904053264194064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,512,0.0070017774899800616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,64,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,51200,0.10099999772177802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,1024,0.06002933449215359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,1536,0.004016889052258598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,7168,0.02290311124589708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,5120,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,3584,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,128,0.17089688777923584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,16384,0.07893866962856717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,3584,0.08852710988786484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,1536,0.010832889212502373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,51200,0.07365866502126057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,4096,0.07576355669233534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,128,0.02126933303144243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,65536,0.13132799996270075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,3584,0.03342577815055847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,512,0.011552888486120435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,12288,0.16193155447642008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,10240,0.0406622224383884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,16384,0.021290666527218286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,256,0.003129777808984121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,3584,0.10667733351389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,65536,0.15085332923465306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,10240,0.014180445008807711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,8192,0.07926844225989448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,10240,0.0671608911620246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,7168,0.10101866722106934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,1536,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,4096,0.029343111647499934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,1536,0.11756800280676948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,256,0.01517599986659156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,256,0.00853777759604984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,256,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,16384,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,4096,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,4096,0.2207324504852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,2048,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,10240,0.06561777989069621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,16384,0.04886044396294487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,51200,0.05267111129230923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,2560,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,32,0.005324444423119227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,64,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,12288,0.015441776977645027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,64,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,10240,0.14395199881659612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,6144,0.00814488861295912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,128,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,4096,0.08195110824373034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,256,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,5120,0.008737777670224508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,3584,0.008178667061858708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,51200,0.6404310862223307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,65536,0.20815732744004992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,10240,1.7272639804416234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,3584,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,64,0.006719111154476802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,32,0.00700444479783376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,4096,0.0729902254210578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,2048,0.012153777811262341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,3072,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,768,0.020748444729381137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,51200,0.22314667701721191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,256,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,51200,0.3152364359961616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,16384,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,10240,0.07133066654205322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,10240,0.20462221569485137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,8192,0.024008888337347243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,2048,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,10240,0.022056000100241765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,1024,0.015043555034531487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,1536,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,2048,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,512,0.035198221604029335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,51200,1.2096008724636502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,7168,0.14108444584740534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,3584,0.01885333326127794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,768,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,2048,0.0395760006374783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,1024,0.015601777368121676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,3584,0.014889778362380134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,65536,0.06938222381803724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,51200,0.41558223300509983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,3072,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,128,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,3072,0.046888887882232666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,7168,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,6144,0.04019999835226271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,256,0.01973600023322635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,8192,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,51200,0.046281778150134616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,512,0.006178666734033161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,1024,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,3072,0.020402666595247056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,256,0.0023173333870040043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,512,0.02421866688463423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,10240,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,3072,0.013041777743233575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,4096,0.02996444371011522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,6144,0.032215111785464816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,6144,0.023297778434223596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,5120,0.016968000266287062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,2048,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,3072,0.06647200054592557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,5120,0.02770044406255086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,65536,0.10256710979673599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,12288,0.10619911220338608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,768,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,3584,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,7168,0.022479999396536086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,65536,0.06397510899437799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,65536,0.07643910911348131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,1536,0.011389333340856763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,10240,0.01239999963177575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,16384,0.04277688927120633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,768,0.0058319999112023254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,3072,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,3072,0.008908444808589088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,512,0.014268444644080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,1024,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,512,0.010975999964608086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,3584,0.022341334157519873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,8192,0.02959288822280036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,1024,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,3072,0.02439733346303304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,1024,0.01442577772670322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,64,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,12288,0.05573155482610067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,6144,0.02609066665172577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,256,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,2560,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,4096,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,2560,0.07747644186019897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,512,0.03937777876853943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,51200,0.408536884519789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,10240,0.06048711140950521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,6144,0.017601778109868366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,3584,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,16384,0.050428443484836154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,4096,0.040932443406846784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,1536,0.015391111373901367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,256,0.020799999435742695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,12288,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,2560,0.021959111094474792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,512,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,8192,0.014920888675583733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,3584,0.015853333804342482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,4096,0.03020533257060581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,3584,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,12288,0.09341777695549859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,51200,0.10527111424340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,512,0.008375110725561777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,2560,0.019183110859658983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,10240,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,1536,0.11574399471282959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,51200,0.11075466871261597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,32,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,512,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,1024,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,1536,0.011136000355084738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,7168,0.11157245106167263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,51200,0.045888887511359326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,8192,0.19134578439924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,2560,0.037532445457246565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,3584,0.038887110021379255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,16384,0.5405813323126899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,128,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,1536,0.012139555480745105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,16384,0.02387199964788225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,3072,0.014104889498816596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,64,0.010904889139864178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,6144,0.010109333528412713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,16384,0.12860177622901067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,2560,0.03189333279927572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,1536,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,5120,0.009975999593734741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,1536,0.05863644679387411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,1024,0.004536889079544279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,1536,0.024950222836600408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,6144,0.020107555720541213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,128,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,3072,0.03071555495262146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,1024,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,5120,0.041821334097120494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,12288,0.2158151070276896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,2048,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,128,0.006807111203670502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,12288,0.013782222237851886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,128,0.021486222743988037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,5120,0.02590755621592204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,10240,0.03656355539957682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,64,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,6144,0.012994666894276937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,65536,0.1746844450632731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,64,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,65536,0.17275644673241508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,6144,0.030433777305814955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,1024,0.05196977655092875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,256,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,5120,0.026866666144794885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,51200,0.29980089929368764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,1536,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,5120,0.044241779380374484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,2048,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,65536,0.427623987197876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,3072,0.026122666067547266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,7168,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,2048,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,128,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,3584,0.06067644225226509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,512,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,1536,0.0038648889296584656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,51200,0.07355822457207574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,32,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,3584,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,65536,0.1638666656282213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,10240,0.027646222048335608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,3584,0.01514755520555708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,64,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,65536,0.017986666825082567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,512,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,16384,0.01802399920092689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,4096,0.04378844300905863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,256,0.01572800013754103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,8192,0.011081777513027191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,16384,0.010427555276287926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,768,0.0070471109615431885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,1536,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,512,0.005497777627574072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,5120,0.07089689042833117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,1024,0.04016710983382331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,7168,0.012289777398109436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,256,0.015258666541841296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,1536,0.02715377840730879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,12288,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,3072,0.004935111022657818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,6144,0.019873778025309246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,6144,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,2560,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,1024,0.020103999310069613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,5120,0.03947110970815023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,12288,0.07546755340364245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,51200,0.03936533464325799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,256,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,10240,0.013832888669437833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,16384,0.12588355276319715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,1024,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,2048,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,4096,0.004250666747490565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,65536,3.750825670030382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,256,0.02068355513943566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,51200,0.20219822724660239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,1536,0.035057776504092746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,5120,0.7282613118489584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,4096,0.11645955509609646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,3584,0.130986664030287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,5120,0.025098666548728943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,1024,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,1024,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,5120,0.02236266599761115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,256,0.018918222851223417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,256,0.00849866701496972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,12288,0.033223999871148005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,32,0.023286221755875483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,2048,0.037965334124035306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,256,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,3072,0.053914666175842285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,1536,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,128,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,1024,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,3072,0.28748801019456655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,4096,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,3584,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,51200,0.5269404517279731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,1536,0.06954755385716756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,1024,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,1024,0.0029013332807355454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,10240,0.09194222423765396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,12288,0.04320799973275927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,4096,0.06014844444062975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,10240,0.012458666331238218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,3584,0.016562667157914903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,2048,0.020219556159443326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,128,0.010518222219414182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,512,0.005415110952324337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,12288,0.049462222390704684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,2560,0.007818666597207388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,8192,0.07083555724885729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,3072,0.09246933460235596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,2560,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,16384,0.08046578036414252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,512,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,16384,0.020618667205174763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,4096,0.009408888717492422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,512,0.026945778065257605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,2560,0.004570666700601578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,65536,0.07597155703438653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,3072,0.04155110981729295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,51200,0.1033146646287706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,2560,0.006107555495368109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,8192,0.010712888505723743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,512,0.0058222222659322954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,5120,0.014479110638300577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,128,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,65536,0.05247911148601108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,65536,0.07626933521694608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,10240,0.02281155520015293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,512,0.008813333180215623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,1536,0.07454488674799602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,10240,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,7168,0.03529422150717841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,128,0.01620444489849938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,10240,0.032574223147498235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,3584,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,65536,0.1931288904613919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,16384,0.049512889650132924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,3072,0.025719111164410908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,2560,0.03311999970012241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,512,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,51200,0.4199546708001031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,2560,0.19809777206844756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,64,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,2560,0.029912889003753662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,4096,0.00869066682126787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,512,0.03816711240344577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,65536,0.0711688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,2048,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,512,0.014412444498803882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,4096,0.16758310794830322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,4096,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,512,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,3584,0.6176088651021322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,768,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,12288,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,12288,0.01997244523631202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,1536,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,5120,0.054789334535598755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,8192,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,5120,0.046987556748920016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,2048,0.04161244299676683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,5120,0.10053066412607829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,4096,0.0852560003598531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,3072,0.048284444544050426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,16384,0.08332000176111858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,51200,11.239339192708334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,7168,0.012128888732857175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,3072,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,3072,0.07349155346552531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,1024,0.0063271112740039825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,5120,0.10604266987906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,2560,0.009521777431170145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,64,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,128,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,10240,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,3072,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,7168,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,65536,0.2459555466969808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,51200,0.14696533150143093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,32,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,65536,0.04821422364976671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,65536,0.20109332932366264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,768,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,256,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,8192,0.013927999469969006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,65536,0.04679377873738607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,3072,0.014797333214018079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,5120,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,10240,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,1024,0.031101332770453557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,2048,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,2560,0.02717155549261305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,128,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,3584,0.024715556038750544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,65536,0.02370044423474206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,32,0.023000889354281958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,128,0.2510471079084608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,128,0.01774222155412038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,5120,0.05069955521159702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,128,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,4096,0.01108000013563368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,5120,0.06924088795979817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,2048,0.02186222208870782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,256,0.003141333245568805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,1536,0.04846488767200046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,768,0.002562666725781229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,16384,0.021377777059872944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,2048,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,16384,0.03204444381925795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,2048,0.11048355367448594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,65536,0.02411199940575494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,512,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,1536,0.026696888936890498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,32,0.01737777723206414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,256,0.011227555572986603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,51200,0.03531733486387465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,12288,0.15579288535647923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,1024,0.006495999793211619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,3584,0.03854044609599643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,1536,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,5120,0.027284445034133062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,2560,0.005352889084153705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,2048,0.01536622146765391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,4096,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,7168,0.013606222139464485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,6144,0.01980266637272305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,6144,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,256,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,768,0.00757422215408749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,3072,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,4096,0.039488001002205744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,128,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,10240,0.020614221692085266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,8192,0.010638222098350525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,512,0.01130222198035982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,768,0.017474666237831116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,8192,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,12288,0.07166400220659044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,10240,0.05286666750907898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,6144,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,51200,0.05446666810247633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,4096,0.07691733042399089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,16384,0.017204445269372728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,8192,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,8192,0.06786044438680013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,51200,0.05437689026196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,8192,0.21558666229248047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,1536,0.006186666587988536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,2048,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,6144,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,65536,0.15538756052652994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,512,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,512,0.2188942167494032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,1536,0.04768088791105482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,12288,0.33918044302198624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,5120,0.15384088622199163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,512,0.01667555504375034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,7168,0.02404622236887614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,1536,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,3072,0.033230221933788724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,6144,0.013965333501497904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,7168,0.055085334512922496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,512,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,65536,0.0982817808787028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,4096,0.6169270939297146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,768,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,4096,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,51200,0.8983644909328885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,3072,0.017099555995729234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,6144,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,2048,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,768,0.09551911221610175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,3072,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,12288,0.4847448666890462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,768,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,3072,0.031298667192459106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,3072,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,65536,1.064710193210178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,3072,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,16384,0.24128799968295625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,128,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,3584,0.09206400314966838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,256,0.020269332660569083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,51200,0.10019289122687446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,3072,0.31431378258599174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,12288,0.03391022152370877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,4096,0.4421351220872667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,5120,0.10735288593504165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,128,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,3072,0.025787555509143408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,7168,0.01590044465329912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,16384,0.07506489091449313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,6144,0.028557333681318495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,10240,0.01620533400111728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,8192,0.02606222199069129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,64,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,65536,0.20242844687567818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,32,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,12288,0.017906667457686532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,65536,0.1445173290040758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,2048,0.01758666667673323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,6144,0.010786666638321347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,16384,0.03463022245301141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,16384,0.2744319968753391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,64,0.025607110725508794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,512,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,12288,0.37090222040812176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,768,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,128,0.014761777387724983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,5120,0.059750219186147056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,5120,0.018736888964970905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,7168,0.3714844385782878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,768,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,128,0.04890399840142992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,16384,0.03811111052831014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,4096,0.00811644477976693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,16384,3.8679945203993054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,51200,0.42869777149624294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,2048,0.05061955584420098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,6144,0.01166044423977534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,128,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,2560,0.0075795559419526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,7168,0.039064887497160174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,2048,0.02104088829623328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,4096,0.007526222202512953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,512,0.0052879999081293745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,3584,0.06634310881296794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,3072,0.020576000213623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,16384,0.11549688710106744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,2560,0.8004070917765299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,64,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,16384,0.03514400124549866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,6144,0.0820924441019694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,256,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,1536,0.01906133360332913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,768,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,64,0.05235911077923245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,3072,0.037979556454552546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,256,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,3584,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,3584,0.20135733816358778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,768,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,3072,0.033061332172817655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,3584,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,6144,0.019902222686343722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,128,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,51200,0.04012622104750739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,128,0.005168888717889786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,4096,0.01272266689274046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,256,0.02444800072246128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,1536,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,51200,0.01348444405529234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,1024,0.006558222075303395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,768,16384,0.11295110649532741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,256,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,3584,0.08085333638721041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,12288,0.2271075513627794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,128,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,1024,0.024292444189389546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,12288,0.2528417905171712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,2560,0.07990666892793444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,7168,0.02510044475396474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,2560,0.03986400034692552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,7168,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,12288,7.811404758029514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,1536,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,1536,0.010132444401582083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,6144,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,512,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,16384,0.0724231137169732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,4096,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,128,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,1024,0.0031431110368834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,128,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,512,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,3072,0.018052443861961365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,4096,0.009101333717505137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,3072,0.425817781024509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,2560,0.030975110001034204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,51200,0.0692053304778205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,5120,0.012474666866991254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,2560,0.033923556407292686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,65536,0.15193155076768664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,12288,0.1080897781583998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,51200,4.532568189832899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,768,0.03451822201410929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,8192,0.4649617936876085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,16384,0.1480275524987115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,65536,0.06413244538837008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,2560,0.2874622344970703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,32,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,8192,0.07546489106284247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,64,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,2048,0.007353777686754863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,7168,0.01202044470442666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,8192,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,10240,0.24788178337944877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,12288,0.03219288918707106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,1024,0.01311911145846049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,8192,0.022658665974934895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,5120,0.051241778665118747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,8192,0.020376889242066275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,3072,0.011286222272449069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,3584,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,5120,0.1579093270831638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,65536,0.06449866957134671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,3584,0.014225777652528552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,6144,0.01853511068556044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,2048,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,4096,0.014838222000333997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,65536,0.21640800105200875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,10240,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,8192,0.01903022163444095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,2560,0.024752888414594863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,2048,0.013432000246312885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,768,0.017735110388861764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,5120,0.12896000014411077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,12288,0.032552000549104475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,5120,0.02401688860522376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,3584,0.04169155491722954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,4096,0.01679377754529317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,4096,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,1024,0.013055110971132914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,3584,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,64,0.11221510834164089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,1024,0.027031110392676458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,512,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,3584,0.023393778337372675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,3584,0.01779911087618934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,16384,0.27121512095133465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,128,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,768,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,2560,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,2048,0.024728889266649883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,10240,0.012094222009181976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,4096,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,4096,0.09411466783947414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,2048,0.012294222083356647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,2048,0.009977777798970541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,3072,0.030243555704752605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,2560,0.15012444390190974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,64,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,2560,0.04050133294529385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,2560,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,51200,0.11887466907501221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,6144,0.0291884442170461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,12288,0.1284844477971395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,8192,0.06921688715616862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,6144,0.05954755677117241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,5120,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,2048,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,2048,0.029439111550649006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,256,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,128,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,6144,0.10509688986672296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,16384,0.05491911040412056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,5120,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,256,0.020370667179425556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,2048,0.006744888921578725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,6144,0.024847111768192712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,2560,0.05418222149213155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,32,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,16384,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,3072,0.03750400079621209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,512,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,4096,0.004936000125275718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,7168,0.8627217610677084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,3584,0.14745155970255533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,51200,2.503199895222982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,8192,0.09187466568417019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,128,0.03502755694919162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,12288,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,7168,0.3214968840281169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,4096,0.018410666121376883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,10240,0.02549422283967336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,4096,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,3584,0.028507557180192735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,3584,0.025791111919615004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,3584,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,12288,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,16384,0.03402488761478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,51200,0.06494577725728352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,3072,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,65536,0.10549333360460068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,4096,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,256,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,512,0.012798222402731577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,8192,0.10966933435863918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,6144,0.01665155589580536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,5120,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,1024,0.011506666739781698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,4096,0.017464000317785475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,32,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,65536,0.0852888888782925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,65536,0.06645955642064412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,2048,0.02399377855989668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,51200,0.05703555875354343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,7168,0.010669333239396414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,8192,0.022612444228596155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,768,0.051480889320373535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,64,0.023616888456874426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,2560,0.011405333048767514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,3072,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,10240,4.239331563313802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,6144,0.031319111585617065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,8192,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,2560,0.026865777042176988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,8192,0.08422222402360703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,10240,0.010114666488435533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,32,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,51200,0.1096995539135403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,128,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,3072,0.03077510992685954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,4096,0.09393511215845744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,2560,0.016566221912701923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,64,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,5120,0.02527644402450985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,128,0.0147724449634552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,8192,0.011416888899273343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,1536,0.017669333351982962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,7168,0.020244444410006206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,1536,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,1536,0.016598222984208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,51200,0.05308177736070421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,2560,0.0069982219073507525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,128,0.043862223625183105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,3072,0.007697777615653143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,512,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,256,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,32,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,1024,0.018256000346607633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,12288,0.14800888962215847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,7168,0.12186400095621745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,64,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,51200,0.06816266642676459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,6144,0.05717421902550591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,4096,0.02369511127471924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,8192,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,51200,0.15875555409325495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,1536,0.03279555506176419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,128,0.014293332894643148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,2560,0.015923556354310777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,16384,0.012507555385430654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,512,0.009200000100665623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,8192,0.05199644300672743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,3584,0.030234667989942763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,10240,0.012295111185974546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,768,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,65536,0.09160000085830688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,32,0.005975111077229182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,10240,0.01615111033121745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,51200,3.024711185031467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,768,0.0051893335249688895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,1536,0.0047022220161226065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,65536,0.0636791123284234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,5120,0.11547466119130452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,5120,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,5120,0.5286666552225748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,1536,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,16384,0.019260444574885897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,512,0.12927556037902832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,51200,0.37007912000020343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,3072,0.021799110703998145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,10240,0.025576000412305195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,51200,0.5378746456570095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,256,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,768,0.020980444219377305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,1024,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,128,0.00757244477669398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,4096,0.010113777385817634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,12288,0.03183733423550924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,6144,0.12239821751912434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,12288,0.03095910946528117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,3072,0.037512888511021934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,5120,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,3072,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,1024,0.016827555166350473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,1536,0.047332445780436196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,1024,0.00453955555955569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,2560,0.011001778145631155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,8192,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,768,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,5120,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,4096,0.07600977685716417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,3584,0.018260444204012554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,256,0.005388444496525659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,51200,0.3105866644117567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,768,0.0380639996793535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,10240,0.02756888833310869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,2560,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,1024,0.006994666324721442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,1024,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,32,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,16384,0.018910222583346896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,2560,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,64,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,1024,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,51200,0.16266932752397326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,128,0.009282666775915358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,3584,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,8192,0.01664977769056956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,16384,0.20364977253807914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,2560,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,10240,0.027786665492587622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,51200,0.1172213289472792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,768,0.040130668216281466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,6144,0.03671911027696397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,4096,0.04968888892067803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,5120,0.010151999692122141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,2048,0.018608889646000333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,512,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,51200,0.06499733527501424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,12288,0.02107200026512146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,128,0.03291555576854282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,1024,0.06098310814963447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,2560,0.006133333262470033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,2560,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,16384,0.08455289072460598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,4096,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,1024,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,3072,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,10240,2.11482302347819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,12288,0.04138133260938857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,512,0.34113510449727374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,256,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,32,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,3584,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,128,0.006903999795516332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,3584,0.047190222475263804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,8192,0.1395217842525906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,4096,0.06338400310940213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,10240,0.8062897788153754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,2048,0.00871022211180793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,768,0.014880888991885714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,12288,0.03242933419015672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,128,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,128,0.03948622279696994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,1024,0.12952978081173366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,512,0.06638755400975545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,51200,0.050979554653167725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,4096,0.016724444097942777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,3584,0.005091555416584015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,256,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,64,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,12288,0.016898666818936665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,5120,0.1313635508219401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,1536,0.018809777167108323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,6144,0.00978133330742518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,32,0.0018159999615616268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,768,0.023398222194777593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,128,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,2560,0.019811555743217468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,3584,0.05252977874543932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,6144,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,6144,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,256,0.014200000299347771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,64,0.010796444283591377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,2560,0.05931378073162503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,6144,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,1024,0.010058666268984476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,3584,0.03996710975964864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,256,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,16384,0.03428533342149522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,768,0.013377777404255338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,10240,0.024358222881952923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,65536,0.06644444333182441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,768,0.0136808885468377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,10240,0.16797510782877603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,5120,0.0038275557259718576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,1024,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,5120,0.02409422232045068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,16384,0.034303112162484065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,64,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,7168,0.012824888858530255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,16384,0.02325155503220028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,8192,0.11487910482618545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,8192,0.013208000196350945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,32,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,12288,1.940880881415473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,4096,0.9107884301079644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,5120,0.03836089041497972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,10240,0.12125421894921197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,64,0.016336888074874878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,1024,0.014535110857751636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,128,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,2048,0.015564444992277356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,3584,0.032758222685919866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,16384,0.10385421911875407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,1536,0.04414044486151802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,1024,0.05339644352595011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,512,0.026387555731667414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,8192,0.03302133414480422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,3072,0.018925334016482036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,256,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,2560,0.01609599921438429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,16384,0.1943653292126126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,6144,0.027821332216262817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,8192,0.028949333561791316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,128,0.021537777450349595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,16384,0.031606223848130964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,256,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,5120,0.38007378578186035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,128,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,64,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,2560,0.024853333830833435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,2560,0.009820444716347588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,256,0.03702400128046671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,65536,0.22156266371409097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,64,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,12288,0.01332800007528729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,1024,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,65536,0.13167556126912436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,1024,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,128,0.12327022022671169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,65536,2.1684320237901473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,12288,0.07007644573847453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,8192,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,51200,0.2014497783448961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,128,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,8192,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,10240,0.012432888978057437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,3072,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,10240,0.3202853202819824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,768,0.020657777786254883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,256,0.00254488881263468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,4096,0.03696711195839776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,3584,0.024550222688251074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,512,0.020258666740523446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,8192,0.11824177371131049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,1024,0.004159111115667555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,12288,0.013449778159459433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,3072,0.012862222062216865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,512,0.0031840000301599503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,2048,0.016819554898473952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,512,0.019881778293185763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,3072,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,10240,0.46618933147854275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,256,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,3072,0.03003466791576809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,1024,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,12288,0.01963644391960568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,8192,0.045817779170142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,6144,0.04239999916818407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,2560,0.03162399927775065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,3584,0.03685777717166477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,256,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,7168,0.24311288197835287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,64,0.01669066647688548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,2560,0.0540284448199802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,256,0.03161155515246921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,1536,0.059352888001335986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,10240,0.037547555234697126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,1024,0.01348355578051673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,7168,0.05125955409473843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,512,0.007045333584149678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,3072,0.025976889663272437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,8192,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,256,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,3584,0.04874222146140205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,4096,0.007870222131411234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,128,0.005440000030729506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,3584,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,2560,0.016162667009565566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,51200,1.337031152513292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,3584,0.01679999960793389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,512,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,7168,0.02607111136118571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,768,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,512,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,1024,0.0672524438963996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,1024,0.013455111119482251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,1536,0.004933333231343163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,128,0.021611554755104914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,6144,0.054465777344173856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,1536,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,51200,1.1733439763387044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,7168,0.20149511761135527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,12288,0.17966132693820527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,3584,0.14443466398451063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,3584,0.021871111459202234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,32,0.004192000048028098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,16384,0.1313306623035007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,2048,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,12288,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,16384,0.01626399987273746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,51200,0.06975555419921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,5120,0.011651555697123209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,768,0.017837334010336135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,10240,0.08755822314156426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,3072,0.008681777450773451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,65536,0.06504978073967828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,64,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,6144,0.02420355545149909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,256,0.003539555602603488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,2560,0.01331733332739936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,64,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,7168,0.01670222149954902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,256,0.03512800070974562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,1536,0.02525866693920559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,2560,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,32,0.057686222924126514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,256,0.014242667290899487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,7168,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,6144,0.15796621640523276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,5120,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,256,0.0031591111587153543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,16384,0.06143289142184787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,1024,0.012366222010718452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,4096,0.04814755585458544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,51200,0.0739955570962694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,7168,0.01369244439734353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,256,0.016805332567956712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,768,0.011782222323947482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,5120,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,6144,0.3852648999955919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,3584,0.04275288846757677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,10240,0.05004088746176826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,512,0.10221066739824082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,3584,0.020777778493033517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,128,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,512,0.010317333042621613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,6144,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,512,0.004621333132187526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,1024,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,768,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,2048,0.031367109881507024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,6144,0.1387466722064548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,128,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,12288,0.017653332816229928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,512,0.024325332707828943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,51200,0.16579643885294595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,65536,1.1236311594645183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,65536,0.20031378004286024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,16384,0.09026133351855808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,3584,0.010100444157918295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,4096,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,10240,0.03631022241380479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,5120,0.35686667760213214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,51200,3.2358656989203562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,2560,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,6144,0.020419556233617995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,256,0.024041778511471216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,4096,0.020118221640586853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,12288,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,6144,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,3072,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,6144,0.007874666816658443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,32,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,6144,0.024838222397698298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,64,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,5120,0.008358221915033128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,4096,0.008032889001899296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,64,0.09522755278481378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,12288,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,2560,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,3072,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,12288,0.0369484457704756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,16384,0.022281777527597215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,5120,0.01074488874938753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,2048,0.013264888690577613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,16384,0.14324177636040583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,12288,0.26618043581644696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,768,0.003185777821474605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,3584,0.008803555534945594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,6144,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,5120,0.18665511078304717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,7168,0.032673776149749756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,8192,0.020663999848895602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,768,0.10947555303573608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,6144,0.13714578416612414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,51200,0.15243555439843073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,512,0.011696000066068439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,65536,0.2779475582970513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,768,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,256,0.0022195554855797025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,4096,0.04378933376736111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,256,0.006730666591061487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,1024,0.03376177615589566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,2048,0.019343111250135634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,3584,0.016556445095274184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,3072,0.0517520010471344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,768,0.034167998366885714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,1536,0.012159999873903064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,8192,0.02365600069363912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,1536,0.006871999965773688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,65536,0.9735822677612305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,768,0.01927199959754944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,51200,0.06927911440531413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,12288,0.017533333765135873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,7168,0.01961688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,6144,0.04296355446179708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,32,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,12288,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,3072,0.12437600559658474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,4096,0.016187555260128446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,8192,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,256,0.11583467324574788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,3584,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,2048,0.00888000014755461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,5120,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,512,0.035558223724365234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,65536,0.13314221964942083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,3584,0.010092444717884064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,4096,0.02343822187847561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,128,0.11389688650767009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,8192,3.4381101396348743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,16384,0.24171998765733507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,6144,0.09888977474636501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,2560,0.02350666622320811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,3072,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,1024,0.03184355629814996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,3584,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,65536,0.27086933453877765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,10240,0.014740443891949125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,1024,0.015815999772813585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,65536,0.07734044392903645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,5120,0.21306578318277994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,4096,0.38010666105482316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,3072,0.022016889519161646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,64,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,8192,26.222230699327255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,64,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,5120,0.01919555498494042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,16384,0.010244444840484196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,5120,0.047113776206970215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,51200,0.5022720230950249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,51200,0.05606666538450453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,65536,0.05751644240485298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,6144,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,256,0.02772888872358534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,768,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,51200,0.5373448795742458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,5120,0.06458933485878839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,1536,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,3584,0.01035199976629681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,65536,3.8489163716634116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,51200,0.01917866700225406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,256,0.009579555855857002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,512,0.016991110311614144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,7168,0.046224001381132335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,5120,0.004868444469239977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,8192,0.01961688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,32,0.013609777722093793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,51200,0.04293244414859348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,1536,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,5120,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,2048,0.28414400418599445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,512,0.020595555504163105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,3072,0.009118222528033787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,16384,0.04907110995716519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,128,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,512,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,2048,0.03194844391610887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,512,0.008804444637563493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,12288,0.018287110659811232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,12288,0.018509333332379658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,32,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,3584,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,3072,0.0322746667597029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,768,0.0992586678928799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,256,0.015476443701320224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,7168,0.02702666653527154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,3584,0.07671111159854464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,768,0.021327111456129286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,6144,0.05309333403905233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,12288,0.029790222644805908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,7168,0.026483555634816486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,8192,0.18654933240678576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,128,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,65536,0.459736877017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,512,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,4096,0.13434399498833552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,128,0.028777778148651123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,512,0.006747555401590135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,16384,0.02049955560101403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,6144,0.018232888645595975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,2560,0.0621395574675666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,32,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,1536,0.025765332910749648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,768,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,7168,0.014125333891974555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,7168,0.20022311475541857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,12288,0.2751999961005317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,51200,0.9508800506591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,2048,0.013817778064144982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,12288,0.02748799986309475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,512,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,5120,0.00943288869327969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,2560,0.010139555566840703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,6144,0.061326218975914844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,3072,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,768,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,4096,0.008801778157552084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,7168,0.023194667365815904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,4096,0.2994115617540148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,4096,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,32,0.012602667013804117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,6144,0.03303822212749057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,12288,0.041356444358825684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,128,0.028993778758578833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,7168,0.01573866605758667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,8192,0.007710222568776872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,8192,0.014491554763582019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,32,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,4096,0.030135999123255413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,12288,0.051557332277297974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,4096,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,64,0.01977155605951945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,8192,0.0374035570356581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,5120,0.12473689185248481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,65536,0.17705867025587294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,256,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,3584,0.028022223048739966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,10240,0.018574222922325134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,7168,0.039509332842297025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,8192,0.05823555257585314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,2048,0.05565244621700711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,10240,0.01847555571132236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,12288,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,6144,0.143175999323527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,16384,0.3039413293202718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,65536,0.10477422343360053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,4096,0.0192248887485928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,2048,0.013510222236315409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,2560,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,8192,0.013754666679435305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,3584,0.020638222495714825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,12288,0.0731182230843438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,12288,0.016062221593326993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,3584,0.023027555810080633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,51200,0.16177333725823298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,3072,0.271542231241862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,6144,0.009240888886981541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,1024,0.006713777780532837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,10240,0.012458666331238218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,6144,0.011546666423479715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,2560,0.06482488579220243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,4096,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,16384,2.487184100680881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,10240,0.015003555350833468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,6144,0.07225333319769965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,3072,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,64,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,6144,0.009970666633711921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,65536,0.09233511156506008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,7168,0.05013333426581489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,4096,0.03015022145377265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,3584,0.01742044422361586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,65536,0.08094844553205702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,32,0.004341333276695675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,1024,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,10240,0.015825778245925903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,1024,0.013528000149461957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,12288,0.013766222529941134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,16384,1.6496515274047852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,7168,0.9087502161661783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,256,0.0041866666740841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,2048,0.00741244438621733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,2048,0.018624888526068795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,10240,1.1703661812676323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,2560,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,3584,0.04385688900947571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,5120,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,65536,5.803051418728298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,512,0.011766222616036734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,64,0.001812444378932317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,1024,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,65536,1.2541981803046334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,12288,0.034428444173600935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,512,0.0028373334142896864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,2048,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,12288,0.01752622259987725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,256,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,1536,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,51200,0.04826133449872335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,65536,0.3646186722649469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,7168,0.01567911108334859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,768,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,512,0.023678221636348303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,4096,0.04497866829236349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,1536,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,8192,0.04117600123087565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,7168,0.04996977912055122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,1024,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,128,0.022283555732833013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,2560,0.14568977885776097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,7168,0.04542488853136698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,1536,0.13658044073316786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,64,0.006214222146405115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,65536,0.06893955336676703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,10240,0.6413137647840712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,512,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,1024,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,65536,0.023023999399609033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,768,0.018944000204404194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,2560,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,2048,0.05707022216584948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,1024,0.054211556911468506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,5120,0.012831110921170978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,51200,0.14746667279137507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,3072,0.11750133832295735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,3072,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,6144,0.8713724348280164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,5120,0.010085333552625444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,128,0.015844444433848064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,256,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,64,0.007824888659848107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,32,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,5120,0.029858665333853826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,512,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,32,0.04296266701486376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,12288,0.020639111598332722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,51200,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,8192,0.02442933287885454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,7168,0.11710666285620795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,6144,0.6370542314317491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,5120,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,65536,0.0836648874812656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,65536,0.15584799978468153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,768,0.0159288893143336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,5120,0.024977778395016987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,768,0.002903999967707528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,65536,0.06339910957548353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,32,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,64,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,10240,0.031147556172476873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,128,0.015491555134455362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,128,0.026693332526418898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,1024,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,32,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,4096,0.040820442967944674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,7168,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,10240,0.01256711118751102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,8192,0.02420888841152191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,10240,0.046999110115898975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,16384,0.12775288687811956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,3584,0.07674400011698405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,10240,0.02411199940575494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,768,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,8192,0.15875911712646484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,10240,0.12421422534518772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,12288,0.026113778352737427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,10240,0.03153955605294969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,3584,0.0781351129213969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,10240,0.013811555173661975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,3072,0.012219555675983429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,1536,0.011983999775515662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,8192,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,5120,0.02811644474665324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,768,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,2560,0.01407999959256914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,512,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,8192,0.011890666352377998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,65536,0.09633421897888184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,32,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,256,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,5120,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,16384,0.022124444444974262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,6144,0.0399875541528066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,512,0.005320888840489917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,1024,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,768,0.03288622366057502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,1024,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,4096,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,4096,0.19684088230133057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,2048,0.028957333829667833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,2048,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,6144,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,3072,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,8192,0.014864888456132678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,64,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,7168,0.011785777906576792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,3072,0.01574577722284529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,512,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,7168,0.006994666324721442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,128,0.0058151111006736755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,2560,0.020929776959949072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,512,0.10395822260114883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,10240,0.201401776737637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,4096,0.05197066730923123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,128,0.0019519999623298645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,2560,0.007037333316273159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,2048,0.1926880015267266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,8192,0.017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,16384,0.024113777610990737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,8192,0.02993688980738322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,1024,0.013964444398880005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,3072,0.028751999139785767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,4096,0.05359466539488899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,3072,0.24265689320034453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,128,0.040067556831571795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,768,0.015855110353893705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,3584,0.08636444144778782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,512,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,4096,0.02022133270899455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,2560,0.014521777629852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,64,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,51200,0.06407644351323445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,256,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,256,0.0052879999081293745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,256,0.006672888994216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,1024,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,65536,0.04762577679422167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,7168,0.12369688351949056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,4096,0.09899644719229804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,3072,0.09605155388514201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,7168,0.03320711188846164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,5120,0.022268444299697876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,4096,0.08484088712268406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,3072,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,6144,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,64,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,768,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,128,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,16384,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,10240,0.026378666361172993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,16384,0.5057733323838975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,128,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,16384,0.09412622451782227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,3584,0.013731555806265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,16384,0.061252441671159535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,1536,0.010964444114102257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,3584,0.011511999699804517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,512,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,7168,0.09151199791166519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,51200,0.27172621091206867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,4096,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,12288,0.116939557923211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,2048,0.015139554937680563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,16384,0.10833333598242866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,1536,0.02848177817132738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,16384,0.015671110815472074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,2048,0.010017777482668558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,65536,0.6569546593560113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,2048,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,3584,0.29486134317186147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,1024,0.0041706665522522396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,5120,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,16384,0.3763742181989882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,16384,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,64,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,768,0.01519466605451372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,12288,0.09020977550082737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,3584,0.0788959993256463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,1536,0.03838222225507101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,2048,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,3072,0.012657778130637275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,2048,0.06040355232026842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,512,0.030400888787375555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,6144,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,768,0.013999111122555204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,2560,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,128,0.015864888827006023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,51200,1.0894746780395508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,10240,0.15996800528632268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,12288,0.02979555394914415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,7168,0.021662222014533147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,512,0.002546666603949335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,5120,0.06631822056240506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,1024,0.026557332939571802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,2048,0.016525333126386006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,51200,0.024021334118313257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,3072,0.009190222455395592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,6144,0.10218666659461127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,256,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,5120,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,256,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,12288,0.028240889310836792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,256,0.015103111664454142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,5120,0.011800000237094032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,512,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,64,0.13736355304718018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,7168,0.04010666741265191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,8192,0.041309333509869046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,16384,0.03441155619091458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,4096,0.06876889202329847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,1536,0.058099554644690625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,3072,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,10240,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,3072,0.07291822301016913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,512,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,65536,0.6493013170030382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,6144,0.015966221690177917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,768,0.03199111090766059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,10240,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,3584,0.024883555041419134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,1536,0.02021511064635383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,4096,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,5120,0.013600000076823764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,2048,0.04253244400024414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,3072,0.011158222125636207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,256,0.014767999450365702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,12288,0.2811448838975695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,1024,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,10240,0.010254222485754225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,10240,0.09765955474641587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,3072,0.020617778102556866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,1024,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,1536,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,12288,0.024061333802011278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,65536,0.04684799909591675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,7168,0.21187822024027506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,64,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,8192,4.125726911756727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,7168,0.026138666603300307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,10240,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,3584,0.026937777797381084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,3072,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,65536,0.054379555914137095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,10240,0.025199999411900837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,2048,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,4096,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,128,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,512,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,4096,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,256,0.011576000187132092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,512,0.011188444164064197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,6144,0.1046631137530009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,51200,17.623265584309895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,1536,0.00835111074977451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,10240,0.022610666023360357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,1536,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,51200,0.119668443997701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,64,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,8192,0.117122663391961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,8192,0.09191999832789104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,1536,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,7168,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,4096,0.09099733167224461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,768,0.027810666296217177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,1536,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,12288,10240,0.035826666487587824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,1536,0.032010667853885226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,768,0.025391111771265667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,6144,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,2048,0.009974222216341231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,7168,0.1464595529768202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,32,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,51200,13.460311889648438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,65536,0.12025066216786702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,3584,0.009525333013799455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,768,0.026234666506449383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,2048,0.024151111642519634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,3072,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,3072,0.016164445214801364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,4096,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,128,0.0019288888821999233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,256,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,65536,0.07322311401367188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,10240,0.02463200026088291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,3584,0.10483555661307441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,128,0.022621333599090576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,7168,0.032124443186653986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,64,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,16384,0.012387555506494312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,256,0.0021955555097924336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,5120,0.1137288941277398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,7168,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,12288,0.7375084559122721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,6144,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,768,0.0064426664676931165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,2560,0.040957334968778826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,128,0.006607111129495833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,128,0.0029120000286234748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,3072,0.03758044375313653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,128,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,5120,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,1024,0.01332622187005149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,512,0.0023102222217453853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,2048,0.039742221434911094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,5120,0.008239111138714684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,7168,0.4048471185896132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,16384,0.030590222941504583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,1536,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,16384,0.013723555538389417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,10240,0.03611733184920417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,64,0.021314667330847845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,10240,0.02160888910293579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,10240,0.01773422294192844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,10240,0.02683999968899621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,4096,0.030230220821168687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,2048,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,5120,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,768,0.006197333335876465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,65536,0.08570310804578994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,12288,0.10099644131130642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,128,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,128,0.00552977787123786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,1536,0.04599110947714912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,512,0.016132444143295288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,3584,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,64,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,5120,0.02442311081621382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,2560,0.0814888874689738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,768,0.017844445175594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,7168,0.058479110399881996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,12288,1.2265893088446724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,4096,0.040523555543687605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,1024,0.007968000239796108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,256,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,8192,0.03790311018625895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,1536,0.010844444235165914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,8192,0.012714666624863943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,3072,0.14030578401353624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,1024,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,16384,0.01568266749382019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,32,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,5120,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,4096,0.023443554838498432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,10240,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,6144,0.020115555988417733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,2048,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,768,0.2524089018503825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,768,0.016812443733215332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,2048,0.01276711126168569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,5120,0.019488000207477145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,8192,0.019691555036438834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,8192,0.27654843860202366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,12288,0.04086133175426059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,3072,0.017062221964200337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,256,0.0022044444663657085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,10240,0.03234577841228909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,1536,0.023553777072164748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,1024,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,7168,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,8192,0.01384622189733717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,4096,0.009287999735938178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,64,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,512,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,5120,0.023443554838498432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,5120,0.019546666079097323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,4096,0.08873599767684937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,8192,0.25381067064073354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,8192,0.03018933203485277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,2048,0.33941155009799534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,512,0.060356444782680936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,1536,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,65536,0.030580444468392268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,12288,0.03677511215209961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,10240,0.09529688623216416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,65536,1.862030241224501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,51200,0.5669217639499241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,16384,0.4532604217529297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,3584,0.03329244587156508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,65536,0.06344177987840441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,8192,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,5120,0.2537626690334744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,1536,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,7168,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,2048,0.013092444174819522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,3072,0.05458933446142408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,1536,0.14731821748945448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,5120,0.009746666583749983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,128,0.011438222395049201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,8192,0.6166853374905056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,128,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,3072,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,10240,0.02552177839808994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,768,0.004857777721352047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,51200,0.05281866590181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,12288,0.14857421980963811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,3072,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,3072,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,16384,0.04331555631425646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,768,0.02717155549261305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,2048,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,2560,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,64,0.015528000063366361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,768,0.045003556542926364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,12288,0.03324710991647508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,64,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,4096,0.019503111640612285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,128,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,32,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,32,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,2560,0.10169155730141534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,7168,0.011939555406570435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,1536,0.025677333275477093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,256,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,16384,0.1171306636598375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,65536,0.12594132953219944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,512,0.003383111208677292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,3584,0.028452446063359577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,51200,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,16384,0.01667644414636824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,64,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,3072,0.03293777836693658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,3584,0.01096088853147295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,12288,0.03988088832961188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,8192,0.20510666900210908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,512,0.07342222001817492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,5120,0.2432924376593696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,65536,0.13798756069607204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,2048,0.03771377934349908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,1536,0.02092622220516205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,7168,0.007033777733643849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,8192,0.01167377746767468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,3072,0.0546960002846188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,1024,0.009424000150627559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,12288,0.0255422227912479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,128,0.11138667000664605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,5120,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,4096,0.008179555336634317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,3584,0.08152000109354655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,768,0.01831022236082289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,4096,0.04135466615358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,7168,0.3983795642852783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,3072,0.015292444162898593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,10240,0.05044000016318428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,3072,0.0887955559624566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,768,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,3072,0.04854933420817057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,512,0.005407111098368962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,16384,4.141152699788411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,3584,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,32,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,768,0.019289778338538278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,7168,0.19437066713968912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,7168,0.019632889164818656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,7168,0.05430488785107931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,5120,0.009639110830095079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,16384,0.010820444259378644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,16384,0.04481955700450473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,3584,0.7348346710205078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,512,0.004029333177540037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,2560,0.01368622233470281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,51200,0.14147822062174478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,51200,0.10523288779788548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,1536,0.025044444534513686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,7168,0.021316443880399067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,5120,0.05165777934922112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,51200,8.498634338378906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,8192,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,5120,0.02717422280046675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,128,0.013677332964208392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,12288,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,512,0.0125928885406918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,3072,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,7168,0.017349332571029663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,32,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,8192,0.48408179812961155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,8192,0.015427556302812366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,64,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,5120,0.02639555599954393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,10240,0.013921777407328287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,3072,0.013696889082590738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,4096,0.12047466966840957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,7168,0.031056000126732722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,8192,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,64,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,16384,0.8630097707112631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,12288,0.04598311252064175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,6144,0.3852702246771918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,2560,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,7168,0.12367199526892768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,1024,0.049419555399152965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,64,0.013743110828929476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,512,0.012780444489585029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,4096,0.00815822184085846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,3584,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,8192,0.03613777624236213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,6144,0.43201422691345215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,512,0.012074666718641916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,3072,0.5124870936075846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,512,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,5120,0.9212782118055555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,12288,0.02004888819323646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,3072,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,1536,0.016574222180578444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,8192,0.0619608892334832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,5120,0.43079643779330784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,256,0.00655644428398874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,3072,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,6144,0.013040000365840064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,1024,0.16300000084771052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,7168,0.06133244434992472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,12288,0.020080889264742535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,65536,0.012328888807031842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,32,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,2048,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,65536,1.0970862706502278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,64,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,1024,0.015207999282413058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,12288,0.026886221435334947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,3584,0.07538666990068224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,6144,0.04007200068897671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,65536,0.07140355639987521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,6144,0.01772177716096242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,12288,0.024381332927280005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,7168,0.01975022256374359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,3584,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,65536,0.5562542279561361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,51200,9.134128994411892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,5120,0.032398223876953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,64,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,32,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,5120,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,2560,0.01036533299419615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,1024,0.02832977639304267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,3584,0.021567111214001972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,256,0.06970133384068807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,10240,0.04860799842410617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,8192,0.026149332523345947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,4096,0.08882577551735772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,3584,0.020041777027977836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,6144,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,12288,0.02382933265633053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,8192,0.1406044430202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,2560,0.006561777657932705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,3584,0.022611555125978258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,4096,0.03872355487611558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,65536,0.13723822434743246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,256,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,5120,0.01014311114947001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,3072,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,12288,0.05239911211861504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,12288,0.06495733393563165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,5120,0.023687111006842718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,512,0.011747555600272285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,128,0.011801777614487542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,8192,1.9086835649278429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,65536,0.344138675265842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,768,0.039134220944510564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,1024,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,12288,0.05419644382264879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,10240,0.16249956025017634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,128,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,2048,0.012656889028019376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,256,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,8192,0.04961511161592272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,768,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,128,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,4096,0.03455822335349189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,768,0.004927111168702443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,10240,0.02090222140153249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,8192,0.02886577778392368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,65536,0.5298595428466797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,768,0.03946844405598111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,256,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,64,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,3072,0.03091555502679613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,8192,0.02237244447072347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,3584,0.01642933322323693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,2048,0.00462666650613149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,7168,0.016300444801648457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,128,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,65536,0.589188469780816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,12288,0.6954595777723523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,5120,0.3261813322703044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,16384,1.9519422319200304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,1024,0.017980444762441847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,65536,0.6465244293212891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,768,0.01698666645420922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,2560,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,3072,0.11919644143846299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,51200,0.06985155741373698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,256,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,3584,0.023386667172114056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,10240,0.021591999464564856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,16384,0.02055111030737559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,4096,0.14176000489128962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,3072,0.02311644454797109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,5120,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,7168,1.7985110812717016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,1024,0.005330666485759947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,128,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,16384,0.4493128988477919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,5120,0.03995733459790548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,65536,4.974040985107422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,10240,2.0196427239312067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,512,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,4096,0.012255999777052136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,10240,0.04919111066394382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,256,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,64,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,512,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,5120,0.0163093325164583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,128,0.010247111320495605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,8192,0.03011911114056905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,512,0.016235555211702984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,1024,0.03960977660285102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,6144,0.017678222722477384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,10240,0.012705778082211813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,256,0.018711111611790128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,2048,0.006559999866618051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,5120,0.043225778473748096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,8192,10.605138142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,2560,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,768,0.02275377843115065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,65536,0.1086275577545166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,65536,8.510845608181423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,6144,0.030760000149408977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,1536,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,16384,0.0281368891398112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,7168,0.04317155480384827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,2048,0.010841777755154504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,1024,0.01722666621208191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,128,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,12288,1.9180933634440105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,1536,0.020209777686331008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,6144,0.01514933341079288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,65536,0.14254132906595865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,768,0.029492444462246362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,2048,0.019765333996878732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,64,0.05795555644565158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,8192,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,1536,0.03956177830696106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,65536,0.07616000043021308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,65536,0.07441333267423841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,3072,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,1024,0.0041644444896115195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,7168,1.3507697847154405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,51200,0.24873956044514975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,5120,0.035418666071361966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,768,0.030788444810443457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,128,0.021186666356192693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,2048,0.030424889591005113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,3584,0.06754755311542086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,7168,0.012059555285506778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,2560,0.011020444333553314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,12288,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,2048,0.005434666656785541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,12288,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,128,0.004848888764778773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,1536,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,512,0.017416889468828838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,51200,0.05050311154789395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,10240,0.12892533673180476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,3072,0.06817333565817939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,12288,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,256,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,6144,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,4096,1.6221226586235895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,1536,0.02680177821053399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,16384,0.14600266350640193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,5120,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,12288,0.32549066013760036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,512,0.01768977774514092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,1024,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,7168,0.03160266743765937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,1536,0.01464355488618215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,12288,1.0440568923950195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,8192,0.01888533267709944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,12288,0.03251111176278856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,8192,0.01072266697883606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,65536,0.05974133147133721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,3584,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,1536,0.02365333338578542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,256,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,5120,0.012364444633324942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,6144,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,1536,0.016356445021099515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,2048,0.018583110637134977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,7168,0.024029332730505202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,1536,0.0040835556056764396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,65536,0.20600267251332602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,3072,0.16403555870056152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,1024,0.07072444756825765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,512,0.011891555454995898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,10240,0.02721244427892897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,5120,0.23754665586683485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,51200,4.825375027126737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,6144,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,65536,0.0655955539809333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,2560,0.031545778115590416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,1536,0.010394666757848527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,128,0.19871021641625297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,768,0.008234666453467475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,2560,0.02609599961174859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,7168,0.046506666474872164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,1536,0.022115555074479844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,5120,0.04986933204862806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,1024,0.018759111563364666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,256,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,512,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,2560,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,3072,0.023499555057949487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,16384,0.025363556212849084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,1024,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,3584,0.016799110505315993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,10240,0.06865421930948894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,51200,0.10495022270414565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,16384,0.1455048852496677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,1536,0.021516443954573736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,256,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,128,0.023173333870040044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,3584,0.007885333564546373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,768,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,10240,0.05478577812512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,4096,0.3512684504191081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,768,0.031832890378104314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,1536,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,6144,4.662752787272136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,16384,0.1638008885913425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,65536,0.0748986668056912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,2560,0.011348444554540845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,3072,0.00453244439429707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,128,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,51200,0.05200533403290642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,512,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,7168,0.009979555176364051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,3584,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,8192,0.315229336420695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,1024,0.02765688962406582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,512,0.002604444407754474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,4096,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,1024,0.11178221967485215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,4096,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,3072,0.00702400008837382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,64,0.0025946667624844448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2048,65536,0.07691822449366252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,10240,0.012430222498046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,51200,0.06815466615888807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,512,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,2560,0.023471999499532912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,5120,0.031417777140935264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,768,0.3201306660970052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,6144,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,1024,0.014890667464998035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,51200,0.05264711048867968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,51200,1.1234124501546223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,2560,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,6144,0.2021182245678372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,256,0.018889778190188937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,4096,0.5595937834845649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,2048,0.005487111293607288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,6144,0.017829333742459614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,65536,0.13792978392706975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,3072,0.059973332617017955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,51200,0.45483912361992734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,1024,0.004601777841647466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,6144,0.04555999901559618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,3584,0.11630045043097602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,5120,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,2048,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,1536,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,1536,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,2048,0.008120888637171851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,5120,0.01755911111831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,256,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,32,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,7168,0.1997671127319336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,16384,1.1025866402520073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,6144,0.03137422270245022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,2048,0.019906666543748643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,12288,7.34037102593316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,16384,0.11161243915557861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,12288,0.22159111499786377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,8192,0.01312800000111262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,3072,0.02109244465827942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,3584,0.026716444227430556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,2048,0.015784000356992085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,64,0.008054222497675154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,3584,0.009375111096435124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,128,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,1024,0.0035048888789282907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,12288,0.02083288960986667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,5120,0.01236177815331353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,6144,0.026890666948424444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,256,0.013075555364290873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,65536,0.2857271035512288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,12288,0.07086933321423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,2560,0.030757334497239854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,512,0.020219556159443326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,6144,0.03494133220778571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,3072,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,3584,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,6144,0.15112622578938803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,2560,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,128,0.016153777639071148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,16384,0.02171733313136631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,51200,0.9691884782579211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,2048,0.01947377787695991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,128,0.002151111140847206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,3072,0.4066346751319037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,1024,0.017231111725171406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,12288,0.04497155547142029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,768,0.017815111411942374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,3072,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,16384,0.05453866720199585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,10240,0.2189262178209093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,3584,0.024328889118300542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,512,0.0038240001433425476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,2048,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,7168,0.014051554931534661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,10240,0.008010666403505538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,6144,0.12098222308688694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,1024,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,768,0.00599911105301645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,65536,0.2948088910844591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,10240,0.01292266696691513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,256,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,5120,0.026328888204362657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,16384,0.3351573414272732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,4096,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,2048,0.019567110472255282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,32,0.0017911110901170308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,32,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,16384,0.03155466583040025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,2048,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,65536,0.07212711042828031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,3072,0.06155111392339071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,3072,0.0084906667470932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,12288,0.535369767083062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,5120,0.11446489228142633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,51200,0.035967999034457736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,8192,0.013080889152155982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,64,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,4096,0.30118576685587567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,51200,0.27039554384019643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,3584,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,4096,0.009688888986905416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,6144,0.04752177662319607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,512,0.015620445211728414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,32,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,12288,0.013920888304710388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,12288,0.02952444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,4096,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,4096,0.05528177817662557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,10240,0.028355555401908025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,2048,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,512,0.030088000827365454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,128,0.021894221504529316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,4096,0.10926844676335652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,51200,0.16244622071584067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,6144,0.0690808892250061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,1536,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,65536,0.08206488688786824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,32,0.0020782223178280727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,2560,0.038733333349227905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,2048,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,2560,0.05701333284378052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,51200,0.06231200032764011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,4096,0.06184799803627861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,3584,0.006537777682145436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,10240,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,65536,0.08624177508884007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,1536,0.00980355590581894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,1024,0.004569777597983678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,256,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,10240,0.1974551147884793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,1536,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,128,0.03028800090154012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,12288,0.04491733180152046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,2560,0.007358222372002072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,7168,0.15335999594794378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,51200,0.08347644408543904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,7168,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,8192,0.03969511058595445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,512,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,8192,0.025751110580232408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,65536,0.13742844263712564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,12288,0.07597777578565809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,64,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,5120,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,12288,1.8717884487575953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,1536,0.03380533390574985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,256,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,65536,0.7122364574008517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,512,0.008018666671382057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,6144,0.013451555536852943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,2560,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,3072,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,512,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,5120,0.008417777717113495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,1536,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,256,0.14941867192586264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,1024,0.02276444435119629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,768,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,5120,0.056984000735812716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,8192,0.32452000512017143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,3584,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,2048,0.02532888948917389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,2048,0.012237333589129977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,4096,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,128,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,256,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,2048,0.061556445227728955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,1536,0.26973599857754177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,8192,0.05631999837027656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,768,0.059192001819610596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,256,0.006127111199829314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,768,0.0028844444702068963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,51200,0.06257244613435534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,256,0.02130844526820713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,1024,0.024318221542570326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,2048,0.006691555596060223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,2560,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,64,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,2048,0.008747555315494537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,2560,0.007345777418878343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,1536,0.005371555685997009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,32,0.013472889032628802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,3584,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,2048,0.012847111456924014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,16384,0.02162933349609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,64,0.013228444589508904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,16384,0.1898791127734714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,8192,0.03978844483693441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,128,0.007734222544564142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,768,0.019296000401178997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,7168,0.012792888614866467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,2048,0.023344889283180237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,5120,0.00904622260067198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,128,0.046133332782321505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,16384,0.018135999639829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,65536,0.13616266515519884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,32,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,65536,0.09156622489293416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,65536,0.1707244449191623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,3072,0.03138577938079834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,6144,0.15868977705637613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,768,0.006063111126422882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,1536,0.10102578004201253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,12288,0.015055111712879606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,1024,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,2560,0.0196079992585712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,7168,0.01499555508295695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,768,0.19712532891167533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,6144,0.05721066395441691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,8192,6.651770697699653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,5120,0.14656355645921496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,10240,0.0314702226055993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,1536,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,5120,0.010575110713640848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,2560,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,1024,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,65536,1.0897466871473525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,65536,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,6144,0.04193066557248434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,128,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,12288,0.034154666794670954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,2048,0.12911378012763128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,2560,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,65536,0.21289155218336317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,10240,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,65536,0.20426132943895128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,768,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,10240,0.04179555508825514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,256,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,2560,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,768,0.015843555331230164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,12288,0.031531555785073176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,6144,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,3072,0.0074826669361856245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,3072,0.17969689104292128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,4096,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,8192,0.02695288923051622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,6144,0.06955733564164904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,32,0.0031840000301599503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,5120,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,12288,0.2390728791554769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,8192,0.03052355514632331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,1024,0.019176888797018263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,5120,0.04478844337993198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,6144,0.02639644510216183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,5120,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,64,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,2048,0.027505778604083594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,32,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,1536,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,65536,0.154767115910848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,12288,0.038204444779290095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,7168,0.017533333765135873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,3072,0.03493688835038079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,128,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,256,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,3072,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,2048,0.029875556627909344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,3072,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,1024,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,10240,0.06644088692135282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,5120,0.009637333452701569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,2560,0.027929776244693335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,128,0.02273599969016181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,768,0.051029334465662636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,2560,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,2048,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,6144,0.2560417652130127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,256,0.015322667029168872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,10240,0.02679377794265747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,256,0.051431109507878624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,2048,0.03251911203066508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,7168,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,2048,0.012808889150619507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,16384,0.08512266476949056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,1536,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,6144,0.018978666928079393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,3072,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,2560,0.027090667022599116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,128,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,65536,7.259398566351996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,256,0.026700443691677515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,4096,0.02054222259256575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,128,0.04053511222203573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,1024,0.025036444266637165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,16384,0.022115555074479844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,4096,0.01923733287387424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,512,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,1024,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,768,0.01923022170861562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,768,0.008210666477680206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,128,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,1536,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,128,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,7168,0.027599111199378967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,6144,0.049067556858062744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,2560,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,1024,0.24469778272840711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,7168,0.2890000078413221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,512,0.01719377769364251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,4096,0.026434666580624048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,128,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,1024,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,2048,0.044159111049440175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,2048,0.018868444694413077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,256,0.008348444269763099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,16384,0.06309600008858575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,7168,0.23672978083292642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,16384,0.7622257868448893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,2048,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,256,0.009372444616423713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,3584,0.01051111105415556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,5120,0.047747555706236094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,2560,0.015963556038008798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,768,0.026551110876931086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,4096,0.05403911074002584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,64,0.003843555433882607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,12288,0.06685866912206014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,4096,0.017880888448821176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,3072,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,3584,0.010318222145239512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,3584,0.03628266519970364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,3072,0.05284800132115682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,10240,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,7168,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,16384,0.07817333274417453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,51200,0.2581457826826307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,51200,0.11048977904849583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,5120,0.027110222313139174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,128,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,65536,0.04685422115855747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,2560,0.004516444272465176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,10240,0.10849777857462566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,51200,0.9881049262152778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,12288,0.02405599918630388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,16384,0.04833066794607374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,5120,0.009407999614874521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,12288,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,4096,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,65536,0.07681155867046781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,3584,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,51200,0.052213334374957614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,512,0.0389342208703359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,16384,0.03151288959715102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,65536,0.09183377689785427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,6144,0.14311022228664821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,256,0.04119999872313605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,10240,0.15846044487423366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,51200,0.4984435505337185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,256,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,128,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,768,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,64,0.01072533345884747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,4096,0.19400533040364584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,10240,0.019881778293185763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,8192,0.20145422882503936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,64,0.0017991111510329777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,51200,0.04699110984802246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,1024,0.023010666171709698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,1536,0.031685332457224526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,1536,0.004449777719047335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,8192,0.050107555256949536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,3584,0.01812088820669386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,10240,0.024060444699393377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,1024,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,128,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,768,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,65536,0.12410310904184978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,10240,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,10240,0.022281777527597215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,5120,0.12036000357733832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,1024,0.022649778260125056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,4096,0.027454222242037456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,1536,0.011384000380833944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,256,0.020950223008791607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,32,0.006520000182920032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,3072,0.048918220731947154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,10240,0.015785778562227886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,3072,0.011195555329322815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,1024,0.03304800060060289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,65536,0.06233244472079807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,10240,0.07112799750434028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,3584,0.011386666860845355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,2048,0.037110222710503474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,5120,0.010660444696744284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,256,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,512,0.1342382166120741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,12288,0.2431253327263726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,12288,0.024559110403060913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,8192,0.020979555116759405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,4096,0.05069244570202298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,64,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,65536,17.319408840603298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,10240,0.3460720115237766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,3584,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,6144,0.022712000542216833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,6144,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,8192,0.060012446509467236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,16384,0.18045511510637072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,16384,0.016082665986484952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,12288,0.03729688790109422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,128,0.03864444295565287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,1024,0.02539288832081689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,6144,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,16384,0.06792000267240736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,6144,0.0514337784714169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,3584,0.13718310991923013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,3072,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,10240,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,128,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,128,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,8192,0.013637333280510373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,51200,0.8058551152547201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,3072,0.055778668986426465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,5120,0.08480444219377305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,256,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,1024,0.018817777434984844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,128,0.01513777838812934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,16384,0.017695999807781644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,2560,0.027248889207839966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,8192,0.050048000282711454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,1536,0.026336888472239178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,6144,0.1502222220102946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,1536,0.32751467492845326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,10240,0.01423822177780999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,2048,0.007734222544564142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,6144,0.015815999772813585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,3584,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,12288,0.03675377700063918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,8192,0.37896354993184406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,4096,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,512,0.0030337776988744736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,51200,20.827368842230904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,12288,0.08726400136947632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,128,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,7168,0.03654844562212626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,1024,0.004285333471165763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,51200,0.4047315650516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,10240,0.1518604490492079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,8192,0.036307556761635676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,3584,0.08578666713502671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,65536,0.558351993560791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,768,0.023330666952663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,12288,0.018184888694021437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,16384,0.024319999747806128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,6144,0.014019555515713163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,2560,0.017138666576809354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,3072,0.010226666927337646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,65536,3.7501686943901906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,1536,0.024680889315075342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,16384,0.016743999388482835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,10240,0.1131475501590305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,768,0.015244444211324057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,10240,0.00488711107108328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,3072,0.019208888212839764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,3584,0.03411022159788344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,1024,0.03740444448259141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,12288,0.030026667647891577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,3072,0.08884266350004409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,3584,0.04551288816663954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,2560,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,256,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,6144,0.033077332708570696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,16384,0.04544533292452494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,4096,0.11779733498891194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,1536,0.01740622189309862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,2048,0.0381644434399075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,16384,0.04542400108443367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,5120,0.027792887555228338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,256,0.008231110870838165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,512,0.02162666618824005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,64,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,512,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,51200,0.654179573059082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,12288,0.14208977752261692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,6144,0.07860089010662503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,51200,8.748463100857204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,7168,0.08663733137978448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,256,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,2560,0.03793066740036011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,5120,0.04794666502210829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,12288,0.14617066913180882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,2560,0.023549333214759827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,6144,0.02111111084620158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,2048,0.012640888492266336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,768,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,3584,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,7168,0.020242666204770405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,1536,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,2048,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,128,0.02784622377819485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,12288,0.030902223454581365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,6144,0.04498577780193753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,51200,0.050380445188946195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,8192,0.05339822173118591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,16384,2.9711680942111545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,16384,0.02385955552260081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,51200,0.056699554125467934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,3584,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,65536,0.07631378041373359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,5120,0.12363110648261176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,2560,0.060284442371792264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,128,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,3584,0.06773244672351413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,2048,0.3128400113847521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,256,0.020994666549894545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,6144,0.1301813390519884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,12288,0.041598220666249595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,51200,0.05454044540723165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,8192,0.1753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,10240,0.047092444366878934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,7168,0.021294222937689886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,2560,0.012410666379663678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,256,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,64,0.01757600075668759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,6144,0.10558399889204238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,10240,0.13389510578579372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,12288,0.13012177414364284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,2048,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,64,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,12288,0.019161777363883126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,6144,0.0882915523317125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,256,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,3584,0.888687981499566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,1024,0.11401866541968451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,3584,0.2648293442196316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,5120,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,128,0.020279111133681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,5120,0.0743324425485399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,5120,0.10650222169028388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,4096,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,10240,0.0514373348818885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,4096,0.028543111350801256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,2048,0.026535110341178045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,65536,0.11509866184658474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,10240,0.34847201241387266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,1024,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,3072,0.013439999686347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,2560,0.010255110760529837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,2048,0.037271999650531344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,512,0.013233777549531726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,51200,0.10846400260925293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,3072,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,128,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,7168,0.02123911182085673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,12288,0.041484445333480835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,2560,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,16384,0.11271554893917507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,8192,0.03160444564289517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,4096,0.032348444064458214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,2048,0.007917332980367873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,8192,0.34572267532348633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,2048,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,8192,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,128,0.012104888757069906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,64,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,51200,0.056794669893052846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,1536,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,512,0.0058204444746176405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,768,0.015301333533393012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,7168,0.030414223670959473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,3584,0.016570667425791424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,512,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,7168,0.01035911093155543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,2560,0.03864088985655043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,51200,0.5089439815945095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,51200,0.04042399923006693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,16384,2.5077910953097873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,3584,0.009352888498041365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,512,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,12288,0.029853334029515583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,51200,0.4853875372144911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,64,0.01094133324093289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,7168,0.011392888923486074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,6144,0.027560888065232172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,768,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,65536,0.7100728882683648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,1536,0.025013332565625507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,5120,0.008650666309727563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,16384,0.022866666316986084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,32,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,10240,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,3072,0.031194667021433514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,7168,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,16384,0.1604000065061781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,4096,0.04309511184692383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,65536,0.07021244366963704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,768,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,2560,0.0074311114019817775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,3072,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,128,0.007696888513035244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,5120,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,1024,0.25461777051289874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,1024,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,256,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,128,0.01549600064754486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,51200,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,10240,0.026740445031060114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,4096,0.08166933059692383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,768,0.014915555715560913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,1024,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,3584,0.10254133409923977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,12288,0.05031555559900072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,128,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,4096,0.017087111870447796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,2560,0.012313777373896705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,16384,0.03240266773435805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,1536,0.014934221903483072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,64,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,128,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,2048,0.010599999792046018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,7168,0.14978222052256265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,2048,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,10240,0.05758400095833672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,5120,0.010328888893127441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,16384,0.0579048858748542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,32,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,3072,0.007370666497283512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,5120,0.012698666916953193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,512,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,64,0.018552889426549275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,10240,0.14180800649854872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,51200,0.03987733191914029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,1536,0.01254488858911726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,512,0.03494933247566223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,1024,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,16384,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,51200,0.45920801162719727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,5120,0.011077333655622272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,3072,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,12288,0.033471110794279314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,32,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,4096,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,64,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,256,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,512,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,10240,0.017064000169436138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,1536,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,768,0.043655110730065234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,1536,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,128,0.010626667075686984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,3072,0.00923644420173433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,5120,0.06569155719545153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,256,0.009456000394291347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,5120,0.04797600044144524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,5120,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,65536,0.9896826214260526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,12288,0.1976942221323649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,16384,0.03738133443726434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,6144,0.01555022266176012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,5120,0.09587644206153022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,10240,0.012494222157531314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,6144,0.0840337806277805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,16384,0.462465763092041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,3584,0.1689280006620619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,4096,10240,0.11777244673834907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,65536,0.6074639956156412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,12288,0.54694578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,1536,0.1550097730424669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,768,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,4096,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,768,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,3584,0.06441155407163832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,128,0.004934222333961063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,2048,0.013567999833159976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,128,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,32,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,12288,0.021735111872355144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,1024,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,12288,0.015788444214397006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,64,0.0016737777946723832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,5120,0.06247022416856554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,2560,0.030244443151685927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,512,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,51200,0.19477778010898164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,512,0.056353780958387584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,51200,0.04032444291644626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,5120,0.04737066560321384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,256,0.032107555203967623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,6144,0.03417955504523383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,65536,0.2653946611616346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,2048,0.22447999318440756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,1536,0.02921066681543986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,65536,0.06452800167931451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,6144,0.024057777391539678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,2048,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,8192,0.18989866309695771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,16384,0.0775137808587816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,3072,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,768,0.02051733268631829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,8192,0.039464887645509504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,128,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,32,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,128,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,3584,0.03365066647529602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,65536,0.44454219606187606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,6144,0.12813244925604925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,1536,0.010383110907342698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,256,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,6144,0.013202666408485837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,512,0.018274666534529794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,2048,0.09023555782106187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,3072,0.03436533278889126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,7168,0.03875911235809326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,5120,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,65536,0.09409777985678779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,16384,0.021464889248212177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,3072,0.006498666687144174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,128,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,4096,0.06692800256941053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,16384,0.06497866577572294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,2048,0.009762666291660732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,768,0.008736888567606607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,1536,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,65536,0.24010221163431802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,3072,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,4096,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,3584,0.03748977846569485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,32,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,4096,0.031141334109836157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,51200,5.639998118082683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,16384,0.05507111218240526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,3584,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,768,0.1565964486863878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,1024,0.003565333369705412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,65536,0.2776533232794868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,65536,0.15740711159176296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,32,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,6144,0.00887022250228458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,51200,0.03381777803103129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,2560,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,3072,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,6144,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,8192,0.9304302003648547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,65536,1.481287956237793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,32,0.015034667319721647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,256,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,1024,0.02456355591615041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,512,0.039432889885372586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,5120,0.031116445859273274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,10240,0.061299555831485324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,12288,0.04663111103905571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,1536,0.2220231162177192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,12288,0.33085423045688206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,16384,0.99600707160102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,16384,1.2770311567518446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,6144,2.011853324042426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,2560,0.010290666586822933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,10240,0.04218577676349216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,10240,0.1810257832209269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,2560,0.010386666489972008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,16384,0.30430576536390513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,5120,0.01979555520746443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,6144,0.7989004453023275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,12288,0.013797333670987023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,51200,0.08868977758619521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,1024,0.0669048892127143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,8192,0.014658666319317289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,6144,0.15278755293952093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,768,0.005676444619894028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,3072,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,1024,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,10240,0.15349066257476807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,65536,0.7102142439948188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,5120,0.025468443830808003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,512,0.022898667388492163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,7168,0.005483555710977978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,8192,0.014755555325084262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,3584,0.11508356200324164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,128,0.008660444782839881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,16384,0.06437244680192736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,4096,0.03862933317820231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,1024,0.08261244164572822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,1024,0.010670222342014313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,12288,0.3672906557718913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,7168,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,1536,0.17281422350141737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,12288,0.10824444558885361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,768,0.020576000213623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,8192,0.013174222575293647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,2560,0.007688889073001013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,16384,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,12288,0.03404266635576884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,1536,0.04080977704789903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,8192,0.07865422301822238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,16384,0.05272266599867079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,16384,0.020390222469965618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,65536,0.23024977578057182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,51200,0.35866310861375594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,1024,0.003863111138343811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,6144,0.011389333340856763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,3584,0.00776533368561003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,512,0.03249955508444045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,512,0.017056889004177518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,8192,0.01182666669289271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,2048,0.02647200061215295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,128,0.018830221560266282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,256,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,6144,0.035755554835001625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,1024,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,512,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,512,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,5120,0.023602666126357183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,3584,0.035608000225490995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,8192,0.45199378331502277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,1024,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,16384,0.04979733294910855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,12288,0.03453244434462653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,256,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,3072,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,8192,0.23511377970377603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,1024,0.0221413340833452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,2560,0.028871999846564397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,6144,0.011437333292431302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,3072,0.011601777540312873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,128,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,3072,0.07770222425460815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,6144,0.026446221603287592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,3584,0.01535999940501319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,12288,0.019741333193249173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,10240,0.0853946672545539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,51200,0.354003561867608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,51200,0.691557354397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,256,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,4096,0.010601777997281818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,2048,0.090774224864112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,65536,1.6047200096978083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,10240,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,5120,0.04862755537033081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,1024,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,1536,0.003934222377008862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,2560,0.07150044706132677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,65536,0.6665146615770128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,512,0.02161688937081231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,5120,7168,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,3584,0.009770666559537252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,12288,0.017472889688279893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,8192,0.01051999959680769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,2560,0.027289777994155884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,16384,0.015828443898095023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,3072,0.0752684473991394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,51200,0.09122133255004883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,3584,0.007837333612971837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,768,0.00351200004418691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,512,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,65536,0.050810668203565806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,51200,0.061434666315714516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,1024,0.0032391111469931076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,5120,2.5330994923909507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,3584,0.01440088947614034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,64,0.010767111347781287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,1024,0.006608888920810487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,12288,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,5120,0.0991191135512458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,256,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,256,0.003912888881233004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,1024,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,1536,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,2560,0.024171556035677593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,6144,0.009671111073758867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,51200,0.5923173162672255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,10240,0.02331022255950504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,3072,0.04056533177693685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,6144,0.023994666006830003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,768,0.029566221767001685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,256,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,2048,0.0303182237678104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,1536,0.0038853333228164246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,8192,0.017920000685585868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,768,0.0437395539548662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,12288,0.018920888503392536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,768,0.015357333752844067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,256,0.0035413333939181436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,64,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,3584,0.010877333581447601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,512,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,512,0.030541333887312148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,512,0.006858666737874349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,1024,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,2560,0.013850666582584381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,4096,0.07922844754325019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,64,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,256,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,1024,0.005979555762476391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,2560,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,1024,0.031344887283113264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,128,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,51200,0.0458568897512224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,3584,0.2269937727186415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,51200,7.1945300632052955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,1536,0.28458576732211643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,7168,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,16384,0.14671377340952554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,4096,0.04863466819127401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,512,0.004515555583768421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,1536,0.0035573335157500375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,256,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,256,0.04412000046836006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,128,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,256,0.01238844460911221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,51200,0.10352800289789836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,768,0.04859911070929634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,8192,0.012071111136012606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,128,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,16384,0.020997333857748244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,2560,0.013578666581047906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,8192,0.033618665403789945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,1024,0.011704000333944956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,4096,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,512,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,65536,0.2450986703236898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,512,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,3584,0.012551999754375882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,51200,0.3358622127109104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,3072,0.014362666341993542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,6144,0.9465902116563586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,2560,0.06072000000211927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,32,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,1024,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,4096,0.06336266464657254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,51200,0.1802426709069146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,5120,0.0309840010272132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,256,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,6144,0.030530667967266504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,4096,0.019920888874265883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,10240,0.025552888711293537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,256,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,4096,0.0435751113626692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,4096,0.02485688858562046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,256,0.038542220989863075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,64,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,8192,0.011381333072980246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,65536,0.4443831178877089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,51200,0.08533422152201335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,2048,0.019359999232821994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,128,0.005404444618357553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,64,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,16384,0.03757333424356248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,768,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,16384,0.059207108285692006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,3072,0.12356889247894287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,10240,0.9532524744669596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,2560,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,6144,0.029920889271630183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,5120,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,256,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,12288,0.029520889123280842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,65536,0.5411635504828559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,2560,0.009935999910036722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,1024,0.010960000256697336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,512,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,7168,0.025536000728607178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,7168,0.6212204297383627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,3584,0.008748444418112436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,6144,0.044814222388797335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,2560,0.0318159990840488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,10240,0.18364444043901232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,1024,0.01750577820671929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,16384,0.06726488802168104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,2048,0.06459555361005995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,16384,0.03292799989382426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,3584,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,10240,0.7541715304056803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,1536,0.004599110947714912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,5120,0.0495777792400784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,7168,0.6123110983106825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,1024,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,256,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,4096,0.04177511069509718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,51200,0.050705777274237744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,2048,0.014747555057207743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,2560,0.014282666974597506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,1536,0.11711111333635117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,16384,0.1409484412935045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,3072,0.015847999188635085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,128,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,64,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,4096,0.03211999932924906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,1024,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,4096,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,65536,0.0503057771258884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,2048,0.02083822256988949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,7168,0.21576800611284044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,3072,0.1452124383714464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,5120,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,5120,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,6144,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,1024,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,6144,0.019303111566437613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,3584,0.03452622228198581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,2048,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,1536,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,8192,0.029763556189007227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,128,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,4096,0.06230311261283027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,1024,0.6122888988918728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,51200,0.3243182235293918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,6144,0.035756445593304105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,8192,0.17794756094614664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,51200,0.05039644572469923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,768,0.2092648877037896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,4096,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,4096,0.035995556248558894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,8192,0.014345778359307183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,4096,0.0690408878856235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,256,0.0277804434299469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,64,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,8192,0.06237510840098063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,65536,0.17384444342719185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,16384,0.032829334338506065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,65536,0.17024978001912436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,6144,0.08039022154278226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,1536,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,1536,0.021928000781271193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,7168,0.028280887338850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,4096,0.015999111864301894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,512,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,8192,0.014581332604090372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,768,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,3584,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,256,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,4096,0.019827556278970506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,4096,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,2048,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,8192,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,4096,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,12288,3.724981519911024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,3072,0.021477333373493616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,51200,0.14878933959537083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,10240,0.2700008816189236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,2560,0.02437955637772878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,16384,0.046077334218555026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,64,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,65536,0.06007288561926948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,128,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,3584,0.015587555037604438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,16384,0.11053422424528335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,4096,0.06468088759316339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,512,0.057598219977484814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,3072,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,8192,0.047760887278450855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,7168,0.016873778568373788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,2048,0.08009244336022271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,12288,0.03825599948565165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,512,0.04221866528193156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,3584,0.005193777796294955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,12288,0.08270933230717976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,10240,0.015280000037617154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,64,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,3584,0.005717333406209946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,768,0.013017777767446307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,128,0.0017884444031450485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,1536,0.013011555704805585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,8192,0.02632444434695774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,3584,0.03581510980923971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,51200,0.17675111028883192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,7168,0.1432160006629096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,10240,0.07310666640599568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,4096,0.7237084706624349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,512,0.0035253332720862497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,512,0.04077688852945963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,10240,0.05899822049670749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,768,0.023224888576401606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,512,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,10240,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,1024,0.002887111157178879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,256,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,2048,0.021968000464969214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,8192,0.4464329083760579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,12288,0.020042666130595736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,51200,0.05264355407820808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,6144,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,8192,0.08227644364039104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,65536,0.44775644938151044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,12288,0.3500888877444797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,768,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,5120,0.013659555878904132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,65536,0.0366248885790507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,3584,0.013383111192120446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,3584,0.014127110441525778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,3072,0.012680889003806643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,1536,0.057647109031677246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,16384,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,256,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,8192,0.020783111453056335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,768,0.11454044448004828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,7168,0.030686222844653662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,65536,0.14543644587198892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,1536,0.0035342222286595237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,32,0.012806221842765808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,256,0.030896888838873968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,65536,0.8507111337449816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,128,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,7168,0.1659288936191135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,2560,0.010807111031479306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,4096,0.03473688827620612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,3072,0.7604942321777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,8192,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,1536,0.12761955791049534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,51200,0.05678666962517632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,128,0.031856000423431396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,512,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,3584,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,10240,0.08731910917494033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,16384,1.0618435541788738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,512,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,8192,0.022970666488011677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,6144,0.19408710797627768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,4096,0.09169422255622016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,3584,0.014112888111008538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,51200,0.1850959989759657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,6144,0.025497777594460383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,12288,0.030943999687830608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,5120,0.026171555121739704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,2560,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,512,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,7168,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,3584,0.03433422247568766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,3072,0.028713779317008123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,4096,0.030079109801186457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,51200,0.1430399947696262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,64,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,1024,0.00572622236278322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,768,0.027609777119424608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,2560,0.013184000220563678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,8192,0.02057422200838725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,2560,0.05580000082651774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,12288,0.013487111363146039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,12288,0.019160888261265226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,7168,0.06344711118274264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,7168,0.020099555452664692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,512,0.015430221954981485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,768,0.02886311213175456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,51200,1.236313819885254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,10240,0.014195554786258273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,256,0.021384888225131567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,128,0.19296622276306152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,12288,0.04018933243221707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,4096,0.02875022093454997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,5120,0.12746311558617487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,3584,0.05280977818701002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,768,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,4096,0.027606222364637587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,3584,0.1060826645957099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,256,0.1250115500556098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,3072,0.05615644322501289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,3072,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,1024,0.15104356076982287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,2560,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,7168,0.02775555517938402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,1536,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,6144,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,7168,0.03436622354719374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,128,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,6144,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,10240,0.5065884590148926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,512,0.004023999803596073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,7168,0.039847112364239164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,2048,0.012481778032249875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,1536,0.07582577731874254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,32,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,7168,0.09838133388095432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,2560,0.007505777809354994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,65536,0.04839644498295254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,16384,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,512,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,512,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,1536,0.6132444275750054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,512,0.009111999637550777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,8192,0.06151377492480808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,12288,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,8192,0.05122488737106323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,65536,0.42951732211642796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,768,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,1536,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,2560,0.06424178017510308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,51200,0.17626133229997423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,512,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,128,0.014860444598727755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,512,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,3584,0.26815467410617405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,5120,0.04089511103100247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,1024,0.022240888741281297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,128,0.028118222951889038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,1024,0.005636444522274866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,768,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,8192,0.019596444235907663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,4096,0.012837332983811697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,3072,0.05353422297371758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,2560,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,16384,0.3714257876078288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,1536,0.0644764436615838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,4096,0.0059279998143513995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,5120,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,10240,0.009716444545321995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,32,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,10240,0.11534489525689019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,65536,0.07223466369840834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,3584,0.05183644427193535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,10240,0.047432889540990196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,3072,0.01387466655837165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,1024,0.04891022377543979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,2048,0.018579555882347953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,512,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,10240,0.13040622075398764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,65536,0.07210399707158406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,768,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,3072,0.11532978216807048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,512,0.09684177902009751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,2560,0.07876266373528375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,512,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,1536,0.015832000308566622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,12288,0.03752266698413425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,4096,0.017715555098321702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,5120,0.024282667371961806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,8192,0.05882222122616238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,4096,0.01683200067943997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,1536,0.029184887806574505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,2048,0.030328889687856037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,5120,0.016605334149466622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,6144,0.0465191106001536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,128,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,128,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,1536,0.2094515562057495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,12288,0.276488012737698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,3584,0.06250222524007161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,256,0.009840000006887648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,2560,0.021298666795094807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,12288,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,3072,0.008283555507659912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,3072,0.5691928863525391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,64,0.013668444421556262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,1024,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,32,0.0021439999755885866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,2048,0.09893511401282416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,512,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,128,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,2560,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,4096,0.14320711294809976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,51200,0.8111377822028266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,10240,0.017210667332013447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,6144,0.018554665976100497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,768,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,1536,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,12288,0.3666702111562093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,16384,0.02251022226280636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,768,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,51200,0.06390488810009427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,1024,0.014635556273990206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,3072,0.06139289008246528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,32,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,128,0.1474284463458591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,3072,0.017083555459976196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,5120,0.013047999805874295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,12288,0.2379217677646213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,4096,0.07424444622463651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,6144,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,6144,0.03462222218513489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,4096,0.009169778062237633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,5120,0.06373866399129231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,32,0.034156444999906756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,512,0.01870133313867781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,3072,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,8192,0.35117332140604657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,5120,0.009628444082207149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,8192,0.013915555344687568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,3072,0.02268444498380025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,3584,0.014073777529928418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,7168,0.08923644489712185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,4096,0.0284222231970893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,65536,0.06759644216961332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,2048,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,3584,0.0618835555182563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,12288,0.25431021054585773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,2560,0.058654222223493785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,8192,0.010783111055692038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,1024,0.054479112227757774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,6144,0.023847111397319373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,7168,0.061478224065568715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,1536,0.06323022312588163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,2048,0.012816000315878125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,2048,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,5120,0.0295724438296424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,2048,0.018944000204404194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,3584,0.08770044644673665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,12288,0.030600888861550227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,12288,0.2527982128991021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,6144,0.016093333562215168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,5120,0.023237332701683044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,128,0.1533377832836575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,3072,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,3584,0.09579733345243667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,2560,0.020272889071040683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,6144,0.12320889366997613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,5120,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,12288,0.12952266799079046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,12288,0.016584888100624084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,65536,0.04897155695491367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,8192,0.02646933330429925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,8192,0.01130311108297772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,1536,0.02699377801683214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,2560,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,51200,0.03144355614980062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,16384,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,1536,0.10947200324800278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,512,0.015185778339703878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,12288,1.1043626997205946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,512,0.008052444292439355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,3584,0.0118231111102634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,6144,0.07839644617504544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,3072,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,2560,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,3584,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,128,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,2560,0.021751110752423603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,768,0.028789334826999243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,768,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,1024,0.01720622181892395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,2560,0.02113688819938236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,5120,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,16384,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,768,0.009576000273227692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,768,0.015786666009161208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,4096,0.017666666044129264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,128,0.009093333449628618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,16384,0.4452488687303331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,512,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,1536,0.12204710642496745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,3072,0.05182133449448479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,2048,0.027679999669392902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,5120,0.008926221893893348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,32,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,8192,0.9240621990627713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,512,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,1024,0.022105778257052105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,3584,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,51200,0.4321902328067356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,12288,0.03644533289803399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,10240,0.03685155510902405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,65536,0.15959821807013616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,16384,0.037893331713146634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,5120,0.03257866700490316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,65536,0.29611288176642525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,6144,0.012135110795497894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,1536,0.043056888712777026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,12288,0.8568711280822754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,768,0.005633777628342311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,256,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,1024,0.006063999815119638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,6144,1.2740835613674588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,2048,0.020939555433061387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,3072,0.06479377879036798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,16384,0.09001600080066258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,128,0.08665955728954738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,3584,0.016023111012246873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,16384,0.13308000564575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,5120,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,12288,0.06564088662465413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,256,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,7168,0.011531555818186866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,768,0.02377333409256405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,2560,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,128,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,5120,0.0957875582906935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,3584,0.020638222495714825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,3072,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,768,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,10240,0.07721244626575045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,3072,0.01980799933274587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,128,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,32,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,2560,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,12288,0.020045333438449435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,3584,0.7112977769639758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,3072,0.04604533314704895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,51200,0.3254471090104845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,5120,0.006701333241330252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,7168,0.05014222198062473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,3072,0.0074444446298811175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,6144,0.26238489151000977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,128,0.018593778212865193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,16384,0.04492622282769945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,128,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,128,0.00666133314371109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,16384,0.038527111212412514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,8192,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,256,0.059281779660118945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,10240,0.023327999644809302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,768,0.00453599997692638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,512,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,51200,0.12135288450453018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,7168,0.014112000664075216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,8192,0.014148443937301636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,3072,0.009576888548003303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,7168,0.04934755629963345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,512,0.012568888564904531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,3584,0.016125332978036668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,2560,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,256,0.01692800058258904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,1536,0.008251555263996124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,768,0.03570666578080919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,128,0.010087110930018956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,4096,1.5624088711208766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,5120,0.041830221811930336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,7168,0.023981332778930664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,6144,0.26516710387335884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,256,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,10240,0.025733333494928148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,2048,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,1024,0.025880888104438782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,768,0.01923377811908722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,768,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,3072,0.08923021952311198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,10240,0.026621333426899377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,51200,0.2175004482269287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,51200,0.3429911136627197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,256,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,65536,0.06338577800326876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,2560,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,3584,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,10240,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,1536,0.020425778296258714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,1024,0.04827466607093811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,2560,0.09507733583450317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,128,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,768,0.034323556555642024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,10240,0.045089777972963124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,1536,0.016931555337376065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,10240,0.33876623047722715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,8192,0.013038222160604266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,3072,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,12288,0.03487733337614272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,3584,0.027406222290462915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,2048,0.0273297776778539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,512,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,5120,0.010022222167915767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,128,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,2048,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,512,0.004853333450025982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,8192,0.027970665031009253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,4096,0.023924445112546284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,768,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,12288,0.026867555247412786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,768,0.0058613332609335584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,51200,1.0810204611884224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,2560,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,128,0.010109333528412713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,128,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,2048,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,7168,0.03850311040878296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,32,0.003499555504984326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,7168,4096,0.009718221922715506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,3584,0.03495999839570787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,3584,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,8192,0.08745421965916951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,1024,0.017320000463061862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,2048,0.02552177839808994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,7168,0.03534577621353997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,1536,0.008176888856622908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,16384,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,8192,0.0424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,512,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,4096,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,65536,0.1393680042690701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,64,0.006824888702895906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,16384,0.12770399782392713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,7168,0.031028442912631567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,3072,0.005426666802830166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,6144,0.0722480018933614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,128,0.010796444283591377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,3072,0.03786844346258376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,2048,0.049294223388036094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,2048,0.15369689464569092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,1024,0.0034977777136696708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,5120,0.012812444733248817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,2560,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,8192,0.2973955472310384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,16384,0.04822133315934075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,512,0.019499555230140686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,8192,0.03321333395110236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,51200,0.5989199744330512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,1024,0.02648266653219859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,3072,0.013381332986884646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,2560,0.8038640022277832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,4096,0.016068443655967712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,128,0.014691554837756686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,8192,0.3523555596669515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,128,0.21009955141279432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,65536,0.6912008921305338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,4096,0.014284445179833306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,12288,0.013898667362001208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,768,0.10783111386828953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,256,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,256,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,3072,0.018166222506099276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,1024,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,5120,0.2835777865515815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,768,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,4096,0.008370666868156856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,32,0.011552000211344825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,51200,0.2646968894534641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,3072,0.006888000087605582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,16384,0.01577422188387977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,2048,0.0336444444126553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,1536,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,3072,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,16384,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,8192,0.11376444498697917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,3072,0.013872000078360239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,3584,0.14596977498796251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,2560,0.22619999779595268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,1536,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,8192,0.11225956016116673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,1024,0.012999999854299756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,8192,0.3020862208472358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,3584,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,2048,0.031137777699364558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,8192,0.030674666166305542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,7168,0.036511109934912786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,16384,0.05923822191026476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,32,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,4096,0.02568799919552273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,51200,0.5134080251057943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,10240,0.08364444308810765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,128,0.007692444655630324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,1536,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,2560,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,4096,0.3090657658047146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,2048,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,2048,0.021597334080272253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,3584,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,4096,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,5120,0.1136453284157647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,32,0.13717333475748697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,7168,0.09997777806388007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,4096,0.026896889011065166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,128,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,1024,0.003559111307064692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,128,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,4096,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,65536,0.1586951149834527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,5120,0.011872000164455838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,10240,0.13088799847496882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,5120,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,768,0.033668445216284856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,1536,0.015300444430775113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,8192,0.1046595573425293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,3584,0.0239982224173016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,16384,0.5693528917100694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,3584,0.00846400029129452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,1024,0.18612355656094023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,10240,0.026708443959554035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,1536,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,256,0.07406310902701484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,3072,0.027170666389995154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,8192,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,65536,0.09213422404395209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,64,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,1024,0.15790488984849718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,32,0.00572533326016532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,65536,0.07835021946165296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,2560,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,3584,0.03453333179155985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,6144,0.058596445454491504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,7168,0.015778667396969266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,3584,0.007537778053018782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,6144,0.20894132720099556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,3584,0.06259200308057997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,768,0.03093866507212321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,2048,0.019521777828534443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,3584,0.06860355536142985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,2560,0.028896000650193956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,64,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,4096,0.017866666118303936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,4096,0.023426666855812073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,65536,0.06637510988447401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,10240,0.583814197116428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,128,0.02051199972629547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,5120,0.027443556321991816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,768,0.03240177697605557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,6144,0.09756000174416436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,64,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,10240,0.1394311057196723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,2560,0.03273422188229031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,7168,0.02097333305411869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,8192,0.08979555633332993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,2048,0.05276088913281759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,12288,0.041153778632481895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,8192,0.02128000060717265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,10240,0.03682044479582045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,5120,0.0468444459968143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,7168,0.03329866793420579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,8192,0.04790844519933065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,3584,0.011600000162919363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,5120,0.012732444538010491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,512,0.02919911013709174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,4096,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,12288,0.09063822031021118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,65536,1.2322400410970051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,2048,0.02402044501569536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,3072,0.025791111919615004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,2560,0.016378665963808697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,8192,0.08560888634787665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,6144,0.026320000489552815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,512,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,1024,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,65536,0.13608977529737684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,1536,0.015484443969196744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,1536,0.20552710692087808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,1536,0.006337777607970768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,1536,0.02440266642305586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,3072,0.03273244367705451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,7168,0.03677599959903293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,128,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,3584,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,5120,0.047416889005237155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,5120,0.008512000242869059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,256,0.04396888944837782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,2560,0.020978666014141504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,128,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,1536,0.03458222084575229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,768,0.05273511012395223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,7168,0.03725333346260919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,65536,0.0551528897550371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,10240,0.1107591125700209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,768,0.0129439996348487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,65536,0.5510835647583008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,128,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,32,0.027441778116756018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,256,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,12288,0.04130222068892585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,2560,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,1536,0.08219289117389254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,10240,0.10105866856045193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,3584,0.016488000750541687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,3072,0.019847111569510568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,4096,0.011424000064531961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,65536,0.3300284544626872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,7168,0.05008800162209404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,8192,0.019267555740144517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,768,0.0028408887899584244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,512,0.005445333404673471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,128,0.009890666438473595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,3072,0.05554044577810499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,7168,0.011086222198274402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,768,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,7168,0.011711111499203576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,3584,0.030517333083682593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,51200,0.05973155630959404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,10240,0.06601155466503568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,2048,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,64,0.16313955518934461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,1024,0.1672044462627835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,65536,0.45657867855495876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,6144,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,3584,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,3072,0.05552177627881368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,16384,0.04144710964626736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,51200,0.9928311241997613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,7168,0.21376089255015054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,10240,0.05038933290375603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,5120,0.06735377841525607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,6144,0.034576889541414045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,768,0.021480000681347314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,7168,0.019945777124828763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,5120,0.18715555138058135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,8192,0.04504088891877068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,3584,0.012077333198653327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,8192,0.021948445174429152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,3072,0.038372443781958684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,64,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,12288,0.02935733397801717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,64,0.02109955582353804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,3584,0.09582044680913289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,128,0.0025351111673646504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,128,0.009574222067991892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,1024,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,32,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,12288,0.08506933185789321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,12288,0.043431109852261014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,8192,0.04184444414244758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,4096,0.004583999928500918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,2560,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,3584,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,512,0.009525333013799455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,2048,0.00941066692272822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,1024,0.015117333994971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,256,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,256,0.05911822451485527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,8192,0.2198613352245755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,512,0.02253155575858222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,1024,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,1536,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,2048,0.005147555636035072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,128,0.005744888964626525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,3584,0.2373164494832357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,4096,0.06332177586025663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,128,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,3584,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,51200,0.434695111380683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,8192,0.5640702247619629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,8192,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,65536,0.1929697725507948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,1536,0.01964000033007728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,12288,0.02865155537923177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,10240,0.026405332816971674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,7168,0.020995555652512446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,65536,0.07534311215082805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,7168,0.02159999973244137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,256,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,51200,0.058989332781897656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,7168,0.07218133078681098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,2560,0.026769777139027912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,2560,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,4096,0.2440213362375895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,256,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,51200,1.3013262218899198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,768,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,2048,0.03421777817938063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,64,0.02397866712676154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,3584,0.08807110786437988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,2560,0.047776887814203896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,51200,0.6076231002807617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,2560,0.019335110982259113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,10240,0.04347022374471029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,32,0.04609688785341051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,256,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,256,0.02438755498992072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,5120,0.02072266737620036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,3584,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,3584,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,5120,0.22646045684814453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,6144,0.10342933071984185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,4096,0.09752088785171509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,6144,0.07204266389211018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,2048,0.015202666322390238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,2560,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,16384,0.06767466995451185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,10240,0.03477688961558872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,2048,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,1536,0.02425777746571435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,256,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,512,0.1983333296245999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,12288,0.048792888720830284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,256,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,1024,0.03685599896642897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,16384,0.2626844512091743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,512,0.002789333255754577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,5120,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,16384,0.02316711180739933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,3584,0.025955556167496577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,3072,0.031178666485680476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,6144,0.012219555675983429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,64,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,10240,1.0224328570895724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,5120,0.008654221892356873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,8192,0.026368889543745253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,128,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,2048,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,2048,0.004863111095296012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,2048,0.006360888895061281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,256,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,256,0.00332266671790017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,32,0.01311288856797748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,6144,0.01000711073478063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,12288,0.5766506724887425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,3584,0.03566222058402167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,7168,0.2620755566491021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,7168,0.4505493375990126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,2560,0.07765777905782063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,65536,0.05325422353214688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,2048,0.02907555633121067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,8192,0.015011555618709989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,32,0.12166133191850449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,16384,0.18586488564809164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,64,0.012952889005343119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,2560,0.01257777793539895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,5120,0.015642666154437594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,8192,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,10240,0.08807644579145645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,512,0.0030986666679382324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,6144,0.010806221928861407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,5120,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,512,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,3072,0.07313422362009685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,12288,0.032227555910746254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,10240,0.016673778494199116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,16384,0.031006223625606958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,8192,0.010785777535703448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,16384,0.019295111298561096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,256,0.0022213332768943575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,6144,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,1024,0.0063786668082078295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,51200,0.049031111929151744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,7168,0.2959333260854085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,10240,0.018167111608717177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,3072,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,4096,0.01795022189617157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,32,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,16384,0.024150222539901733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,5120,0.9364453421698676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,2560,0.012272000312805176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,3072,0.05024888780381945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,6144,0.01772888832622104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,1536,0.010314666562610203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,5120,0.06531466378106011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,1024,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,64,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,3072,0.03201155530081855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,256,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,512,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,128,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,1024,0.028821332587136164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,512,0.06422844198015001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,3072,0.016227554943826463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,512,0.0031946665710873077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,768,0.0032222221295038858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,3072,0.007376000285148621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,16384,1.0232906341552734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,768,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,65536,0.37471821573045516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,16384,0.02289955483542548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,512,0.009924444059530893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,8192,0.014371555712487964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,5120,0.11793333954281277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,7168,0.014360888136757744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,8192,0.205749339527554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,51200,13.52595435248481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,5120,0.6560737821790907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,3072,0.01819644371668498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,16384,0.18404444058736166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,256,0.010771555205186209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,2048,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,3072,0.028315554062525432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,5120,0.1434453328450521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,768,0.8930550681220161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,5120,0.008832000195980072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,7168,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,5120,0.025303110480308533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,1024,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,8192,0.0326186650329166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,1536,0.022293332550260756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,8192,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,128,0.0251208891471227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,1024,0.003142222141226133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,65536,0.06230044364929199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,10240,0.4426604376898871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,3584,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,12288,0.021684444612926904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,10240,0.013136000268989138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,3584,0.14537689420912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,7168,0.02365511159102122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,3584,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,16384,0.08584888776143391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,7168,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,512,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,768,0.013349333571063148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,1024,0.018801778554916382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,256,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,12288,0.028515554136700098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,64,0.012823111481136747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,16384,0.16644533475240073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,51200,3.140725241767036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,8192,0.10975466834174262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,8192,0.02437244521247016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,768,0.1154515610800849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,2048,0.06216710805892944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,51200,0.0780586666531033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,65536,1.7559786902533636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,1024,0.0447324448161655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,512,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,51200,0.04383111000061035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,2560,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,2048,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,7168,0.37473511695861816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,4096,1.6759724087185328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,128,0.07389511002434625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,256,0.005506666584147348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,768,0.013122667041089801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,512,0.03312266535229153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,10240,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,65536,0.020232000284724765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,512,0.0069777775141927935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,3072,0.035170667701297335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,16384,0.06538133488761054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,1024,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,65536,0.14720000161064997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,768,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,12288,0.013520888984203339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,51200,0.037979556454552546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,2560,0.07929333051045735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,65536,0.0651822222603692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,12288,0.020076443751653034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,256,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,2048,0.05195822318394979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,2048,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,2560,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,7168,0.1218186616897583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,16384,0.0461662213007609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,10240,0.02898400028546651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,256,0.023807111713621352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,128,0.014838222000333997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,5120,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,4096,0.02604177759753333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,4096,0.035960889524883695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,3072,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,12288,0.020489777127901714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,8192,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,6144,0.03144177794456482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,7168,0.03475111060672336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,2560,0.4396213425530328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,7168,0.02888800038231744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,4096,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,7168,1.6856124666002061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,6144,2.2107777065700955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,5120,0.021570665968788996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,1024,0.0038817777401871155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,12288,0.019931554794311523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,10240,0.10619288682937622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,2048,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,256,0.006786666810512543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,256,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,51200,0.5760693550109863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,16384,0.9418915642632378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,2560,0.04291022155019972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,5120,0.5080604553222656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,12288,0.16774400075276694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,512,0.05617688761817085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,12288,1.483057763841417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,512,0.024596444434589807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,10240,0.586657789018419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,12288,1.8776018354627821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,2048,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,4096,12288,0.053724444574779935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,2560,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,2560,0.7640577952067057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,1024,0.8951075871785482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,65536,0.2802657816145155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,3584,0.014543111125628153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,128,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,10240,0.11518932713402642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,4096,0.05095110999213325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,2560,0.01787911189926995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,10240,0.04974133438534207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,10240,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,2560,0.018624888526068795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,2048,0.07892622550328572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,16384,0.0664657751719157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,12288,0.062936888800727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,1536,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,1024,0.011715555356608497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,2560,0.03592177894380357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,65536,0.3820195462968614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,32,0.007060444189442529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,1024,0.006688888702127669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,512,0.015209777487648858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,12288,0.12315111690097386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,1024,0.007459555235173967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,4096,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,512,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,65536,0.18501688374413383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,12288,0.04397333330578274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,51200,0.30721065733167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,65536,0.21211822827657065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,768,0.02035288843843672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,768,0.025638222694396973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,3072,0.04635644290182325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,3072,0.09238400061925252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,128,0.015173332558737861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,65536,0.21029244528876412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,51200,0.05004444387223986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,6144,0.08363378047943115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,2560,0.009974222216341231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,32,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,51200,0.038330667548709445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,65536,0.02328355610370636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,6144,1.45233334435357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,65536,2.0607005225287542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,768,0.011066666907734342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,10240,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,7168,0.06389422549141778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,16384,3.7322506374782987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,10240,0.048503110806147255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,256,0.00564533347884814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,16384,0.5876444180806478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,4096,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,4096,0.11311111185285781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,65536,0.3061724503835042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,1024,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,5120,0.012778666284349231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,64,0.04845155609978569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,4096,0.008105778031879002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,8192,0.055000000529819064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,6144,0.02404088940885332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,32,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,768,0.018618666463428073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,2560,0.024680000212457445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,128,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,6144,0.0435697767469618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,1536,0.003536000019974179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,512,0.020868443780475195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,2048,0.048707554737726845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,768,0.03057510985268487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,3072,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,256,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,3584,0.028751999139785767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,51200,2.442201826307509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,4096,0.021765333082940843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,768,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,256,0.018453333112928603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,4096,0.04645688997374641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,12288,2.694009780883789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,51200,0.7414115799797906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,51200,1.115195592244466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,12288,0.030905776553683813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,65536,0.5254213545057509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,2048,0.03159822358025445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,4096,0.018968888454967074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,7168,0.021996445126003687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,512,0.002720000015364753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,2048,0.025757332642873127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,51200,0.4842053519354926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,2560,0.07491911119884916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,5120,0.01074222226937612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,5120,0.20856711599561903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,768,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,768,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,128,0.01721955504682329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,7168,0.02640177806218465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,65536,0.06975199778874715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,65536,0.09337155686484443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,65536,0.06848444541295369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,10240,0.015549333559142219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,128,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,5120,0.11394933859507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,5120,0.004323555363549126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,4096,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,256,0.023694222172101338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,3072,0.03677155574162801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,1024,0.04085155659251743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,1024,0.009415999882751042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,6144,0.3926551077100966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,2560,0.012359110845459832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,512,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,6144,0.029564443561765883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,7168,0.0532915559079912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,1536,0.006151111175616582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,2048,0.020840888222058613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,2048,0.03459466828240289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,4096,0.015789333317014907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,4096,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,512,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,3584,0.02307022280163235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,1024,0.01833066675398085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,4096,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,12288,0.02998044424586826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,3584,0.029376887612872656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,5120,0.01657777859105004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,10240,0.08457777897516887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,3072,0.012789333032237159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,2560,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,1024,0.0625600020090739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,16384,0.0195777780479855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,51200,0.0568444463941786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,16384,0.014711111783981323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,4096,0.040836443503697716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,128,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,12288,0.020798222886191476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,65536,0.6918222109476725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,1536,0.006151999864313338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,512,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,6144,0.03570400012864007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,3584,0.02223288847340478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,7168,0.2591208881802029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,5120,0.05063199996948242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,16384,0.1698613299263848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,12288,0.022388445006476507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,16384,0.11024622122446696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,512,0.03595733311441209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,7168,2.728212356567383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,1536,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,768,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,10240,0.024737776981459722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,3584,0.06005511018964979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,2560,0.3905591169993083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,4096,0.21154399712880453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,16384,0.03502577874395583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,65536,0.23930221133761934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,128,0.0063999998900625445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,10240,0.0967751079135471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,16384,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,8192,0.015505777464972602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,2560,0.025063110722435847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,65536,0.3047608799404568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,128,0.002543111021320025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,12288,2.113520940144857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,2048,0.009763555394278632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,2048,0.03278311093648275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,1536,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,4096,0.012278222375445895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,4096,0.024396444360415142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,256,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,128,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,51200,0.16171822283003065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,128,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,1536,0.0064222220745351576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,16384,0.08256622155507405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,16384,0.04122044311629401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,3584,0.004551110996140374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,256,0.08053511381149292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,51200,0.11379199557834202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,768,0.014889778362380134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,2048,0.047374222013685435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,128,0.00965244405799442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,2048,0.010398222340477837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,7168,0.01685333251953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,64,0.0018800000349680583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,51200,0.14052355289459229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,768,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,2560,0.015102222561836243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,16384,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,51200,0.11208532916174994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,10240,0.020257777637905545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,32,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,12288,0.017729777428838942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,4096,0.04626399940914578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,5120,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,1024,0.003220444545149803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,2560,0.01367377738157908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,64,0.012605333493815528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,7168,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,768,0.0029893333299292457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,10240,0.012539555629094442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,12288,0.15913955370585123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,10240,0.049936887290742665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,1536,0.0721857746442159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,2560,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,32,0.0022026666750510535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,51200,0.12452977233462864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,128,0.01612444387541877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,512,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,1536,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,8192,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,51200,1.4760168923272028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,51200,256,0.055594669448004834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,16384,0.06023199690712822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,1024,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,5120,0.5065671073065864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,32,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,2560,0.04279911186960009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,8192,0.17109866937001547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,10240,0.01167377746767468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,7168,0.018174222773975797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,6144,0.15699022346072727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,512,0.11135378148820664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,51200,0.03802488909827338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,3584,0.20035733116997612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,51200,12.211941189236112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,3584,0.010994666980372535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,4096,0.07785599761539035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,128,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,3072,0.034098668230904475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,4096,0.014183999763594734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,64,0.00850399997499254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,16384,0.28405600123935276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,2560,0.07681244611740112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,6144,0.018596443865034316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,256,0.044235554006364614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,4096,0.024205333656734888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,1536,0.09301600191328262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,2048,0.01369866645998425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,7168,0.0311155551009708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,4096,0.019381332728597853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,5120,0.319943110148112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,768,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,10240,0.017277333471510146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,768,0.071633776028951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,32,0.010656000011497073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,3072,0.7431786855061849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,768,0.00600888869828648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,4096,0.09501866499582927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,256,0.02186044388347202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,6144,0.1534151103761461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,768,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,51200,0.48047468397352433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,128,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,1536,0.015815110670195687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,2048,0.07863199710845947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,2048,0.056919998592800565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,8192,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,768,0.013370667066839008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,128,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,5120,0.4572764502631293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,4096,0.02551199992497762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,6144,0.03156355685657925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,7168,0.08606044451395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,256,0.008372444245550368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,256,0.03533066643608941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,3072,0.17103200488620332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,3072,0.07623822159237333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,1024,0.09642577833599514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,16384,0.3229706552293566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,7168,0.14480800098843044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,1536,0.01055466632048289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,6144,0.1502479977077908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,12288,0.04236622320281135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,2048,0.015453333655993143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,3584,0.043712000052134194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,1024,0.0028773333049482773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,65536,4.102919260660808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,5120,0.879842652214898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,51200,1.015507592095269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,10240,0.21667467223273384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,51200,0.0586986674202813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,256,0.0140346669488483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,768,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,2560,0.03240622083346049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,64,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,4096,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,3072,0.03624533282385932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,12288,0.016361777981122334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,1024,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,64,0.0032666667054096856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,65536,0.11419910854763454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,65536,0.06994489166471693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,16384,0.14099021752675375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,4096,0.12503733899858263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,12288,0.06836622291141085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,512,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,10240,0.01369955556260215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,1536,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,1536,0.05754399961895413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,128,0.02444533341460758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,128,0.00508622204264005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,3584,0.007729777859316931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,16384,2.286717308892144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,2560,0.008375110725561777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,3072,0.016695110334290397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,2560,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,256,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,6144,0.05235022306442261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,32,0.36758489078945583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,4096,0.054916444751951426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,768,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,51200,0.4013875590430365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,1536,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,10240,0.288595543967353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,4096,0.016542222764756944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,8192,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,12288,2048,0.012140444583363004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,256,0.020622221959961783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,51200,0.1399679978688558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,1024,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,1024,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,3584,0.02847644355561998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,5120,0.012269333004951477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,7168,0.03348799877696567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,512,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,7168,0.010871111518806882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,3584,0.11238133907318115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,6144,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,65536,0.11155554983350967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,2048,0.04950755503442553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,10240,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,1536,0.015433778365453085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,51200,0.11515466372172038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,128,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,3072,0.09159288803736369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,6144,0.22562132941351998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,65536,1.7717475891113281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,256,0.0032462223122517266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,12288,0.04131288992034064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,3584,0.05883466535144382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,16384,0.017163554827372234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,16384,0.07433866792254977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,2048,0.004596444467703502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,3072,0.01332711097266939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,6144,0.015585778488053216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,1024,0.049787554475996226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,32,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,65536,0.057166225380367704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,4096,0.012306667036480375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,1536,0.05391022231843737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,12288,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,768,0.017345777816242643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,256,0.15102756023406982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,5120,0.026231999198595684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,512,0.02348533272743225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,4096,0.2204586664835612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,5120,0.08007466793060303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,5120,0.06451111369662814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,64,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,16384,0.0292231109407213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,256,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,6144,0.10918577512105306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,1024,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,10240,0.13743110497792563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,6144,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,256,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,6144,0.012057777908113269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,768,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,128,0.020101333657900494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,1536,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,51200,0.1294026639726427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,7168,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,10240,0.9930684831407336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,10240,0.26671377817789715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,3072,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,65536,0.07148622141944037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,51200,0.534830199347602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,512,0.01960444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,4096,0.020098666350046795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,32,0.009731555150614845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,768,1.2357519997490776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,3072,0.27793600824144155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,6144,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,10240,0.045496000183953174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,10240,0.07055733601252238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,12288,0.013068444199032254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,128,0.024512888656722173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,4096,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,8192,0.015343111422326831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,2560,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,51200,0.15238222810957167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,1024,0.013847110999955071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,51200,0.048130667871899076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,3072,0.0047377778424157035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,1536,0.0063697778516345555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,3584,0.043493333790037364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,1536,0.004040889028045866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,7168,0.02918044394916958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,1536,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,16384,0.2366968790690104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,51200,0.23884799745347765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,3072,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,512,0.02416177756256527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,128,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,2560,0.006992888947327931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,512,0.011011555790901184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,64,0.057314667436811656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,8192,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,128,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,768,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,128,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,3584,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,16384,0.006865777903132969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,1024,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,128,0.003162666741344664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,1536,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,2560,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,16384,0.02304977840847439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,128,0.07742577791213989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,2048,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,64,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,256,0.022786666949590046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,51200,0.36019643147786456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,16384,0.03492000036769443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,256,0.02362755603260464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,2048,0.07611378033955891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,10240,0.03728444377581278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,5120,0.10206222534179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,32,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,3584,0.008279110822412703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,256,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,12288,0.057168887721167676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,7168,0.01106844428512785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,1024,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,16384,1.0366204579671223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,10240,0.10316889153586493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,51200,0.17724977599249947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,51200,0.046239111158582896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,128,0.0024177777684397167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,128,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,128,0.021399999658266704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,1024,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,768,0.002999111182159848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,51200,2.9322427113850913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,3072,0.02954933378431532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,128,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,256,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,16384,0.5703537729051378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,128,0.015266666809717814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,4096,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,768,0.005663110978073544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,768,0.06201777855555216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,1536,0.029230223761664495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,4096,0.008268444074524773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,12288,0.0185244447655148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,3072,0.09508622354931301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,10240,0.07209599680370755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,10240,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,6144,0.014706666270891825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,51200,0.0529022216796875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,3072,0.011256888508796692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,4096,0.11600532796647813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,64,0.004263999975389904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,8192,0.24922755029466417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,7168,0.011416888899273343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,4096,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,128,0.021073778470357258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,256,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,10240,0.047071999973720975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,128,0.018413333429230582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,128,0.013165333204799227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,4096,0.06578133503595988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,2048,0.020571556356218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,6144,0.04366933306058248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,64,0.03204444381925795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,64,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,51200,0.3062942292955187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,12288,1.1422026952107747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,6144,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,512,0.026183999247021143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,8192,0.01494311127397749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,6144,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,512,0.1448791159523858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,512,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,8192,0.003900444342030419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,3584,0.20678667227427164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,8192,0.010664888554149203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,12288,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,3584,0.034662223524517484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,2560,0.013186666700575085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,3584,0.006531555619504716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,5120,0.009404444032245213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,65536,0.13024800353580052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,64,0.6129440201653374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,3584,0.030821333328882854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,1536,0.02472977836926778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,64,0.002206222257680363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,256,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,2048,0.011151110960377587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,2560,0.060621334446801074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,3584,0.055175999800364174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,8192,0.0266977780395084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,1024,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,768,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,3584,0.05225422316127353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,16384,0.029503110382292006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,3584,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,8192,0.022625777456495497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,256,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,6144,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,10240,0.27492888768513996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,6144,0.037393778562545776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,12288,0.1336266729566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,4096,0.053026666243871055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,6144,0.10567288928561741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,2560,0.017245334055688646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,6144,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,3072,0.010458666417333813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,8192,0.05841866466734144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,1536,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,32,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,256,0.027131555808915034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,10240,0.05764800310134888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,51200,0.08493510882059734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,6144,1.1524391174316406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,2560,0.04422400063938565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,2048,0.02121155626244015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,3584,0.019858666592174105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,512,0.01718933383623759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,6144,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,1024,0.09478044509887695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,128,0.014966222974989148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,16384,0.03317244516478644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,3584,0.1736053360833062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,7168,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,6144,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,512,0.02125866711139679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,512,0.012772444221708508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,16384,0.14350933498806423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,16384,0.06362844175762601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,3584,0.03008355696996053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,128,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,768,0.07834755712085299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,51200,0.20630755689409044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,10240,0.06553244590759277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,3584,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,12288,0.03232266505559286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,3072,0.01996799972322252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,2048,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,16384,0.07097599903742473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,1536,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,2560,0.014087110757827759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,12288,0.013746666411558786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,1024,0.026363554928037856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,1024,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,768,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,5120,0.0932604471842448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,8192,0.08743911319308811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,3584,0.017308443784713745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,65536,0.4032879935370551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,16384,0.018170666363504197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,3584,0.01704266667366028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,65536,0.2628266546461317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,6144,0.12821955151028103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,1024,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,512,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,8192,0.5640222231547037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,512,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,8192,0.012714666624863943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,10240,0.15622933705647787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,7168,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,256,0.01901155544651879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,10240,0.11005066500769721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,256,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,16384,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,12288,0.17031910684373644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,3584,0.008151999778217739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,2048,0.015783111254374187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,7168,0.007249777515729268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,1536,0.07161777549319796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,16384,0.03878133495648702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,10240,0.04381777842839559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,65536,0.09371911154852973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,1024,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,5120,0.01645155582163069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,128,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,7168,0.020005333754751418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,2048,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,256,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,2048,0.01965688831276364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,768,0.06258933411704169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,128,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,5120,0.12137333552042644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,65536,0.06093421909544203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,65536,0.08006133635838826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,64,0.0027831111931138565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,12288,0.06006399790445963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,3072,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,16384,0.22628090116712782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,4096,0.06985955768161349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,1536,0.38210309876335996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,10240,0.14134755399492052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,1536,0.086952891614702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,16384,0.028612444798151653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,65536,8.46461656358507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,8192,0.7158551216125488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,256,0.0029128889242808023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,512,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,768,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,1024,0.0418488879998525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,512,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,2560,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,128,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,1536,0.012426666915416718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,32,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,5120,0.14295289251539442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,7168,0.012668444050682915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,64,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,768,0.13239644633399114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,3072,0.03852977686458164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,51200,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,1536,0.020263999700546265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,10240,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,2560,0.02664088871743944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,128,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,12288,0.020992888344658747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,2560,0.013780444032616086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,768,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,2560,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,12288,0.017159110969967313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,16384,0.5042311350504557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,7168,0.012833777401182385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,768,0.004524444540341695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,10240,0.33124621709187824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,65536,2.793780432807075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,1536,0.3630399968889024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,7168,0.024176888995700415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,3072,0.161244445376926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,1024,0.026683555708991155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,10240,0.5759991010030111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,3584,0.0945697757932875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,10240,1.5986062155829535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,6144,0.01310488912794325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,512,0.005976000179847081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,8192,0.010655110908879174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,16384,0.034210665358437434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,8192,0.2131484482023451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,256,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,5120,0.013430222041077085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,7168,0.39896000756157773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,768,0.017857778403494094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,2560,0.05637066894107395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,64,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,10240,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,3072,0.025335111551814612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,768,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,128,0.022988445229000513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,65536,0.2410222159491645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,12288,0.018174222773975797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,7168,0.023538667294714186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,5120,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,2560,0.030951112508773804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,3072,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,3072,0.032171554035610624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,3072,0.014499555031458536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,65536,0.3661484453413222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,7168,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,256,0.019780443774329293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,4096,0.023767999476856653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,8192,0.04348977737956577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,12288,0.032277332411872015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,1024,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,1024,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,16384,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,4096,0.015447111593352424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,5120,0.1591457790798611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,2560,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,768,0.0653031137254503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,65536,0.06911200284957886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,256,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,7168,0.06633689006169637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,5120,0.034556445148256086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,8192,0.015928000211715698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,3584,0.011456888582971362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,3584,0.03666222095489502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,3072,0.06634133391910128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,256,0.09842133522033691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,5120,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,32,0.006032889087994893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,65536,0.14210844039916992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,3584,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,8192,0.09327466620339288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,2560,0.022504000200165644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,2560,0.02715733316209581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,256,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,3072,0.0075777777367168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,16384,0.01613066593805949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,2048,0.06856089168124728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,5120,0.0799422197871738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,32,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,1536,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,8192,0.014663111832406787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,8192,0.02126311096880171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,3584,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,3072,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,128,0.0028231110837724474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,51200,0.7430151303609213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,128,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,3072,0.13726754983266196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,32,0.012843555874294706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,2560,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,6144,0.15722044308980307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,12288,0.06459022230572171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,12288,0.02145244512293074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,16384,0.19612977239820692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,65536,0.24228445688883463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,256,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,128,0.08602755599551731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,16384,0.046042667494879834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,12288,0.01771022213829888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,4096,0.028411553965674505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,2560,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,1536,0.020784000555674236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,10240,0.03630933496687148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,16384,0.054676443338394165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,2048,0.02835822105407715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,3584,0.07681333356433444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,16384,0.03779733180999756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,2048,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,2048,0.008264888491895463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,3072,0.03052177694108751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,65536,0.25244800249735516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,256,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,1024,0.0032097777972618737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,65536,0.13278667132059732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,512,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,7168,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,3072,0.08836533625920613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,2560,0.07950755622651842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,12288,0.04361866580115425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,3584,4.282947540283203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,3584,0.13854044013553196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,7168,0.08209866947597928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,64,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,256,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,256,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,512,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,256,0.00812977800766627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,7168,0.036074668169021606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,512,0.02308888898955451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,768,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,3584,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,51200,0.10056800312466091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,3584,0.017370666066805523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,3072,0.017048888736300997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,5120,0.040222220950656466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,3584,0.013232889274756113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,51200,0.1887955533133613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,51200,0.02064622276359134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,768,0.009232000344329411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,3584,0.009340444372759925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,1536,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,7168,0.08196355236901177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,512,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,7168,0.023423999547958374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,1536,0.004506666627195147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,12288,0.08599733644061619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,768,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,51200,0.0424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,12288,0.032461331950293645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,2048,0.03579911258485582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,7168,0.010268444816271463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,2048,0.038634667793909706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,6144,0.05222577850023905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,1536,0.014840000205569796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,12288,0.05092888739373949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,16384,0.13641511069403753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,7168,0.010021333065297868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,16384,0.0568737785021464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,32,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,2560,0.44628265168931747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,8192,0.0292195545302497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,12288,0.13047733571794298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,12288,0.012246222131782107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,2048,0.035678221119774714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,64,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,256,0.002542222125662698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,16384,0.26288800769382054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,51200,0.049676444795396596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,2560,0.012058667010731168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,3072,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,1536,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,256,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,16384,0.0203840004073249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,2560,0.007713778151406183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,7168,0.12213777171240912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,2048,0.022841778066423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,1536,0.044030222627851695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,128,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,3584,0.06754310925801595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,1024,0.017497777938842773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,4096,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,12288,0.019139554765489366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,16384,0.025145777397685583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,10240,0.05301955673429701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,3584,0.21643822722964814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,51200,1.8929990132649739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,6144,0.01364177796575758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,256,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,51200,0.09590933057996963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,2560,0.009664889011118147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,4096,0.01146488885084788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,512,0.010774222513039907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,12288,0.014463111758232117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,512,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,5120,0.008748444418112436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,2560,0.041730665498309664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,8192,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,10240,0.011749332977665795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,8192,0.010472888747851053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,16384,0.6076782014634874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,1024,0.1813546684053209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,65536,0.22139199574788412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,2048,0.004461333155632019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,768,0.017236444685194228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,8192,0.04247733288341098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,3072,0.03900711072815789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,8192,0.045091556178198926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,256,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,6144,0.04752177662319607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,1024,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,3584,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,768,0.0031022222505675424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,16384,0.01958666741847992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,768,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,32,0.003458666718668408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,128,0.005114666703674528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,12288,0.020270221763186984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,128,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,768,0.026371555195914373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,256,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,3072,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,1024,0.0273324449857076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,16384,0.018904889623324077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,6144,0.027810666296217177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,4096,0.02111022174358368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,128,0.009665778113736046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,4096,0.03334755698839823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,3584,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,7168,0.04416888952255249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,8192,0.061363557974497475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,16384,0.08937155538135105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,51200,0.08834311034944321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,7168,0.08213155799441867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,768,0.16493511199951172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,512,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,65536,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,1024,0.02852977646721734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,51200,0.36765244272020126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,3584,0.03527999917666117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,512,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,51200,0.1949297719531589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,65536,0.05944355328877767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,7168,0.019648889700571697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,768,0.018769777483410306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,512,0.017775111728244357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,1024,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,8192,0.04110399881998698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,2560,0.01643288963370853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,256,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,2048,0.39398754967583555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,768,0.017645332548353408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,2048,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,16384,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,4096,0.014284445179833306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,4096,0.04751288890838623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,51200,1.342685381571452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,3584,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,2048,0.004685333205593957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,8192,0.014007111390431723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,5120,0.009115555220180089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,32,0.0029804443733559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,256,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,8192,1.3740746180216472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,768,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,768,0.053872889942593045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,5120,0.037857777542538114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,6144,0.054904000626669995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,512,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,12288,1.2224017249213326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,5120,0.08262399832407634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,512,0.013449778159459433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,10240,0.028491556644439697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,768,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,256,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,512,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,65536,0.1449999941719903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,10240,0.028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,10240,0.07328444719314575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,768,0.024881778491867915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,65536,0.08734133508470322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,65536,0.04688177837265862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,1536,0.03750933210055033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,6144,0.02770488957564036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,32,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,8192,0.06116444534725613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,256,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,512,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,64,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,12288,0.06935644149780273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,1536,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,51200,1.9490167829725478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,768,0.021532444490326777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,512,0.02216266592343648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,16384,0.020945777495702107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,2048,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,64,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,128,0.005663110978073544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,51200,0.1557475593354967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,2560,0.011348444554540845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,6144,0.17290311389499238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,7168,0.10886133379406399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,8192,0.010676444404655032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,3072,0.008413333031866286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,3584,0.019088000059127808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,51200,2.529510286119249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,6144,0.10270222028096516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,6144,0.013642667068375481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,12288,0.03081688947147793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,768,0.01407911048995124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,2048,0.01681155628628201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,4096,0.2655937671661377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,65536,2.361647923787435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,512,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,12288,0.022587555978033278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,512,0.01428711083200243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,1024,0.003921777837806278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,2048,0.052348444859186806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,3584,0.0075786668393347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,2560,0.012479110724396177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,3072,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,2048,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,128,0.01755199995305803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,128,0.006341333190600078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,51200,0.23501510090298125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,2560,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,3584,0.02974133359061347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,768,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,4096,0.7353084352281359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,12288,0.1278631157345242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,128,0.019552888141738046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,768,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,65536,0.0758568909433153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,256,0.03424266642994351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,2560,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,4096,0.01846577723821004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,3584,0.027599111199378967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,6144,0.01955911186006334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,8192,0.020337777005301583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,5120,0.041987554894553296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,7168,0.010285333626800114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,65536,0.0958799984720018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,2048,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,512,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,5120,0.0935937762260437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,1024,0.03819200065400865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,51200,0.05155644483036465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,4096,0.011428444749779172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,1024,0.024539555112520855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,16384,0.17345066865285239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,2048,0.007149332927332983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,1024,0.005519999812046687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,6144,0.03977866636382209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,2048,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,16384,0.037977778249316745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,3584,0.022838221655951604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,3584,0.09044888946745132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,3072,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,12288,0.03392533461252848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,64,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,12288,0.05310933457480537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,3584,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,32,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,4096,0.20595999558766684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,8192,0.019920888874265883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,2560,0.02531733281082577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,32,0.003549333247873518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,8192,0.12108622656928168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,768,0.019261333677503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,4096,0.026202667090627883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,5120,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,65536,0.5083457628885905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,6144,0.1165448824564616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,7168,0.04083555605676439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,128,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,4096,0.0463137792216407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,2048,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,32,0.030468444029490154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,3072,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,1536,0.011968000067604912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,2048,0.053732444842656456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,4096,0.11106755336125691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,2560,0.007461333440409766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,65536,0.15667643811967638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,1536,0.007238222493065729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,128,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,5120,0.05279199944602119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,16384,0.47379202312893337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,1536,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,5120,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,1024,0.0307626658015781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,1024,0.2221813334359063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,12288,0.010018666585286459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,1536,0.05844533443450928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,4096,0.2616951200697157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,256,0.005672889037264719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,2560,0.024417777856190998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,10240,1.1989786359998915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,51200,0.16722310913933647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,4096,0.032681776417626276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,3584,0.010440000229411654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,8192,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,3584,0.04072355561786228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,3584,0.03593333231078254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,512,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,1024,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,3584,0.016910221841600206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,1024,0.01426577733622657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,7168,0.0392577780617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,65536,0.07294133636686537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,16384,0.10822666353649563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,3072,0.015881778465376962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,2560,0.08776266707314385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,64,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,2048,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,10240,0.10869155327479045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,512,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,2048,0.012730666332774691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,1536,0.006208000083764394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,128,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,12288,0.12957333193884954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,1024,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,7168,0.02312711046801673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,5120,0.5796480178833008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,8192,128,0.029155555698606703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,8192,1.068152003818088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,3072,0.042479998535580106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,12288,0.23430045445760092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,10240,0.35422221819559735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,128,0.019896000623703003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,2048,0.011335999601417117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,1536,0.03183733423550924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,128,0.007247111035717859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,12288,0.04299555553330315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,256,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,4096,0.0074755557709270054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,16384,0.09739466508229573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,6144,0.05124888817469279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,1024,0.03451111250453525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,8192,0.1369431151284112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,7168,0.028064890040291682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,2560,0.1506791114807129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,10240,0.04957155717743767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,32,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,5120,0.020228443874253165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,3072,0.012364444633324942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,10240,0.017111111018392775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,4096,0.039975110027525164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,12288,0.07287733422385322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,32,0.015977778368526034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,5120,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,12288,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,3072,0.12802578343285456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,3072,0.05619821945826212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,3072,0.005978666659858491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,8192,0.031712889671325684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,5120,0.018040888839297824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,65536,0.061828447712792285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,1536,0.0038684445122877755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,256,0.02036711076895396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,3584,0.09198755688137478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,7168,0.053568889697392784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,768,0.07053422265582614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,6144,0.03500977820820279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,4096,0.03559644354714288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,128,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,51200,0.18209066655900744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,51200,0.02756355537308587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,3584,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,6144,0.11136533154381646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,768,0.07467644744449191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,1024,0.016885333591037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,1024,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,2048,0.025620443953408137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,512,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,128,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,7168,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,2560,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,7168,0.049808889627456665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,2048,0.003842666745185852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,128,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,3584,0.24079200956556532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,6144,0.008741333252853816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,1536,0.008184000021881526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,3072,0.09296000003814697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,5120,0.03427111109097799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,7168,0.013221333424250284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,5120,0.038888000779681735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,2048,0.007713778151406183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,768,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,512,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,32,0.02443466583887736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,16384,0.07959377765655518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,1536,0.016196444630622864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,16384,0.17222400506337485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,2560,0.07072799735599093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,1536,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,6144,0.08969777822494507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,256,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,16384,1.3073707156711154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,3072,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,6144,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,3072,0.1665279997719659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,1024,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,6144,0.02958933181232876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,51200,0.18947199980417886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,8192,0.015836444165971544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,51200,1.1606924268934462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,8192,0.057557331191168896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,2560,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,5120,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,2048,0.01273155543539259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,1024,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,768,0.035544001393847995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,8192,0.16070577833387586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,256,0.023949333363109167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,8192,0.0457893345091078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,16384,0.3577920065985785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,768,0.8597973187764486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,65536,12.204182094997831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,12288,0.13158844576941595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,4096,0.01612977849112617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,1024,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,4096,0.03683822353680929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,12288,0.017820444371965196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,1024,0.023609777291615803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,10240,0.03762844536039565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,5120,0.019319110446506076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,256,0.009576000273227692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,16384,0.044694221682018705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,6144,0.00997688869635264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,128,0.0031519999934567344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,768,0.020589333441522386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,3584,0.02715555495686001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,32,0.013737777868906656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,4096,0.011667555405033959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,256,0.0035048888789282907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,256,0.0038337777886125776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,512,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,4096,0.3915999995337592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,128,0.01646933290693495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,6144,0.014692443940374585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,51200,0.32000356250339085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,12288,0.013727111121018728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,768,0.019204444355434842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,3072,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,2560,0.04593066705597771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,51200,0.12589066558414036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,128,0.01756977703836229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,7168,0.022913777165942725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,3072,0.02152266601721446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,7168,0.021920000513394673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,5120,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,4096,0.01143022212717268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,10240,0.3040880097283257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,1536,0.006843555718660355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,16384,0.023279110590616863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,1024,0.024792888098292883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,6144,0.055585775110456676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,3072,0.026758222116364375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,51200,0.25732355647616917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,12288,0.030003554291195337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,8192,1.1268035040961373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,2560,0.015903110305468243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,768,0.040122667948404946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,256,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,256,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,2560,0.015372445185979208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,1024,0.027083555857340496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,768,0.03308799862861633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,3584,0.09894577662150066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,6144,0.01149866647190518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,12288,0.10455999771753947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,2560,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,8192,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,16384,0.37200533019171816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,1536,0.0063475556671619415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,64,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,4096,0.1915502283308241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,2560,0.03884799944029914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,65536,1.8162888420952692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,8192,0.017032000753614638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,512,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,1024,0.007686222592989604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,2560,0.027081777652104694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,8192,0.07109422153896756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,256,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,256,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,3584,0.03693422343995836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,51200,0.053547554545932345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,1536,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,128,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,256,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,51200,0.050267553991741605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,1536,0.5404471291436089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,512,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,512,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,12288,0.0542755557431115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,256,0.01956622302532196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,2560,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,7168,0.051092445850372314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,10240,0.23430310355292427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,3584,0.75510130988227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,768,0.030978666411505804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,65536,0.32205067740546334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,2560,0.008608889248636033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,12288,0.043755554490619235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,6144,0.015263110399246216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,3584,0.06292000081804064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,5120,0.06326933039559259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,128,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,3072,0.10071733262803818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,12288,0.3801235622829861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,51200,0.012326222327020435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,2048,0.03080000148879157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,512,0.011039111349317761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,3072,0.01790844400723775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,768,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,1536,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,51200,0.04688266581959195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,6144,0.034557332595189415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,128,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,16384,0.03312355611059401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,7168,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,51200,0.143248889181349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,1024,0.056112892097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,512,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,512,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,6144,0.028697778781255085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,32,0.0017662222186724346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,768,0.01814311080508762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,6144,0.009295110901196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,10240,0.06053600046369764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,65536,0.5613359875149196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,51200,0.04968533251020643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,256,0.002546666603949335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,512,0.00574933323595259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,6144,0.017934223016103108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,1024,0.018435556027624343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,3584,0.018239110708236694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,5120,0.02167644434505039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,10240,0.09180355734295315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,65536,0.0644328859117296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,3584,0.01681511104106903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,3584,0.1618115504582723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,2048,0.006511111226346757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,512,0.027835554546780054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,1536,0.006258666515350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,16384,0.5468622313605415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,2048,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,256,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,5120,0.008240000241332585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,10240,0.042747557163238525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,768,0.0326533317565918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,5120,0.2111253341039022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,128,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,64,0.01757422255145179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,128,0.004640000147951974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,512,0.021722666091389124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,16384,0.016154666741689045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,5120,0.03196888830926683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,4096,0.10302578078375922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,128,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,1024,0.016216889023780823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,2560,0.12580533822377524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,16384,0.4908160103691949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,3584,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,4096,0.03768622212939792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,32,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,256,0.002222222172551685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,8192,0.08500622378455268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,4096,0.02121777832508087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,2560,0.020089776979552377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,1024,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,51200,0.053730666637420654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,65536,0.1488542159398397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,64,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,768,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,65536,0.7978097597757975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,1536,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,512,0.06376533375846015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,16384,1.0439981884426541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,51200,0.09390399853388469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,7168,0.021817778547604878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,64,0.009397333694828881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,12288,0.03492444422509935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,8192,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,65536,0.06313421991136339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,2560,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,128,0.020672889219390023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,16384,0.4327840010325114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,4096,0.0766213337580363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,51200,0.1761315531200833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,12288,0.5482106738620335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,5120,0.0819644464386834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,768,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,16384,0.03719288773006863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,65536,0.07722577783796522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,65536,0.06802133056852552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,3072,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,16384,1.9363884396023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,512,0.022791999909612868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,1536,0.003548444559176763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,10240,0.047727998760011464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,5120,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,8192,0.06360799736446805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,1024,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,5120,0.02533600065443251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,1024,0.022877333892716303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,256,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,16384,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,7168,0.030581331915325586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,128,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,3584,0.043790221214294434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,16384,0.04524088899294535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,512,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,2560,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,1024,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,65536,0.874326229095459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,65536,0.05565422111087375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,768,0.04851644568973117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,10240,0.042528887589772545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,512,0.14981688393486872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,51200,0.08768000205357869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,128,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,7168,0.052705778015984424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,512,0.04435733291837904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,7168,0.019702222612169053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,768,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,3072,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,2560,0.005059555586841372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,2048,0.019124444988038804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,1024,0.010031110710567897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,128,0.03009777930047777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,1024,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,51200,0.06069688664542305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,1024,0.03140888942612542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,7168,0.011791999969217511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,3584,0.027452444036801655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,512,0.02162222233083513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,3584,0.01605244477589925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,12288,0.016089777151743572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,7168,0.010395555860466426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,8192,1.3460346857706706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,2560,0.03552000059021844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,8192,0.030239999294281006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,65536,0.16663200325436062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,512,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,16384,0.017988445030318368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,10240,0.11190666092766656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,768,0.02885777751604716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,64,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,3584,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,1536,0.006844444407357111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,256,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,1024,0.012225777738624148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,5120,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,256,0.01720177796151903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,32,0.012958221965365939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,2048,0.02368533280160692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,1024,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,16384,0.2880364523993598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,2560,0.41947823100619847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,768,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,512,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,6144,0.05197066730923123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,12288,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,3584,0.02658933401107788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,512,0.0028311111446883944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,3072,0.030243555704752605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,3072,0.0110880004035102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,51200,0.09529599878523086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,7168,0.07634933127297296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,7168,0.035071111387676664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,512,0.011730666789743634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,3584,0.026354667213228013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,256,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,7168,0.02394666605525547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,51200,0.06912000311745538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,65536,0.9894427193535699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,2048,0.02405333353413476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,128,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,1024,0.003563555578390757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,1024,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,16384,0.11003822088241577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,16384,0.49726491504245335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,512,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,64,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,1024,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,1024,0.032219555642869734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,64,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,1024,0.07421066363652547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,4096,0.017610667480362784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,256,0.04088355435265435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,2560,0.017701332767804463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,65536,3.074992073906792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,4096,0.048046221335728966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,64,0.03404444456100464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,4096,0.05751733647452461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,5120,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,8192,0.04593333270814684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,128,0.007641777396202087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,3584,0.04463466670778063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,12288,0.023970666858885024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,64,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,256,0.008446222378147973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,65536,0.04856444398562113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,2560,0.007327111230956183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,8192,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,16384,0.387609773212009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,2560,0.40881776809692383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,12288,0.057137780719333224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,8192,0.017312889297803242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,51200,0.24542400572035047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,7168,0.019907555646366544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,16384,0.01958577831586202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,2048,0.029429333077536687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,51200,2.2602861192491317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,10240,0.4326168960995144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,4096,0.009482666850090027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,2560,0.3849777645534939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,512,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,51200,1.0470924377441406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,128,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,16384,0.06911555263731214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,4096,0.04309422108862135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,5120,0.21170488993326822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,12288,0.0398613346947564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,8192,2.3794167836507163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,12288,0.03394577900568644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,51200,0.04855822192298042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,512,0.02951999836497837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,6144,0.567122671339247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,5120,0.029826667573716905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,10240,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,32,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,512,0.01109866632355584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,7168,6.967011345757379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,32,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,4096,0.9136151207817925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,65536,0.06725155644946627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,6144,0.08681333065032959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,1024,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,256,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,16384,0.05777244435416328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,3072,0.031352887550989784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,128,0.01663466625743442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,65536,0.06772622134950426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,16384,0.2975697782304552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,7168,0.2237128946516249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,16384,1.068603515625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,65536,0.030607110924190942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,4096,0.01828622155719333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,65536,2.128479215833876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,512,0.02163911031352149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,1024,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,2560,0.018576888574494254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,1024,0.011243555280897351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,128,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,6144,1.0127289030287001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,5120,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,1024,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,65536,0.06383911106321546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,512,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,3584,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,6144,0.055162668228149414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,1536,0.017273777061038546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,1024,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,64,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,65536,0.20380711555480957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,4096,0.09582933452394272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,768,0.011472889118724398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,1536,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,8192,0.03678755627738105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,51200,0.759400897555881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,1536,0.0588666664229499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,10240,0.16976443926493326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,65536,0.027706666125191584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,768,0.048080888059404164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,5120,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,12288,0.07916533284717135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,51200,13.393533494737413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,1536,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,65536,0.1399031082789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,12288,0.019314666589101154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,3072,0.2233964337242974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,1536,0.06986577643288507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,65536,1.2909893459743924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,3072,0.00922400007645289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,32,0.002858666703104973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,64,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,1024,0.1645804378721449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,256,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,2048,0.02554399934079912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,51200,1.4147449069552953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,1536,0.03940355446603563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,51200,0.05128088924619886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,65536,1.7239262262980144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,3072,0.020651555723614164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,10240,0.07956800195905897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,256,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,8192,0.017761778500345018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,1536,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,3584,0.09472088681326972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,65536,0.20455200142330596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,3072,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,768,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,16384,0.008216888540320927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,1536,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,6144,0.05309244328074985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,10240,1.0856142044067383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,6144,0.02089866664674547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,2048,0.011870221959220039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,768,0.01649688846535153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,4096,0.04825599988301595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,6144,0.015460444821251763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,256,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,4096,0.033189333147472806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,8192,0.020429333051045735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,65536,0.14686666594611272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,2048,0.027978665298885767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,2048,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,3072,0.008044444852405125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,8192,0.15352266364627415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,2048,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,1536,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,1024,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,16384,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,128,0.01204622205760744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,1536,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,65536,0.10244266854392158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,3584,0.01161688897344801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,5120,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,1024,0.02703911066055298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,8192,0.2257822354634603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,3584,0.2577928966946072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,4096,0.02214488883813222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,4096,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,65536,1.9443910386827257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,7168,0.12495999866061741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,256,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,3072,0.01425155500570933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,128,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,6144,0.0830471118291219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,512,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,7168,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,6144,0.014416888356208801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,256,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,1024,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,7168,0.04354666670163473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,6144,0.03795822130309211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,64,0.029764443635940552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,4096,0.02276888820860121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,51200,0.051051557064056396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,3584,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,65536,0.07660622066921659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,3584,0.024333332975705464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,65536,0.14555733733707005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,32,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,8192,0.035453332795037165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,512,0.0185208883550432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,4096,0.02477688921822442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,256,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,51200,0.06451911396450467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,3584,0.010111110905806223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,2048,0.022634666826989915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,512,0.0035235554807715942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,768,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,64,0.022544888986481562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,256,0.04434577624003092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,8192,0.01202933324707879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,5120,0.017697778013017442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,8192,0.03295466634962294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,65536,31.806515163845486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,1024,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,8192,0.1563804414537218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,64,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,3072,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,7168,1.5123920440673828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,8192,0.031630221340391375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,65536,0.3681520091162787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,64,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,2048,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,51200,0.5200008816189235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,16384,0.02374222212367588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,64,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,3072,7.575903998480903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,3072,0.015384000208642749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,2048,0.046048889557520546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,3072,0.0997048881318834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,2560,0.06427111228307088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,2560,0.011503111157152386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,5120,0.04768800073199802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,3072,0.0844808883137173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,2048,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,65536,1.2272390789455838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,51200,1.025992923312717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,64,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,16384,0.043150222963756986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,51200,0.05976088841756185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,2048,0.04833777745564779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,3072,0.11548533704545762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,1024,0.06395110819074842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,2048,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,8192,0.01740977830357022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,7168,0.2237013445960151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,12288,5.431907653808594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,8192,0.02351733379893833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,128,0.005714666512277391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,32,0.010049777726332346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,7168,0.006549333532651265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,2048,0.010302222437328762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,2048,0.03057866626315647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,128,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,128,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,256,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,1536,0.029948443174362183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,32,0.013164444102181328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,3584,0.03910488883654276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,1536,0.017131555411550734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,3584,0.01660711069901784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,4096,0.010644444160991244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,65536,0.06260977851019965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,512,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,256,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,51200,0.14067822032504612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,4096,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,512,0.008575110799736446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,768,0.03245333168241713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,2048,0.13919911119672987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,768,0.008847110801272923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,2048,0.02143200072977278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,64,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,51200,1.28965515560574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,2560,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,3584,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,5120,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,2048,0.022435555855433147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,512,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,2048,0.03490133417977227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,51200,0.2754097779591878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,256,0.014133334159851074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,32,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,1536,0.09901422262191772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,8192,0.11200177669525146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,12288,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,8192,0.04562755425771078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,10240,0.03087555699878269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,4096,0.015034667319721647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,5120,0.5961190859476725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,1024,0.03591466612286038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,3584,0.060805333985222705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,5120,1.516840934753418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,1024,0.011101333631409539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,1024,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,16384,0.34036978085835773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,128,0.0041777777175108595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,5120,0.13128177324930826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,256,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,1536,0.02016088863213857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,8192,0.06942400005128649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,51200,0.19663377602895102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,64,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,1536,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,10240,0.02639644510216183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,5120,0.012447111308574677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,2048,0.013194666968451606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,5120,0.0476800004641215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,1536,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,6144,0.022971555590629578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,16384,0.2695057657029894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,128,0.004330666528807746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,3584,0.054077333874172635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,4096,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,3584,0.007831110722488826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,3584,0.05959910816616482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,7168,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,5120,0.03180266751183404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,65536,0.06220799684524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,12288,0.21241066190931532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,768,0.02346399923165639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,5120,0.01657777859105004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,3584,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,5120,0.13760799831814235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,6144,0.020971554848882887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,256,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,12288,0.15156977706485325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,64,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,2048,0.045401778486039907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,65536,0.09140711360507542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,128,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,16384,0.8095048798455132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,5120,0.17092355092366537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,3072,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,10240,0.0608195530043708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,6144,0.12071555190616184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,12288,0.44024178716871476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,8192,0.0555902189678616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,51200,0.06264533599217732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,256,0.009523555636405945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,2560,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,10240,0.021720000439220007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,1536,0.020615110794703167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,3072,0.01984355515903897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,512,0.014286221729384528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,65536,0.025051555699772302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,6144,0.07261421945359972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,3072,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,128,0.059161775641971164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,16384,0.18795466423034668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,32,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,6144,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,512,0.02375288969940609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,5120,0.008449777960777283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,3072,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,768,0.018373333745532565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,10240,0.19976888762580025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,8192,1.8157617780897353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,2048,0.01475644442770216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,5120,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,768,0.03227377931276957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,51200,0.014295111099878946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,10240,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,1024,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,512,0.029399110211266413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,128,0.07627999782562256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,51200,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,6144,0.1699191066953871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,2048,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,256,0.009733333355850643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,2560,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,2560,0.02739199995994568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,128,0.05961866511238945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,256,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,1536,0.01977333260907067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,12288,0.03177511029773288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,128,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,512,0.02762577765517765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,8192,0.2597964339786106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,768,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,51200,0.059623108969794385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,2048,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,7168,0.0324924455748664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,2560,0.01686844395266639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,12288,0.02738044493728214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,4096,0.016539555456903245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,5120,3.847200181749132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,8192,0.12814311186472574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,2560,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,256,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,768,0.010296888649463654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,4096,0.02977777851952447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,16384,1.0533884896172416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,5120,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,2048,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,16384,0.5332720014784071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,32,0.003847111016511917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,10240,0.03011466728316413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,512,0.04720711045795017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,12288,0.022657778528001573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,2048,0.02134311033619775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,1024,0.04326488905482822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,128,0.0018968888454967076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,64,0.00794666674402025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,64,0.0020275554723209804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,3072,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,7168,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,16384,0.3695226775275336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,256,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,8192,0.081386665503184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,16384,0.016982222596804302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,5120,0.07864177889294095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,5120,0.08832089106241862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,3584,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,2048,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,256,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,512,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,2048,0.07113155391481188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,65536,0.23229334089491102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,5120,0.00903733323017756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,64,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,4096,0.008267555799749162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,512,0.013524444566832649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,7168,0.13216533925798205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,3072,0.11524444156222874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,10240,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,3072,0.08243199851777819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,51200,0.12719021903143987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,12288,0.04924355612860786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,12288,0.3441840012868245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,64,0.017080000705189176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,12288,0.018160000443458557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,256,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,64,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,768,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,6144,0.01944533321592543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,16384,0.14361955059899226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,3584,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,256,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,1024,0.017966222431924608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,2048,0.004278222305907143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,10240,0.016708445217874315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,128,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,256,0.014123555686738757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,6144,0.012351999680201212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,512,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,32,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,1536,0.003896000070704354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,10240,0.01960000064637926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,3584,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,8192,0.01442399952146742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,2560,0.007231999602582719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,1024,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,768,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,10240,0.026963555150561865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,5120,0.010031110710567897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,10240,0.03147733211517334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,12288,0.11639022827148438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,768,0.029947555727428857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,7168,0.018094221750895183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,512,0.002951111023624738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,768,0.033604443073272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,768,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,5120,0.009387555221716562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,16384,0.010928000013033548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,2048,0.017478222648302715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,51200,0.3649848832024469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,3584,0.07300266954633924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,1024,0.015744000673294067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,1536,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,256,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,5120,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,1024,0.02929511004024082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,512,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,5120,0.011529777612951068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,3072,0.04364888866742452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,5120,0.04740088846948412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,5120,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,5120,0.020245333512624104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,8192,0.07060088713963826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,51200,0.1435528861151801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,128,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,16384,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,1024,0.00332622230052948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,4096,0.08574399683210586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,3072,0.011712888876597086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,16384,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,8192,0.030528889762030706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,16384,0.06048711140950521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,12288,0.021537777450349595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,512,0.005527110977305307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,2560,0.027122666438420612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,1024,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,768,0.02772177755832672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,1024,0.005526222288608551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,64,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,2560,0.011412444214026133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,3072,0.027834667099846735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,768,0.018580444984965853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,7168,0.1671511067284478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,1024,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,2048,0.014596444037225513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,6144,0.022397332721286353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,256,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,10240,0.023747555083698694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,128,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,3072,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,512,0.009152889251708984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,16384,0.3177564409044054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,128,0.017845332622528076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,4096,0.02085866696304745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,4096,0.18783377276526558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,5120,0.12226311365763347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,768,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,128,0.008589333130253686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,768,0.10557866758770412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,256,0.026750221848487854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,1024,0.021820444199774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,1024,0.004286222159862518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,3584,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,7168,1.3066106372409396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,1024,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,1024,0.042027556233935885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,65536,2.60480711195204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,7168,0.10183822446399265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,7168,0.14441155062781438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,512,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,2560,0.20752444532182482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,16384,0.5116853184170193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,3584,0.023379556006855432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,51200,0.32567376560635036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,128,0.00674044465025266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,2048,0.006694222076071634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,1536,0.030657778183619182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,16384,0.04720800121625265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,1536,0.016470222009552848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,512,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,256,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,1024,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,512,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,768,0.011895111037625207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,256,0.022367111510700647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,1024,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,51200,0.12266311380598281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,512,0.016578666037983365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,768,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,3072,0.027670222851965163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,10240,0.14135555426279703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,128,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,768,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,256,0.09154755539364284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,4096,0.016851555969980028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,65536,0.06781244277954102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,65536,0.08671200275421143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,3072,0.006833777659469181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,65536,0.021317332983016968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,1024,0.06883111264970568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,3584,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,2560,0.02309866746266683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,1024,0.308297766579522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,51200,0.044079999128977455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,16384,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,16384,0.053420444329579674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,4096,0.008052444292439355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,3584,0.10739644368489583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,10240,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,6144,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,2560,0.12457866138882107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,4096,0.08173155784606934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,3584,0.4147164556715224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,7168,0.14151377148098415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,16384,0.018458666072951425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,12288,0.027140445179409448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,8192,0.04979999860127767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,16384,0.06087199846903483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,2560,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,3584,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,1024,0.06981422503789265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,512,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,2560,0.18050132857428658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,1024,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,128,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,2560,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,1024,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,3584,0.0177831103404363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,6144,1.7502791086832683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,256,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,256,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,5120,0.06315377685758802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,65536,127.66022406684027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,65536,2.009444342719184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,1024,0.025455999705526564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,32,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,3072,0.49649688932630753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,128,0.05201866560512119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,8192,0.07140711281034681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,5120,0.014126222994592456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,4096,0.022987556126382615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,256,0.01720622181892395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,64,0.04303377866744995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,8192,1.6585715611775715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,256,0.02167466613981459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,7168,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,512,0.002712888850106133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,16384,0.19371733400556776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,12288,0.18145599630143908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,8192,0.19731556044684517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,16384,0.04866310954093933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,6144,0.10477866729100545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,16384,0.04088888896836175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,1024,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,10240,0.10502844386630589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,51200,0.1466239955690172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,768,0.008985777695973715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,12288,0.15295555856492785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,1024,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,5120,0.10528444581561619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,5120,0.02201955517133077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,256,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,64,0.010608889162540436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,512,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,2048,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,8192,0.06938933663898043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,51200,0.10391200251049465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,512,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,10240,0.07578844494289823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,256,0.01645599967903561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,51200,0.11055288712183635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,65536,0.13856089115142822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,3072,0.007927999728255803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,8192,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,16384,0.028267555766635474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,1536,0.031081779135598078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,7168,0.027481777800454035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,7168,0.024023110667864483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,12288,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,768,0.29457955890231663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,3072,0.007394666473070781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,512,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,32,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,2048,0.01586933268441094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,3072,0.022617777188618977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,6144,0.018002667360835604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3584,12288,0.028127110666698877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,51200,0.16990666919284395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,10240,0.20292354954613578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,3584,0.07849244276682536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,5120,0.02274311085542043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,65536,0.09453155597050984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,6144,0.009721777505344814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,16384,0.1515102254019843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,1024,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,16384,0.008795555267069075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,128,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,4096,0.040849778387281634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,1024,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,768,0.027090667022599116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,2560,0.027079999446868896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,512,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,512,0.007130666739410824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,12288,0.13466755549112955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,4096,0.03306311037805345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,7168,0.30085600747002494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,16384,0.08618399831983779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,65536,0.678416887919108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,1536,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,128,0.004519111166397731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,1024,0.015545777148670621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,51200,0.5218142403496636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,7168,0.21006400055355498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,3584,0.050944000482559204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,4096,0.11099466350343491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,51200,0.40319111612108016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,128,0.0038942222793896994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,7168,0.010496888723638324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,51200,0.1040302250120375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,65536,0.4277448919084337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,1024,0.015722667177518208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,65536,0.11753688918219672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,6144,0.08374310864342584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,7168,0.012914666699038612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,16384,0.16291377279493544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,1536,0.026271111435360376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,1536,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,768,0.035691556003358625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,10240,1.117120001051161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,2560,0.08663555648591782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,7168,0.09567022323608398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,3584,0.029422223567962646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,3072,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,5120,0.022107554806603327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,4096,0.07071022192637126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,512,0.03568444318241543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,51200,0.07388710975646973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,10240,0.015432000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,16384,0.046762665112813316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,768,0.002801777794957161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,2048,0.01111644423670239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,65536,0.41460445192125106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,3072,0.10075644652048747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,768,0.016000888413853116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,64,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,768,0.118651549021403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,65536,0.0862364437845018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,1024,0.006142222219043308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,12288,0.0704231129752265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,16384,0.28934044308132595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,65536,2.0576604207356772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,10240,0.018138666947682697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,16384,0.04887466629346212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,128,0.025736000802781846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,2048,0.39142221874660915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,12288,0.06161866585413615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,16384,0.09323733382754856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,6144,0.17371643914116752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,16384,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,7168,0.019550222489568923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,512,0.055995557043287486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,12288,0.14710933632320827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,65536,0.5743093490600586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,16384,0.020483555065260995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,1024,0.024069334069887798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,768,0.5682497554355198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,256,0.007682667010360294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,2560,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,12288,0.051161779297722705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,4096,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,6144,0.026504889130592346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,65536,0.062485330634646945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,1536,0.013996443814701505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,2560,0.4790364371405707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,51200,0.4500551223754883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,5120,0.08296710915035672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,7168,0.05163733495606316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,128,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,4096,0.012671111358536614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,51200,0.2918764485253228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,768,0.07240088780721028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,4096,0.11260533332824707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,256,0.030532442861133154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,256,0.0038008888562520346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,10240,0.0207013338804245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,3584,0.029116445117526587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,10240,0.07266755236519708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,16384,0.3391457663642035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,12288,0.031417777140935264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,10240,0.976594712999132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,5120,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,32,0.010824888944625854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,5120,0.8123920228746203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,3072,0.0074471111098925276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,3072,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,2048,0.012499555945396423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,256,0.06241866615083483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,6144,0.01645155582163069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,65536,1.2332258224487305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,1536,0.073216888639662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,1024,0.019938665959570143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,8192,0.0484773317972819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,51200,0.08522933059268528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,1536,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,3072,0.06257688999176025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,3584,0.015442666080262927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,1536,0.06042489078309801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,3584,0.40378310945298934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,2048,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,768,0.010326222413116032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,1024,0.032784889141718544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,4096,0.02917333443959554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,1024,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,5120,0.004576888763242298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,256,0.023312888211674158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,12288,0.017140444782045152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,65536,0.10434310966067845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,1024,0.030694223112530176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,2048,0.006558222075303395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,10240,0.012903110848532783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,51200,0.0645315580897861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,5120,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,16384,0.018984888990720112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,3584,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,8192,0.14226222038269043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,1536,0.023006222314304773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,1536,0.006004444426960415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,64,0.0017973333597183228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,1536,0.0314684444003635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,3072,0.05726044707828098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,12288,0.11996355321672227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,5120,0.03227555420663621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,5120,0.025588444537586633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,32,0.0021235555824306277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,512,0.01256622208489312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,65536,0.20304889149136016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,65536,1.4176968468560114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,12288,0.020610666937298246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,65536,0.2776346736484104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,512,0.02223288847340478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,256,0.0022479999396536085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,1536,0.030256887276967365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,6144,0.04704888992839389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,16384,0.019662222928471036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,2560,0.010977778169843884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,256,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,5120,0.05714933077494303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,512,0.01720622181892395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,512,0.06487733125686646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,32,0.0016888889173666637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,51200,0.25871822569105357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,256,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,32,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,1536,0.051057779126697116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,8192,0.021662222014533147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,3584,0.013738666971524557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,1536,0.0063493334584765965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,10240,0.08413866493437026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,12288,0.01756000022093455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,32,0.002485333424475458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,256,0.019149333238601685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,3584,0.017228444417317707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,5120,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,1536,0.08535110950469971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,1536,0.011001778145631155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,3072,0.05125955409473843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,128,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,4096,0.06412355767356025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,6144,1.183920012580024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,6144,0.07148444652557373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,2560,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,65536,0.07090044683880277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,768,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,64,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,51200,0.014292443792025248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,2560,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,16384,0.04140355520778232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,6144,0.02626488937271966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,65536,0.39000799920823837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,5120,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,6144,0.030571556753582422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,10240,0.018049778209792245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,1536,0.02887822190920512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,64,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,3072,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,3072,0.045555555158191256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,5120,0.021406221720907424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,3584,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,10240,0.034613334470325045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,64,0.004201777693298128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,2048,0.009704888694816166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,65536,0.7206160227457682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,5120,0.030998223357730444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,7168,0.021912000245518152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,32,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,51200,1.2469039493136935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,768,0.021835555632909138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,65536,0.17088533772362602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,256,0.034873776965671115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,2560,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,7168,0.019853333632151287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,1024,0.03163466519779629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,8192,0.04279466801219516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,32,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,16384,0.0359280010064443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,4096,0.3509093390570746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,256,0.010648888846238455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,51200,0.05319999986224704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,6144,0.060623996787601046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,1536,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,51200,0.13411200046539307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,2560,0.028778665595584448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,4096,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,2048,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,3072,0.0099653336736891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,65536,0.41409601105584043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,512,0.01791733337773217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,5120,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,4096,0.05243822269969516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,256,0.0038168889780839286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,3072,0.1265875498453776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,6144,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,6144,0.1411608854929606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,1536,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,65536,0.11899556054009332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,64,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,1024,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,5120,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,6144,0.09777422083748712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,2048,0.004951999833186467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,6144,0.004941333499219683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,768,0.030777778890397813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,2048,0.007633777956167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,512,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,1536,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,768,0.005519111123349931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,12288,0.15076888932122126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,256,0.024658666716681585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,16384,0.019697777099079557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,10240,0.02845422095722622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,10240,0.2516293260786268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,2560,0.09796533319685195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,512,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,51200,64,0.046385778321160205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,1024,0.03978755407863193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,1536,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,128,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,6144,0.018996444013383653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,1024,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,512,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,7168,0.0310515562693278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,51200,0.4475173420376248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,51200,0.05778133206897312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,64,0.0018275556051068837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,4096,0.028745777077145044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,512,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,12288,0.026240888569090102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,2048,0.09250489208433364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,2048,0.01808088852299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,128,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,51200,0.4355795648362901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,10240,0.052489777406056724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,7168,0.024495111571417913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,4096,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,65536,0.0621617767545912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,51200,0.06127555502785576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,5120,0.041272001134024724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,64,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,512,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,64,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,2048,0.018552000323931377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,128,0.016913778252071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,64,0.010443555812040964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,1536,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,128,0.01477599971824222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,128,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,65536,1.6799661848280165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,256,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,512,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,1024,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,2560,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,128,0.01087288889620039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,6144,0.204605327712165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,512,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,768,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,1024,0.03221599923239814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,128,0.7790737681918674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,8192,0.2081537776523166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,32,0.0019191110299693213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,128,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,4096,0.04382133483886719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,10240,0.18818399641248915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,256,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,7168,1.4686951107449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,3072,0.04079733292261759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,256,0.02860355708334181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,2560,0.02163822286658817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,128,0.022083555658658344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,4096,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,768,0.0074924445814556545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,1536,0.024572443630960252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,16384,0.03700177868207296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,128,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,16384,0.11916889084710015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,2560,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,65536,0.05958133273654514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,16384,0.08111555708779229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,12288,0.01754044493039449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,10240,0.024103111690945093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,16384,0.1400417751736111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,512,0.12785066498650446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,512,0.008347555167145198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,2048,0.0042257776690853965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,51200,0.07538488838407728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,512,0.011770666473441653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,2560,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,4096,0.02278222143650055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,2048,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,10240,0.06955555412504408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,128,0.014064000712500678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,10240,0.25578400823805064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,32,0.00254488881263468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,6144,0.03219199842876858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,65536,0.1268355581495497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,6144,0.023177777727444965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,2560,0.02397955622937944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,1536,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,65536,0.06264266702863905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,51200,0.48835288153754336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,10240,0.03756622142261929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,51200,17.244808620876736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,12288,0.017749332719379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,2048,0.008094222181373173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,4096,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,2560,0.020282665888468426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,128,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,768,0.01109955542617374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,7168,0.02757511039574941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,1024,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,3584,0.01791200041770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,12288,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,10240,0.020938666330443487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,6144,0.031256887647840716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,128,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,65536,0.1293706629011366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,512,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,5120,0.021193777521451313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,4096,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,512,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,1024,0.023409777217441138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,51200,0.07042577531602648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,7168,0.018783110711309645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,4096,0.05412888858053419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,3072,0.01692800058258904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,10240,0.026360000173250835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,65536,0.18881689177619088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,65536,1.7943191528320312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,16384,0.05471733212471008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,51200,0.22119288974338105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,6144,0.018233777748213876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,64,0.004407111141416761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,32,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,7168,0.01819644371668498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,128,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,768,0.0028853333658642242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,5120,0.08715911044014825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,512,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,2560,0.2109840048684014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,10240,0.6590933269924587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,12288,0.09619733360078599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,16384,0.02232355541653103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,4096,0.022667555345429316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,3584,0.014819555812411837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,32,0.0017848888205157386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,32,0.062259554862976074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,4096,0.010413332945770687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,65536,0.0247715562582016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,2048,0.0041751112374994485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,51200,0.37198400497436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,7168,1.9763253529866536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,65536,2.8885813819037547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,768,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,8192,0.02181600034236908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,64,0.0259253333012263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,5120,0.48167732026841903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,7168,0.0630355543560452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,2560,0.013219555219014486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,32,0.012498666842778524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,3584,0.028505778974956934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,65536,0.017954667409261067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,16384,0.021564443906148274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,12288,0.04042133357789781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,10240,0.17400533623165557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,32,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,16384,0.016306666864289176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,256,1.1745484670003254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,8192,0.03013777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,4096,0.04958666695488823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,1536,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,3072,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,16384,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,5120,0.03874489002757602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,7168,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,3072,0.012247111234400006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,7168,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,256,0.011034666664070554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,12288,0.2808319992489285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,3072,0.0684551133049859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,4096,0.4389964474572076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,16384,0.33189688788519967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,64,0.009031111167536842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,2560,0.07687466674380831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,1536,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,16384,0.020966221888860066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,51200,0.09562133418189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,6144,0.00906044410334693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,65536,0.2728817727830675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,768,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,12288,0.018723555737071566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,5120,0.05899822049670749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,5120,0.008739555875460306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,768,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,5120,0.06549955738915338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,7168,0.08179110950893827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,128,0.016321778297424316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,12288,0.034430222378836736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,7168,0.17833333545260957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,65536,0.20378043916490343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,65536,0.14366222752465144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,3072,0.031522668070263334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,1024,0.02309866746266683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,3072,0.011508444117175208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,12288,0.1890782250298394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,32,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,128,0.009539555344316695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,1024,0.012450666891203986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,3584,0.04315733247333103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,256,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,768,0.01828799976242913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,1024,0.19227022594875762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,4096,0.0716080003314548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,512,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,256,0.015160888433456421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,3584,0.11372444364759658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,10240,0.026008889079093933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,1024,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,6144,0.02925866511132982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,1536,0.018560889694425795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,32,0.012679110798570844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,256,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,12288,0.013556444810496436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,51200,0.06470133198632134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,3072,0.36509066157870823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,65536,0.1537946727540758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,128,0.016516443755891588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,7168,0.030680888228946265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,768,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,16384,0.028575999869240656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,4096,0.637696901957194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,64,0.013047999805874295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,128,0.005478222337034013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,1024,0.05884266561932034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,2048,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,768,0.014510222607188754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,12288,0.07589422331915961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,3072,0.4603297975328233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,6144,0.051161779297722705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,5120,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,2560,0.012269333004951477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,65536,0.1479146613015069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,768,0.031368888086742826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,3584,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,3072,0.1725697782304552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,768,0.008915555146005418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,12288,0.022408889399634466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,768,0.035712000396516584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,10240,0.03582311007711623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,512,0.022289777795473736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,16384,0.022787556052207947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,32,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,10240,0.09897777769300674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,10240,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,51200,0.33564265569051105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,768,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,5120,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,10240,0.020628444022602506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,8192,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,2048,0.017610667480362784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,32,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,768,0.0036559998989105225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,256,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,256,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,6144,0.6740621990627713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,5120,0.09340800179375543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,12288,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,3072,0.012336889074908363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,5120,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,1024,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,32,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,768,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,6144,0.12419377432929145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,51200,0.7576648924085828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,51200,12.235454983181425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,10240,0.03156444430351257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,51200,0.06884266932805379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,3584,0.026904889278941687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,128,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,1024,0.00776444458299213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,7168,0.1325333383348253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,1024,0.006125333408514659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,2048,0.006770666688680649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,10240,0.012481778032249875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,2048,0.01494311127397749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,6144,0.01664444473054674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,12288,0.023189332750108507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,1536,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,128,0.001965333397189776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,8192,0.011408888631396823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,3584,0.029712888929578993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,12288,0.44628535376654727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,4096,0.04072533382309808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,5120,0.0554017788834042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,3584,0.020187555087937247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,3072,0.047747555706236094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,256,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,1536,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,65536,0.07561155822541979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,1536,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,65536,0.5015608999464247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,12288,0.011402666568756104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,16384,0.3585982322692871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,3584,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,2048,0.009008000294367472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,128,0.01499466598033905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,5120,0.031882666879230075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,2560,0.00759733302725686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,7168,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,51200,1.2030835681491427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,51200,0.2583555645412869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,10240,0.051979555024041064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,512,0.01753866672515869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,512,0.0042168887125121225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,10240,0.018021333548757765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,2048,0.018295110927687753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,12288,0.021368000242445204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,256,0.022999111149046157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,128,0.01755911111831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,10240,0.02167822255028619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,1024,0.03085600005255805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,768,0.029329776763916016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,5120,0.03122133347723219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,256,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,5120,0.03165510959095425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,16384,0.12772622373369005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,256,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,8192,0.12726488378312853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,4096,0.08429066340128581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,10240,0.04456088940302531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,51200,0.0573600000805325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,65536,1.6870479583740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,1536,0.0041831110914548235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,5120,0.02306933369901445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,12288,0.0380311111609141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,1024,0.012335999972290464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,3584,0.015407999356587728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,1024,0.13944889439476862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,16384,0.2978959878285726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,4096,0.05281244383917915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,3584,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,256,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,768,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,1536,0.024512888656722173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,1536,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,65536,0.16922667291429308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,16384,0.17479022343953451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,2560,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,64,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,4096,0.021342222889264423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,1536,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,2048,0.02776088813940684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,51200,0.1440026627646552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,3584,0.05091822147369385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,3072,0.044024000565210976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,12288,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,10240,0.7172595659891764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,12288,0.07458311319351196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,256,0.00544888898730278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,5120,0.01039644413524204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,5120,0.08970844745635986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,5120,0.4710382355584039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,2048,0.01124533348613315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,51200,0.042479998535580106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,3072,0.008046222229798635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,1536,0.10706844594743516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,12288,0.01089511149459415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,1536,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,65536,0.3617546558380127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,768,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,16384,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,12288,0.01698844465944502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,2560,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3072,1024,0.02847111225128174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,768,0.015853333804342482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,7168,0.04121333360671997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,1024,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,2560,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,3584,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,1536,0.0413680010371738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,8192,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,256,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,51200,0.056892441378699414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,12288,0.01885333326127794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,3584,0.0656026668018765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,16384,0.26718934377034503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,768,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,128,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,3072,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,10240,0.01443288889196184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,8192,0.14512444867028132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,65536,0.15379288461473253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,7168,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,16384,0.018519111805491976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,2048,0.04068444503678216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,512,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,768,0.07566489113701715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,768,0.02149244480662876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,1536,0.2442257669236925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,2048,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,64,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,512,0.07467466592788696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,3584,0.04208266735076904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,8192,4.897186703152127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,512,0.25378666983710396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,12288,0.045722666713926524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,768,0.030234667989942763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,768,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,1024,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,2560,0.02680355476008521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,12288,0.029192888074451025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,8192,0.01198933356338077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,2048,0.022661333282788593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,2560,0.02609066665172577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,51200,0.1787795490688748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,3584,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,7168,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,768,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,128,0.02151022189193302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,3072,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,1536,0.03243644369973077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,51200,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,256,0.015564444992277356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,128,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,1536,0.008376888930797577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,12288,0.01742488808102078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,16384,0.26148443751864964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,64,0.008031110796663497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,2560,0.007560888926188151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,1024,0.008247999681366814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,8192,0.01637066735161675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,512,0.025519111090236243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,65536,0.1416071123547024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,256,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,5120,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,65536,0.06225066714816623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,64,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,768,0.02292622294690874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,10240,0.2470160060458713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,10240,0.012777778009573618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,16384,0.024309333827760484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,2048,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,12288,0.11958400408426921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,51200,1.2470151053534615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,5120,0.12498754925198025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,12288,0.15271377563476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,32,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,51200,0.850470225016276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,2560,0.04584177666240268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,256,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,10240,0.029701332251230877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,512,0.014690667390823364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,16384,0.016480889585283067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,8192,0.03109244505564372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,256,0.01685422162214915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,6144,0.012513777448071374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,1536,0.01737244427204132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,4096,0.028167112006081477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,6144,0.010695111420419482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,1536,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,7168,0.09500710831748115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,3072,0.011984888878133563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,12288,0.01772533357143402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,3584,0.24133777618408203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,1536,0.016846223009957206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,256,0.0023626666516065598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,256,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,65536,0.16752444373236763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,16384,0.13313510682847765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,64,0.016622222132152982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,10240,0.043895999590555825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,1536,0.00407911092042923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,1536,0.013140444126394061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,512,0.023992000354660883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,6144,1.5248924891153972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,512,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,8192,0.046006222565968834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,3072,0.016224000189039443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,3072,0.04716089036729601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,6144,0.03496711121665107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,51200,0.09712711307737563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,65536,1.546733326382107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,7168,0.040341334210501775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,256,0.005534222142563925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,3584,0.08551555871963501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,64,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,256,0.0148062225845125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,16384,0.03621777892112732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,10240,0.2905297809176975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,6144,0.050649778710471265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,1024,0.016562667157914903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,51200,0.10388533274332683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,256,0.009870222045315636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,12288,0.018396443790859647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,512,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,4096,0.017197334104114108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,6144,0.01088711122671763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,64,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,1536,0.004226666771703296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,12288,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,8192,0.15414310826195612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,6144,0.07407555315229628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,16384,6.420305040147569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,12288,0.022823999325434368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,16384,0.13785955641004774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,2048,0.027501333090994094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,512,0.013495999905798169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,12288,0.040116445885764226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,1536,0.014083556003040738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,32,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,8192,0.041048887703153826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,51200,0.03661511010593838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,1536,0.007087111473083496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,2048,0.061567107836405434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,6144,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,64,0.008955555657545725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,3584,0.013049778011110095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,6144,0.02649066680007511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,3584,0.02365600069363912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,1536,0.0063564446237352155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,128,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,3072,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,512,0.0721342232492235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,3072,0.024488000406159297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,12288,0.02973066767056783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,1024,0.025459556115998164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,1536,0.011767999993430244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,7168,0.12248978349897598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,4096,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,3072,0.15573067135281032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,768,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,5120,0.0052666668262746595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,6144,0.15680978033277723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,2560,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,6144,0.05420533484882779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,12288,0.04946844445334541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,8192,0.011863999896579318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,3072,0.01553066737122006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,12288,0.20732800165812174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,12288,0.5293075773451064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,2560,0.031325333648257785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,7168,0.026144888665941026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,1024,0.025608888930744592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,65536,0.07603733407126533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,2048,0.023370666636361018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,5120,0.7588604291280111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,65536,0.11655644575754802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,10240,0.07959200276268853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,10240,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,12288,0.03958844476275974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,3072,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,3584,0.018246221873495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,64,0.01625422139962514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,2560,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,7168,0.02775466607676612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,10240,0.034640888373057045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,6144,0.01945244438118405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,5120,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,5120,0.031303998496797346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,12288,0.1024435559908549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,32,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,10240,0.022879110442267522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,65536,4.787834591335721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,16384,8192,0.6680897606743706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,6144,0.2262044482760959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,4096,0.011459555890825061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,2560,0.035176889763938055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,512,0.0038622220357259116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,2048,0.016901334126790363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,16384,0.042103111743927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,5120,0.024002666274706524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,65536,0.07190577851401435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,256,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,6144,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,5120,0.0875706672668457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,768,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,2560,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,7168,0.06235644552442762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,10240,0.08715644147660996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,2048,0.04328622089491951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,768,0.006423111177153057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,16384,0.02068444424205356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,256,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,12288,0.05018933282958137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,4096,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,7168,0.013516444298956128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,5120,0.4469662242465549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,7168,0.059337774912516274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,2048,0.007175111108356052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,10240,0.0836906698015001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,10240,0.8217288653055826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,128,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,512,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,3584,0.01774577796459198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,4096,0.04365333252482944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,6144,0.20679822232988146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,2560,0.04862133330769009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,2560,0.025283555189768474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,2560,0.010099555883142684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,5120,0.008777778181764815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,10240,0.05199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,5120,0.03156000044610765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,2048,0.014754666222466363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,8192,0.016368889146380954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,1536,0.07974311378267077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,51200,0.12971111138661703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,512,0.02626222206486596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,3584,0.025115556187099878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,7168,0.04051999913321601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,5120,0.018182223041852314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,16384,0.0440008905198839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,128,0.005342222336265776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,12288,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,8192,0.023003555006451074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,2048,0.020344000723626878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,5120,0.019890666007995605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,10240,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,256,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,6144,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,1024,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,65536,0.1052888896730211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,8192,0.052060445149739586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,10240,0.034796443250444196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,7168,0.08256266514460246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,8192,0.039839999543295965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,12288,0.029968000120586816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,2560,0.007381333245171442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,4096,0.014357333381970724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,2560,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,3584,0.023631999890009563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,2048,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,768,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,2560,0.03389599919319153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,5120,0.01961511042382982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,51200,0.34200178252326113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,256,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,1024,0.3684977690378825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,3584,0.015758223003811307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,16384,0.03681777914365133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,768,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,7168,0.013665777941544851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,512,0.003907555507289039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,128,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,64,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,51200,0.46277687284681535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,8192,0.22168889310624865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,3584,0.013789333403110504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,12288,0.3430462148454454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,32,0.008223999705579545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,256,0.014642667439248828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,2560,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,128,0.03129955463939243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,256,0.014693333042992486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,51200,0.5221635500590006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,65536,0.39471644825405544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,2560,0.06249244345559014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,7168,0.016758221719000075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,768,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,256,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,2048,0.009552888572216034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,2048,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,5120,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,1536,0.009369778136412302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,8192,0.10689155260721843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,1536,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,6144,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,5120,0.05098399851057264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,7168,0.011010666688283285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,4096,0.016231111354298063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,768,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,512,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,51200,1.0829155180189345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,3072,0.13205333550771078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,1536,0.010679110884666443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,7168,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,4096,0.030272000365787085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,65536,0.12286133236355251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,10240,0.021876444419225056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,16384,0.7099128829108344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,512,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,5120,0.04840977655516731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,16384,0.022610666023360357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,7168,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,16384,0.5262053277757432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,65536,0.4403519895341661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,2560,0.015775110986497667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,32,0.0026640000028742682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,6144,0.019066666563351948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,3584,0.019409777389632333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,512,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,12288,0.04247822364171346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,12288,0.014470222923490735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,64,0.011335111326641507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,7168,0.14562043878767225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,64,0.008464888566070134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,12288,0.017424000634087455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,12288,0.0051857779423395796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,8192,0.05343822307056851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,8192,0.039360887474483915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,128,0.0019466665883858998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,16384,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,51200,0.284479988945855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,8192,0.03602133194605509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,7168,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,512,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,1024,0.01865244408448537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,6144,0.009761778016885122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,10240,0.07721510860655043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,5120,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,51200,0.5852337943183051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,1536,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,16384,0.195469339688619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,7168,0.02392888896995121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,3584,0.06266222397486369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,51200,0.4461706479390462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,2048,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,12288,0.8264488644070096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,10240,2.0201564364963107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,51200,0.9280586242675781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,4096,0.010260444548394945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,12288,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,5120,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,16384,0.02658844490845998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,128,0.009875555833180746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,2048,0.01882577770286136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,7168,0.02252888845072852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,10240,0.59406402375963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,1024,0.01237066669596566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,3072,0.00489066665371259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,768,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,1536,0.003546666767862108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,8192,0.015326221783955893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,5120,0.030958222018347845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,64,0.020618667205174763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,16384,0.025378665990299646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,3584,0.06580977969699435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,7168,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,3072,0.02819377846188015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,1024,0.014882667197121514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,768,0.01793155570824941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,64,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,7168,0.14317333698272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,6144,0.09469688600964016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,5120,0.03024622135692173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,32,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,3072,0.007873777714040544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,51200,8.362715827094183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,32,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,7168,0.25480445226033527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,1024,0.04119199845525954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,512,0.019135110908084445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,32,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,128,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,12288,6.200495825873481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,10240,0.32008621427747935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,4096,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,6144,1.4818444781833227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,1536,0.05105066630575392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,51200,0.06270400020811293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,5120,0.03398311138153076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,4096,0.01904711127281189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,2560,0.03949688871701559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,3072,0.06745244397057427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,1024,0.0041680000722408295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,128,0.010562666588359408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,256,0.5607573191324869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,4096,0.014120888378885059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,1024,0.028344889481862385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,2560,0.014827556080288358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,1536,0.03103999959097968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,768,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,64,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,3072,0.01350311107105679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,5120,0.19101156128777397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,3072,0.0899439983897739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,51200,0.08408888843324448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,65536,0.06379555331336127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,512,0.0026622222115596137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,2048,0.06855911016464233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,3072,0.016680889659457736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,2048,0.014280888769361707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,128,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,4096,0.023061333431137934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,3072,0.02166755497455597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,8192,0.01999288962946998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,12288,1.458456039428711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,16384,0.2559351126352946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,256,0.005671111245950063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,8192,0.028655999236636694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,6144,0.05585955248938667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,512,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,4096,0.0567528870370653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,3072,0.14089066452450222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,256,0.015543111496501498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,51200,0.049963557057910495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,3072,0.05123200019200643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,7168,0.02334755493534936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,51200,0.828071117401123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,10240,0.011895111037625207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,10240,0.01927199959754944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,1024,0.01442577772670322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,7168,0.014476444986131458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,1536,0.01921422282854716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,1536,0.33843644460042316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,1536,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,65536,0.022343110707071092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,7168,0.12268088923560248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,128,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,256,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,2560,0.02125866711139679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,8192,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,51200,0.3853253258599176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,768,0.07123644484413995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,5120,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,65536,0.06982488764656915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,12288,0.08599733644061619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,10240,0.07288177808125813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,128,0.0023866666273938283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,3072,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,8192,0.025962667332755193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,16384,0.05945777893066406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,3584,0.020747555626763236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,1536,0.06167288621266683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,7168,0.03333955672052171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,3584,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,128,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,3584,0.03059911065631443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,65536,0.38069600529140896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,3072,0.01721155477894677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,256,0.022890667120615642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,8192,0.6163102255927192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3584,4096,0.08781688743167454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,256,0.015233778291278414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,16384,0.02628444466325972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,128,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,128,0.002620444529586368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,256,0.007903111477692923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,256,0.011352889239788055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,12288,0.5001449055141872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,12288,0.021423111359278362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,64,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,7168,0.24202577273050943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,128,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,1024,0.003564444267087512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,512,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,6144,0.10706133312649196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,16384,0.06964800092909071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,1536,0.33333688312106663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,6144,0.03269155489073859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,12288,0.34248712327745223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,1536,0.019517333971129525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,10240,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,2048,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,64,0.06132266918818156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,768,0.016217778126398723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,1024,0.010436444646782346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,8192,0.05222310953670078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,12288,0.07904089159435697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,7168,0.13230844338734946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,10240,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,768,0.012614222036467658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,5120,0.08186755577723186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,5120,0.09518577655156453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,256,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,51200,0.06548888815773858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,16384,0.22937867376539442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,3584,0.08590755860010783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,64,0.034210665358437434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,65536,0.11673777633243138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,768,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,5120,0.10711555348502265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,512,0.011423110961914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,51200,0.05046933227115207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,1536,0.014322666658295525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,51200,0.13561511039733887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,65536,0.1873555580774943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,768,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,256,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,2560,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,3584,0.018560889694425795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,128,0.014787554740905762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,2048,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,128,0.009920888476901585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,65536,0.22893688413831922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,32,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,5120,0.03439111179775662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,8192,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,10240,0.04021155503061082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,16384,0.004937777916590373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,16384,0.020488889680968392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,5120,0.03559199968973795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,128,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,1024,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,65536,0.06212266948488024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,128,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,4096,0.016192888220151264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,2048,0.012107555237081317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,10240,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,8192,0.005689777847793367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,10240,0.1937262217203776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,10240,0.2834675576951769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,512,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,2048,0.2715840074751112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,32,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,16384,2.5034302605523004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,8192,0.03047910994953579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,5120,0.11147822274102105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,3072,0.13203288449181452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,8192,1.913488811916775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,5120,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,768,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,768,0.014095111025704278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,7168,0.24856355455186632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,2048,0.031049778064092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,2048,0.013941332697868347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,5120,0.021291555629836187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,65536,0.193648894627889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,5120,0.012695999609099494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,65536,0.17367288801405165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,3072,0.016156444946924847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,8192,0.12820622656080458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,256,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,6144,0.020581333173645865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,2048,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,128,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,7168,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,3584,0.008600888980759515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,2048,0.006364444477690591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,256,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,3584,0.010786666638321347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,512,0.021423111359278362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,51200,0.07907822397020128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,32,0.0018773333479960759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,2560,0.024816888901922438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,51200,0.05075733198059929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,51200,0.9222186406453451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,5120,0.02216000027126736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,2048,0.012500444220172035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,1536,0.04120355513360765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,32,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,5120,0.07704355319341023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,512,0.0058453331391016645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,8192,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,16384,0.02925866511132982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,3584,0.009319110876984065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,1024,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,256,0.017143110434214275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,6144,0.08872799740897284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,5120,0.008604444563388824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,2048,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,512,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,10240,0.022431999444961548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,768,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,2560,0.01180622229973475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,2560,0.013972444666756524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,1536,0.0041786668201287585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,64,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,1024,0.013167111410035027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,12288,0.029177778297000464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,16384,0.03660088777542114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,16384,0.12386755148569743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,512,0.012100444071822695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,7168,2.202379650539822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,3584,0.07437066899405585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,12288,0.04014133413632711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,128,0.02676711148685879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,768,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,256,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,512,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,128,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,64,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,4096,1.5652159584893122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,1024,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,7168,0.3005422221289741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,128,0.6409058041042751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,10240,0.20277155770195854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,51200,0.4180106586880154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,65536,8.933135986328125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,10240,0.02029244436158074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,51200,1536,0.05992444356282552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,1024,0.025075554847717285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,2560,0.06844355662663777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,64,0.01536177761024899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,128,0.01551644504070282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,1024,0.014174222946166992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,4096,0.008921778036488427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,768,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,5120,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,6144,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,512,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,12288,0.08490755822923447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,6144,0.04519111249181959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,7168,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,768,0.0074720001882976955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,7168,0.014203555054134794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,12288,0.055719110700819224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,5120,1.1509386698404949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,768,0.024969778127140466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,65536,0.05834222502178616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,6144,0.010795555180973478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,32,0.1535822285546197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,1536,0.008641777767075432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,3072,0.031367109881507024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,51200,0.10569777753618027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,8192,1.0605910619099934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,6144,0.029483556747436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,3584,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,16384,0.03559644354714288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,32,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,32,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,4096,0.03523644473817613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,768,0.019663110375404358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,7168,0.028070221344629925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,4096,0.04071111149258084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,2048,0.06497510936525133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,16384,0.01626044511795044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,64,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,64,0.01687733332316081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,3584,0.08888533380296494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,768,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,5120,0.08435110913382636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,256,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,5120,0.18101599481370714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,256,0.016566221912701923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,32,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,64,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,512,0.007484444313579136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,256,0.01886666648917728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,12288,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,65536,0.23795821931627062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,1536,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,6144,0.028689778513378564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,65536,0.19481155607435438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,5120,0.04779200090302361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,4096,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,6144,0.0376160012351142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,6144,0.013080000049538083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,65536,0.06174221966001722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,3584,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,4096,0.03163733416133457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,16384,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,10240,0.008711111214425828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,2560,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,5120,0.22926221953497994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,8192,0.1332026720046997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,256,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,2048,0.0480951103899214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,1024,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,65536,10.954010009765625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,51200,0.4065973228878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,5120,0.23193332884046766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,1536,0.028757333755493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,8192,0.02058844433890449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,3584,0.008017777568764156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,51200,0.14884532822502985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,8192,0.016170667277442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,1024,1.038691520690918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,10240,0.08656444152196248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,2048,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,51200,0.39694844351874453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,128,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,512,0.1678417788611518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,4096,0.11012444231245254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,32,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,5120,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,7168,0.051295999023649425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,5120,0.05343288845486111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,10240,0.025690666503376428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,7168,0.12235111660427517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,5120,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,65536,0.07758666409386529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,256,0.030440890126758154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,512,0.15820266140831843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,2560,0.03182399935192532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,1536,0.29556534025404185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,10240,0.0540284448199802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,7168,0.012995555169052549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,256,0.015505777464972602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,7168,0.013655111193656921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,2048,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,8192,0.38866045739915633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,256,0.016439111696349252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,256,0.03250577714708116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,4096,0.05964889129002889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,12288,0.0449813339445326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,8192,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,2560,0.006103111224042044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,6144,0.018213333355055917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,5120,0.009717333647939894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,768,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,51200,0.12696444988250732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,2560,0.026485333840052288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,16384,0.37252267201741535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,5120,0.27241865793863934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,10240,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,8192,0.024047111471494038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,10240,0.028295109669367473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,2048,0.01651822196112739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,1536,0.014505777094099255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,1536,0.006028444402747684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,7168,0.44786490334404844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,3072,0.10811644130282932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,64,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,768,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,2560,0.05786666605207655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,4096,0.010664888554149203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,32,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,3584,0.027796443965699937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,7168,0.05632266733381483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,12288,0.07699733310275607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,65536,0.11760621600680882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,128,0.010484444598356882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,2560,0.13008533583747015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,1536,0.028654221031400893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,512,0.010423999693658618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,256,0.005474666754404704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,12288,0.06689955790837605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,16384,0.03404622276624044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,2560,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,512,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,4096,0.161171555519104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,5120,1.4478089014689128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,12288,0.597067568037245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,1536,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,768,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,4096,0.032391998502943255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,6144,0.024480889240900677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,2560,0.03439555565516154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,65536,0.06746488809585571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,2560,0.0831128888660007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,4096,0.009409777820110321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,3072,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,1024,0.007804444266690149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,10240,0.015461333923869662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,12288,0.0472871098253462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,10240,0.015460444821251763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,8192,0.03597777750757005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,51200,0.4796702067057292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,16384,0.03163822160826789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,2048,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,1536,0.010035555395815108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,12288,0.017501332693629794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,5120,0.01365600029627482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,768,0.029915554655922785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,6144,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,2048,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,1536,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,1024,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,768,0.5305768648783366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,16384,0.01926488843229082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,65536,1.1029493543836806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,512,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,51200,0.05717066923777262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,768,0.00425155543618732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,32,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,1536,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,7168,0.013207111093733044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,65536,0.09449422359466553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,6144,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,256,0.008016889293988546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,7168,0.33165155516730416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,6144,0.1288799974653456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,2048,0.014556444353527494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,7168,0.13248444928063288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,32,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,128,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,6144,0.016598222984208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,128,0.026370666093296472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,10240,0.5684844652811686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,256,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,3584,0.013167999684810638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,16384,0.06111199988259209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,64,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,12288,0.03706489006678263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,12288,0.020676443974177044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,3072,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,4096,0.008164444731341468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,3072,0.041894220643573336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,3584,0.010395555860466426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,10240,0.015150222513410779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,3584,0.056143111652798124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,32,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,64,0.0026559999419583213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,5120,0.12016888459523518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,3072,0.02436977790461646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,3072,0.11620977189805771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,4096,0.01869511107603709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,2560,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,12288,0.017222222354676988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,6144,0.010943999720944298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,2560,0.035275555319256253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,3072,0.01938311093383365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,10240,0.31891822814941406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,768,0.01789422167672051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,65536,2.4442630343967013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,768,0.014119111829333834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,64,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,1024,0.45297331280178493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,32,0.003859555555714501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,128,0.010439111126793755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,6144,0.023774221539497375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,1024,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,1024,0.00676711110605134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,10240,0.400582234064738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,3584,0.030991110536787245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,2048,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,1536,0.022670222653283015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,4096,0.7669830852084689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,8192,0.016864000095261466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,768,0.030192888445324365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,2048,0.0068959999415609576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,51200,1.3469279607137044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,2048,0.013129777378506131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,128,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,256,0.027594667341974046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,16384,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,4096,0.0328382220533159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,5120,0.0896000001165602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,256,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,3584,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,8192,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,3584,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,7168,0.06668266985151503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,5120,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,2560,0.008573333422342936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,2560,0.010450666977299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,1536,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,7168,0.02088088956144121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,512,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,10240,0.009583999713261923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,1024,0.01275022245115704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,128,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,8192,0.029845333761639063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,10240,0.015026667051845126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,4096,0.15012533134884304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,1024,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,512,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,64,0.03604711095492045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,768,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,256,0.011733333269755045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,8192,0.1280453337563409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,12288,0.18602932824028862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,256,0.015115555789735583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,512,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,12288,0.01833066675398085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,2560,0.007132444116804335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,1024,0.019557333654827543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,1024,0.0159599996275372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,1536,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,256,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,2048,0.008930666579140557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,7168,0.842775132921007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,2560,0.020954666866196528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,8192,0.21408532725440133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,10240,0.06878399848937988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,4096,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,51200,1.217590226067437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,8192,0.0331857767370012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,4096,0.020818667279349435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,4096,0.018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,128,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,2048,0.01457244488928053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,768,0.06733955277336968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,128,0.02036622166633606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,8192,0.2771608829498291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,5120,0.05437155564626058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,10240,0.015119110544522604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,2560,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,51200,0.07541155815124512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,2560,0.024880000286632117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,51200,0.7290195359124078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,4096,0.1558311118019952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,12288,0.027832001447677612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,7168,0.012471111284361945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,12288,0.013440888788965015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,256,0.012987555729018318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,768,0.030054221550623577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,256,0.0045422220395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,64,0.005984000033802456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,65536,0.25004177623324925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,12288,0.2725306616889106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,51200,2.6715005238850913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,3584,0.08152711391448975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,4096,0.039507554637061224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,768,0.03480089041921828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,2048,0.0738791094885932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,1024,0.014132445057233175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,16384,0.03370577759212918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,10240,0.023916444844669763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,4096,0.07358311282263862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,512,0.011564444336626263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,8192,0.00903200027015474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,1536,0.02051999999417199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,16384,0.16869866847991943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,3584,0.01350311107105679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,51200,0.11710755030314128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,256,0.0023048888478014204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,3072,0.08421422375573052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,4096,0.039156443542904325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,64,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,12288,0.013454222016864352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,10240,0.012233778006500669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,512,0.045371555619769625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,8192,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,3584,0.05786844756868151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,768,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,1536,0.00646666685740153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,51200,2.106503168741862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,51200,0.483317322201199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,512,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,1024,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,4096,0.016470222009552848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,128,0.014106666048367819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,51200,14.129874335394966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,51200,0.06581066714392768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,4096,0.03432888785998026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,51200,1.127378675672743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,768,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,5120,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,64,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,16384,0.4100026554531521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,65536,0.24566756354437935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,1024,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,8192,0.09962666696972317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,32,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,6144,0.14231467247009277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,256,0.016156444946924847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,8192,0.03631822268168131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,1024,0.016602666841612924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,64,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,1024,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,32,0.00905866672595342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,2048,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,128,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,128,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,65536,0.8287902408175998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,2048,0.011339555184046427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,4096,0.044722666343053184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,8192,0.061828447712792285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,4096,0.05797511339187622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,10240,0.10724088880750869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,256,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,10240,0.02166666587193807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,7168,0.056299553977118603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,768,0.011988444460762871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,512,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,768,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,5120,0.09154399898317124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,10240,0.2550862100389269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,128,0.028295109669367473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,8192,0.16480710771348742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,51200,0.9466791152954102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,1536,0.03295022249221802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,51200,0.1772728893491957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,512,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,512,0.002867555452717675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,1536,0.028484443823496502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,7168,0.034487111700905695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,1536,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,128,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,3584,0.08356799681981404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,65536,0.12350222799513076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,512,0.03240711159176297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,51200,0.02970400121476915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,32,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,5120,0.08512266476949056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,768,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,256,0.011900444825490316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,51200,0.09445422225528294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,768,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,5120,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,3584,0.009959110783206092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,8192,0.03193155593342251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,32,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,10240,0.02608355548646715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,64,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,5120,0.07984622319539388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,4096,0.03275733192761739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,3584,0.028032888968785603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,16384,0.06270311276117961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,51200,0.04994755652215746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,1024,0.011705777711338468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,4096,0.01686844395266639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,8192,0.030574222405751545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,1536,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,2560,0.015352000792821249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,3072,0.4286773204803467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,2560,0.01629866659641266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,7168,0.03497688968976339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,8192,0.1978000005086263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,256,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,7168,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,16384,0.24811198976304796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,1024,0.016089777151743572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,65536,0.5183759795294868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,256,0.10937777492735122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,12288,0.017875555488798354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,16384,0.027142221728960674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,128,0.012127111355463663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,16384,1.3894347084893122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,3072,0.020557334025700886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,512,0.021922666165563796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,8192,0.8523004319932727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,65536,0.05266133281919691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,10240,0.016903999778959487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,512,0.010621333287821876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,51200,0.058407107988993325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,256,0.01759733259677887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,4096,0.012117333710193634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,4096,0.04663377669122484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,256,0.004519999855094486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,4096,0.05474399858050876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,12288,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,256,0.02049333353837331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,6144,0.014408000641398959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,16384,0.4420648945702447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,65536,0.09012889199786717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,3584,0.04903644323348999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,3072,0.4838159879048665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,6144,0.0232933329211341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,8192,0.010687111152542962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,64,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,4096,0.0359226663907369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,16384,0.03478577733039856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,128,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,32,0.012282667060693106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,8192,0.03323466579119364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,768,0.006861333217885759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,8192,0.07414666811625163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,128,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,16384,0.04025333457522922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,2560,0.02477600011560652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,3584,0.0370506677362654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,5120,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,6144,0.04027377896838718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,5120,0.010021333065297868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,3072,0.007693332930405934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,2048,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,8192,0.017465778523021273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,7168,0.02787733409139845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,8192,0.01310488912794325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,2048,0.030777778890397813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,51200,0.011902222202883827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,2048,0.015159999330838522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,1024,0.0037191112836201987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,7168,0.14077156119876438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,8192,0.22325510448879668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,4096,0.027983999914593164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,10240,0.2505902184380425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,16384,0.028117332193586562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,6144,0.10865422089894612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,5120,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,512,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,768,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,16384,0.08225599924723308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,4096,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,2560,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,2048,0.07007466422186957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,51200,0.04011111127005683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,32,0.48598045772976345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,1024,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,8192,0.010411555568377176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,12288,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,3584,0.015221332510312399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,8192,0.015658666690190632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,3072,0.020000888241661918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,256,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,768,0.018220444520314533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,51200,0.12373156017727321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,3584,0.023023999399609033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,12288,1.2836240132649739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,64,0.0032044444233179092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,12288,0.06878755489985149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,3072,0.050952888197369046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,16384,0.7706906530592176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,3584,0.09408177932103474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,8192,0.02160177793767717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,768,0.03415200114250183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,768,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,8192,0.25714045100741917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,3072,0.030269334713617962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,10240,0.0581173300743103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,8192,0.014700444208251106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,7168,0.06713955269919501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,768,0.023194667365815904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,10240,0.8498364554511176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,10240,0.0075866662793689305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,10240,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,16384,0.5235839949713813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,1536,0.009295110901196798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,6144,0.010973333484596677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,5120,0.1366000043021308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,32,0.01132266637351778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,64,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,768,0.19214044676886666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,512,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,7168,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,768,0.008691555923885768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,512,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,128,0.024490666058328416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,2048,0.11650221877627903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,12288,0.34264622794257266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,2048,0.03851200143496195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,16384,0.02772888872358534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,64,0.002554666664865282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,2560,0.05366222063700358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,128,0.01789155602455139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,8192,0.036532445086373225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,12288,0.03832533293300205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,7168,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,6144,0.023823110593689814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,3584,0.09603110949198405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,10240,0.05264977614084879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,256,0.022647110952271357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,2048,3.5731057061089406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,2560,0.07919022109773424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,1024,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,6144,0.1274062262641059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,8192,0.014013333453072442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,51200,0.07044355736838447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,1536,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,3584,0.0222053329149882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,7168,0.31245067384507924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,51200,0.10851289166344537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,65536,0.5975359810723199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,32,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,512,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,4096,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,7168,0.019852444529533386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,3072,0.03630133469899496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,128,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,3584,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,512,0.030236442883809406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,512,0.01889244384235806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,64,0.013087111214796701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,128,0.008375999828179678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,1536,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,2048,0.06433155801561144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,768,0.005981333139869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,5120,0.021990221407678392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,10240,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,512,0.0028524444335036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,512,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,5120,0.006903111106819577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,7168,0.03992444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,4096,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,64,0.013920888304710388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,64,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,65536,13.58538818359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,1024,0.02849777870708042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,128,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,512,0.004263111286693149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,1024,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,32,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,1024,0.01145066652033064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,256,0.014378666877746582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,65536,0.05365511112742954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,512,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,2048,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,2560,0.006697777658700943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,3072,0.9045866860283746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,8192,0.013815110756291283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,8192,0.028079112370808918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,128,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,1024,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,512,0.019073777728610568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,256,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,512,0.008422222402360704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,3584,0.028339554866154987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,1024,0.053378668096330434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,1024,0.003524444583389494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,10240,0.1395288838280572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,32,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,7168,0.0998035536872016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,2048,0.006335111127959357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,12288,0.026126222478018865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,10240,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,128,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,7168,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,3584,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,51200,0.10274489058388604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,3072,0.031215111414591473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,64,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,7168,0.05475733346409268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,6144,0.11281955242156982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,3584,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,2048,0.024121777878867254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,4096,0.014909333652920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,4096,0.015259555644459195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,768,0.022448889083332483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,768,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,4096,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,8192,0.1914693382051256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,512,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,1024,0.07187022103203668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,3584,0.064538664287991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,5120,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,2560,0.004961777892377642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,128,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,2560,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,8192,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,768,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,3584,0.02763644523090786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,65536,0.08234044578340319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,512,0.021329777108298406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,65536,0.061589333746168345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,768,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,16384,0.020577778418858845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,10240,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,7168,0.09255288706885444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,51200,0.055768887201944985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,5120,0.01701244380739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,2560,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,1024,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,1536,0.011674666570292579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,3072,0.025443555580245122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,6144,0.027373333772023518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,6144,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,10240,0.9744808408949109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,65536,0.1155431138144599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,16384,0.1648728847503662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,2560,0.019292443990707397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,6144,0.023755555351575215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,768,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,5120,0.0427706672085656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,16384,0.023980443676312763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,10240,0.15930399629804823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,3584,0.04004089037577311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,5120,0.01384800010257297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,256,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,64,0.4692506790161133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,1024,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,3072,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,1536,0.02442577812406752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,4096,0.02664888898531596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,1024,0.02748799986309475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,7168,0.012734221915404001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,12288,0.01348355578051673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,3584,0.03429777754677667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,3072,0.1230328877766927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,768,0.032820443312327065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,4096,0.10735111104117499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,4096,0.020454222957293194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,1024,0.01623288955953386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,3584,0.05488533443874783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,768,0.03280177712440491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,1536,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,65536,0.16267377800411648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,4096,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,1024,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,64,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,256,0.025394666526052687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,51200,0.028353777196672227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,768,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,65536,0.3959173361460368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,5120,0.5631084442138672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,8192,0.06573510832256742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,5120,0.10276266601350571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,128,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,5120,0.008652444514963362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,7168,0.13956977261437312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,65536,0.08817511134677464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,3072,0.05800444549984402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,12288,1.1567920049031575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,1024,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,3584,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,512,0.028318223026063707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,2048,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,3584,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,5120,0.008963555925422246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,768,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,768,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,5120,0.05992444356282552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,16384,0.03289244572321574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,256,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,7168,0.136409772766961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,4096,0.05166222320662605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,6144,0.05274933245446947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,128,0.0731182230843438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,2048,0.028873778051800195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,512,0.005399999933110342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,5120,0.12461688783433701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,51200,0.0473235547542572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,6144,0.03264710969395108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,8192,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,51200,0.2602550983428955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,1024,0.07590666744444105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,1024,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,768,0.022200889057583276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,16384,0.024452444579866197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,4096,0.03568355573548211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,3584,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,1024,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,512,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,1024,0.22540712356567383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,768,0.029598222838507757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,3584,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,7168,0.019612444771660697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,12288,0.1443342235353258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,64,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,7168,0.009938666390048133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,3072,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,3072,0.02405422263675266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,65536,0.12164178159501816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,3584,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,7168,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,6144,0.026735999517970618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,32,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,16384,0.06393422020806207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,3584,0.27984176741706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,16384,4.65068096584744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,16384,0.01594577729701996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,1024,0.009709333380063375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,16384,0.13806400034162733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,2560,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,1536,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,128,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,512,0.01789866718981001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,7168,0.01993955506218804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,1024,0.030001779397328694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,3072,0.010599111517270407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,8192,0.14822844664255777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,10240,0.018868444694413077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,768,0.04105955693456862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,6144,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,3072,0.01622577839427524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,8192,0.15809422069125706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,64,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,5120,0.03254666593339708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,3072,0.01832444469134013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,2560,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,32,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,51200,0.451982233259413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,12288,0.05886933538648817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,3072,0.2735919952392578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,6144,0.11065510908762614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,768,0.02404888967672984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,51200,0.028304000695546467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,3584,0.4091262287563748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,1536,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,7168,0.01772888832622104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,7168,0.04497333367665609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,128,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,2048,0.0429057776927948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,256,0.03919466667705112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,512,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,1024,0.029143999020258587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,51200,0.139408893055386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,2048,0.014091556270917257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,3584,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,16384,0.04398933384153578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,4096,0.09281422032250299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,1024,0.29658667246500653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,7168,0.022472888231277466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,4096,0.13522311051686606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,1024,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,2048,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,5120,0.009333333207501305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,1024,0.02476088868247138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,6144,0.08480533626344468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,7168,0.11821688546074761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,768,0.03707911239729987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,768,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,32,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,6144,0.033324443631702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,65536,0.07810666826036242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,3584,0.4310035440656874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,51200,0.017836444907718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,64,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,1536,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,6144,0.010175111393133799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,16384,0.030652443567911785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,8192,0.039276444249682956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,3584,0.007175111108356052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,1024,0.06273688872655232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,3584,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,7168,0.025765332910749648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,5120,0.29913245307074654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,7168,0.025849777791235182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,3072,0.021330666210916307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,3072,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,16384,0.21728177865346274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,6144,0.09501066472795273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,2048,0.03678044345643785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,768,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,3072,0.018217777212460835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,51200,0.07530399825837877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,10240,0.01277511070171992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,12288,0.049561778704325356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,1024,0.0031884445084465873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,1024,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,2048,0.00423200014564726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,1024,0.01717155509524875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,128,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,3072,0.20143911573621961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,3584,0.4635919994778103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,51200,0.04313244422276815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,768,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,128,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,768,0.01926666663752662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,5120,0.02614044480853611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,128,0.010061333576838175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,65536,0.154767115910848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,5120,0.027479110492600337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,3072,0.22717510329352486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,65536,0.13482133547465006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,256,0.023987554841571387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,256,0.019273777802785236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,2560,0.008095111283991072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,7168,0.01861422260602315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,5120,0.04929955469237434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,256,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,256,0.006454222318198945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,51200,0.9130702548556857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,10240,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,4096,0.00831200016869439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,4096,0.8834417661031088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,256,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,3072,0.01997599999109904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,128,0.014904888139830695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,5120,0.016924444172117446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,6144,1.975951936509874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,32,0.010663111176755695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,8192,0.02756711178355747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,16384,0.023004444109068975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,3072,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,512,0.012621333201726278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,12288,0.050108446015252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,51200,0.23416532410515678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,256,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,7168,2.8160461849636502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,51200,0.15027199851142034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,10240,0.08415022161271836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,16384,0.0831928915447659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,32,0.08146399921841092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,16384,19.891845703125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,10240,0.03336266676584879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,768,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,12288,0.1328008837170071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,8192,0.025285333395004272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,3072,0.05808889203601413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,65536,0.06330133146709867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,12288,0.2878826724158393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,6144,0.023745778534147475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,1536,0.028612444798151653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,512,0.013568888935777875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,64,0.13303644127315944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,3584,0.0332951115237342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,51200,0.21159911155700684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,16384,0.02568977740075853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,7168,0.053555554813808866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,8192,0.026084444589085046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,768,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,5120,0.060573332839541964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,128,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,5120,0.04491555690765381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,8192,0.025628444221284654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,7168,0.023779556155204773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,2048,0.0064026667839950984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,1024,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,6144,0.030942221482594807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,3584,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,1024,0.00350577798154619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,3072,0.030067556434207495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,256,0.01512888901763492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,768,0.027088888817363318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,128,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,51200,0.06548800071080525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,2048,0.07190399699740939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,12288,0.03155466583040025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,7168,0.034443553951051496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,64,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,6144,0.0281333327293396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,1536,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,51200,0.37471911642286515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,6144,0.05391022231843737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,128,0.02590755621592204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,16384,0.04851288927925957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,4096,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,2048,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,12288,0.2730960051218669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,16384,0.2939937909444173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,2048,0.0087333329849773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,3584,0.01893422173129188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,768,0.0027946666296985415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,4096,0.023434667123688593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,512,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,128,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,16384,0.03681688838534885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,10240,0.01145511120557785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,10240,0.014699555105633207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,6144,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,4096,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,256,0.019880889190567862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,12288,3.5933333502875437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,1024,0.028472890456517536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,256,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,6144,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,5120,0.01182044463025199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,6144,0.04145155681504144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,3584,0.2185493310292562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,2048,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,5120,0.019967110620604623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,32,0.002642666714058982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,12288,0.07396799988216825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,1536,0.005208888815508948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,65536,0.05033333433998955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,768,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,10240,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,1024,0.007254222200976477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,12288,0.030117332935333252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,3584,0.01646488904953003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,3584,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,16384,0.3033742109934489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,4096,0.01167555567291048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,1024,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,6144,0.17304533057742646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,6144,0.9244604110717773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,8192,0.01202755586968528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,2560,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,12288,0.02401688860522376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,128,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,3072,0.04708800050947401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,65536,51200,10.527667575412327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,32,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,7168,0.018279999494552612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,51200,6.749559190538194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,16384,0.05570399761199951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,1536,0.027095999982621934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,7168,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,32,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,3072,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,3072,0.013578666581047906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,256,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,16384,0.07025155756208631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,128,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,65536,0.0950195524427626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,8192,0.16725599765777588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,7168,0.045596443944507174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,128,0.008746667040718926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,4096,0.013772444592581855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,3584,0.03825689024395413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,64,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,128,0.007671111159854465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,128,0.043177776866488986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,12288,0.042542222473356456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,4096,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,8192,0.014329777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,51200,0.3005795478820801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,10240,0.22643645604451498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,256,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,512,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,3072,0.016202666693263583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,6144,0.01827911039193471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,128,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,7168,0.017603556315104168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,12288,0.02176711128817664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,32,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,64,0.0021946666141351065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,3584,0.009181333084901174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,12288,0.2220968935224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,10240,1.1747075186835396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,8192,0.05349333418740166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,7168,0.03102488981352912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,2048,0.0073928890956772705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,512,0.021853332718213398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,7168,0.050160888168546885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,6144,0.041060444381501936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,6144,0.06053511301676432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,512,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,16384,1.0746853086683485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,6144,0.13178755177391901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,64,0.003543111185232798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,768,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,2048,0.03934755590226915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,256,0.026269333230124578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,12288,0.019118222925398085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,12288,0.014845333165592618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,2048,0.028267555766635474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,3584,0.06438222196367052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,51200,6.16010242038303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,512,0.02365600069363912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,2560,0.07368977864583333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,12288,0.020337777005301583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,8192,0.0120328888297081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,16384,0.01847288840346866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,10240,0.12081510490841335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,65536,0.12067200077904595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,512,0.01218844453493754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,64,0.025192888246642217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,2560,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,128,0.017720000611411203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,512,0.013628444737858243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,8192,0.03738844394683838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,768,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,51200,0.06996444198820326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,2560,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,64,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,1024,0.01706133286158244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,51200,1.9321653578016493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,128,0.014548444085650973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,5120,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,7168,0.032109333409203425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,3584,0.010470222267839642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,16384,0.06029511160320706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,768,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,65536,0.13326756159464517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,10240,0.20168711079491508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,8192,0.04971111151907179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,1024,0.026271111435360376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,3584,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,2560,0.007358222372002072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,2048,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,10240,0.012141333685980903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,12288,0.6427591111924913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,8192,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,5120,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,6144,0.019261333677503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,12288,0.016570667425791424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,3584,0.12007821930779351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,64,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,2560,0.04403466648525662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,5120,51200,0.7889439794752332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,256,0.021976888179779053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,1024,0.013891556196742587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,12288,0.13682933648427328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,7168,0.024686222275098164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,1024,0.0031742221779293488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,7168,0.20015377468532988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,7168,0.06258844667010836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,51200,0.05885244740380181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,51200,5.563961876763238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,51200,0.9977253807915581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,1536,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,16384,0.02611644400490655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,1024,0.017535999417304993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,2560,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,10240,2560,0.016358221570650738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,6144,0.03691733214590285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,768,0.0029066666546795103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,16384,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,16384,0.8278657595316569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,12288,0.029955555995305378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,2560,0.02861066659291585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,1024,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,8192,0.015632000234391954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,12288,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,51200,0.04045955671204461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,8192,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,768,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,3584,0.03621066610018412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,2048,0.033956444925732084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,4096,0.018283555905024212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,2048,0.014228444960382251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,128,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,65536,0.09073599841859604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,16384,0.07397866911358304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,4096,0.05992888742023044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,51200,1.2331742180718315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,65536,1024,0.1642728911505805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,3584,0.21387199560801187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,1536,0.01276533305644989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,65536,0.27515554428100586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,2048,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,768,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,12288,0.05769689215554131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,6144,0.03146933184729682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,2560,0.058668441242641874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,7168,0.004824000100294749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,3584,0.02084355552991231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,1536,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,8192,0.02089244458410475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,768,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,256,0.0155102229780621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,3584,0.009907555249002244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,7168,0.015976000163290236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,1024,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2560,51200,0.33567378256056046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,51200,0.283348454369439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,65536,0.07161777549319796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,16384,0.05579288800557455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,64,0.0019946667469210094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,128,0.01792977750301361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,512,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,512,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,16384,0.42275023460388184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,10240,0.04707555638419258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,512,0.022563555174403723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,2048,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,512,0.010909332997269101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,3072,0.0302471121152242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,512,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,16384,0.46153688430786133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,12288,0.11078133185704549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,6144,0.12557155556148952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,3072,0.5819733407762315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,2560,0.019150222341219585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,128,0.008736888567606607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,65536,0.06671022044287787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,1536,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,7168,0.03882222374280294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,512,0.019926221834288705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,3072,0.04201689031389025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,1536,0.02834577692879571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,3072,0.05507911245028178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,65536,0.08333510822719997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,10240,0.0321324434545305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,5120,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,6144,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,768,0.003140444349911478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,2048,0.02072000006834666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,256,0.005730666634109285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,128,0.01789688898457421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,6144,0.012551999754375882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,2560,0.007516444557242923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,1024,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,5120,0.022336888644430373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,4096,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,2048,0.05360710952017042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,256,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,768,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,256,0.003154666680428717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,6144,0.012803555362754397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,2560,0.33348443773057723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,1536,0.021356445219781663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,7168,0.025776889589097764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,7168,0.007820444802443186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,64,0.008628444539176093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,65536,0.021892444954978094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,12288,0.013743110828929476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,768,0.019113777412308585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,768,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,51200,0.07054666678110759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,2560,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,4096,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,32,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,65536,0.3506017790900336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,256,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,16384,0.06953244739108615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,16384,0.03677689035733541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,8192,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,2560,0.015640000502268474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,1024,0.14388088385264078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,65536,0.07663822174072266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,1536,0.03228977653715346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,2048,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,12288,0.028018666638268366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,10240,0.02493155499299367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,51200,0.03827111257447137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,64,0.00833777752187517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,3072,0.05523733297983805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,256,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,512,0.018146667215559218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,4096,0.03756444321738349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,16384,0.06991733445061578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,32,0.0028088889602157804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,1024,0.021691555778185528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,6144,3.6638391282823353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,12288,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,6144,0.6395653088887533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,3072,0.011582222249772815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,4096,0.05122577812936571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,16384,0.1364133358001709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,64,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,10240,0.03188088867399428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,5120,0.030755556292004053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,10240,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,10240,1.517624855041504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,7168,0.10853333605660333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,51200,9.856591966417101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,8192,0.12386043866475423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,4096,0.08179644743601482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,10240,0.027979556057188246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,2048,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,2560,0.028287112712860107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,12288,0.06568444437450834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,16384,0.038161777787738375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,64,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,768,0.04433155390951368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,8192,0.0602088901731703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,256,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,10240,0.0381004446082645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,16384,1.521534177992079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,16384,0.6432479752434624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,1024,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,8192,0.013433777623706393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,10240,0.004848888764778773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,7168,0.05286488930384318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,5120,0.013796444568369122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,16384,1.5482684241400824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,4096,0.092448890209198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,4096,0.04071822100215488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,6144,0.026326222552193537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,1536,0.010142222046852112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,1024,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,4096,0.08337066570917766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,2560,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,512,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,1536,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,2560,0.4126106633080377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,12288,0.02850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,7168,0.44447024663289386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,2048,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,65536,0.33847822083367246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,1536,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,51200,0.46771377987331814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,51200,0.2800275484720866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,16384,0.05374933282534281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,12288,0.11090488566292657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,1536,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,3584,0.016050666570663452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,16384,0.48538843790690106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,12288,0.019305777218606737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,768,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,2560,0.04388977752791511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,5120,0.011852444046073489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,256,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,8192,0.013635555903116861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,4096,4.838558197021484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,1024,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,7168,0.15758932961357966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,8192,0.2878071202172173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,3584,0.07628889216317071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,2048,0.01802399920092689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,10240,0.01423288881778717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,3584,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,12288,0.22803288035922578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,10240,0.020959999826219346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,10240,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,6144,0.01529600057337019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,1024,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,16384,0.06213066975275675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,1536,0.024345777100986902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,1024,0.018948444061809115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,51200,3.273538589477539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,128,0.03222666515244378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,12288,0.08288711309432983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,12288,0.1031493345896403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,3584,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,16384,0.025028443998760645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,7168,0.07446222172843085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,2560,0.04596977763705783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,128,0.007755555212497711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,16384,0.04119199845525954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,2560,0.3174915578630236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,768,0.007802666889296637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,16384,0.34264177746242946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,1536,0.0063573333124319715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,2560,0.03021866745418972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,256,0.0123511114054256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,16384,0.30039199193318683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,4096,0.031543999910354614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,6144,0.05247555507553948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,8192,0.11479111512502034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,10240,0.07107110818227132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,3072,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,3072,0.007358222372002072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,4096,0.029127111037572224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,3584,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,3584,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,5120,0.008584000170230865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,7168,0.026025777061780293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2560,65536,7.4888517591688375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,5120,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,16384,0.016480889585283067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,4096,0.28657955593532985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,3072,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,32,0.0118568887313207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,1024,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,12288,0.23165599505106607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,12288,0.030118223693635728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,6144,0.013275555438465543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,128,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,2560,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,512,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,65536,0.07798044549094306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,8192,0.02885599931081136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,12288,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,512,0.015356444650226168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,6144,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,256,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,10240,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,7168,0.051707555850346885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,2048,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,1024,0.05797511339187622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,5120,0.016639111770523917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,1536,0.061885330412122935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,7168,0.08983021974563599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,1024,0.02243999971283807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,2048,0.032454222440719604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,4096,0.06544266806708442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,16384,0.03299999899334378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,3584,0.03743377659055922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,6144,1.0003973642985027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,4096,0.09102933274375068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,1536,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,1024,0.004222222086456087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,65536,1.1031057569715712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,12288,0.028182221783532038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,3072,0.011564444336626263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,7168,0.1857751078075833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,5120,0.012870222330093384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,512,0.017800889081425138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,1536,0.019753777318530612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,512,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,3072,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,3072,0.04906310968928867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,4096,0.021789333886570398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,768,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,512,0.006555555595291986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,65536,0.04993333419164022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,12288,0.31662045584784615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,768,0.011897777517636618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,8192,0.07685599724451701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,16384,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,768,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,2048,0.024111110303137038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,4096,0.012717333104875354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,4096,0.06648088826073541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,4096,0.037137776613235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,2048,0.010288000106811523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,3584,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,2560,0.020207999481095206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,5120,0.010737777584128909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,51200,0.020421332783169217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,8192,0.02770044406255086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,16384,0.0073626670572492815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,8192,0.0464897784921858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,65536,0.19870932896931967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,16384,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,1536,0.026518222358491685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,3584,0.014490667316648694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,128,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,16384,1536,0.035423109928766884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,3584,1.750483512878418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,16384,0.028654221031400893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,16384,0.06895822286605835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,512,0.010404444403118556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,3072,0.01146577795346578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,51200,0.010737777584128909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,3072,0.009543110926946005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,768,0.002765333279967308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,768,0.01887822151184082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,51200,0.05222310953670078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2560,3584,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,256,0.027875555886162653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,256,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,7168,0.01939377850956387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,32,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,8192,0.02251288957066006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,10240,0.008711111214425828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,8192,0.013794666363133324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,4096,0.008292444050312042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,3584,0.00601333338353369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,2048,0.020449777444203694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,1536,0.006069333189063602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,768,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,16384,0.056240889761182994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,7168,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,51200,0.12114132775200738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,64,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,2048,0.0063928887248039246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,2560,0.06462844212849934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,8192,0.18450489309098986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,7168,0.4681102434794108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,1024,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,7168,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,256,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,10240,0.05105333195792305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,51200,0.2477182282341851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,256,0.024359110328886244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,1024,0.0031822222388452957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,6144,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,12288,0.1812631156709459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,10240,0.029283556673261855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,3584,0.05021689004368252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,32,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,5120,0.1398622194925944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,8192,0.3665582338968913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,512,0.015344889627562629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,768,0.012597333225939008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,5120,0.030516445636749268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,12288,0.05277422070503235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,128,0.006024000131421619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,51200,0.10188355710771348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,8192,0.027813331948386297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,12288,0.035071998834609985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,5120,0.07215999894671969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,32,0.011426666544543372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,6144,0.008108444511890411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,6144,1.2865226533677843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,10240,0.014312888185183207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,5120,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,51200,0.053377777338027954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,5120,0.021928888228204515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,128,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,6144,0.10035110844506158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,12288,0.013399110900031196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,1024,0.026328888204362657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,7168,0.03797244363360935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,1024,0.01553244392077128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,65536,0.18044711483849418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,8192,0.3366702132754856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,512,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,768,0.05467555589146084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,768,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,768,0.0028195555011431375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,3072,0.020259555843141343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,64,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,768,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,10240,0.016785777277416654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,7168,0.13579377863142225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,1536,0.01995111174053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,7168,0.14067822032504612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,4096,0.029511110650168523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,10240,0.012895999683274163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,32,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,512,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,4096,0.15322577953338623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,512,0.013403555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,65536,0.8697528839111328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,4096,0.02035199933581882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,7168,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,1536,0.07473778062396579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,1536,0.012757333616415659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,256,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,2560,0.011543999943468304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,5120,0.908932473924425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,1536,0.022133333815468684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,10240,0.2856711016760932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,4096,0.10080799791547988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,3072,0.15220177173614502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,256,0.00639111093348927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,8192,0.024217777782016333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,512,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,51200,0.17070044411553276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,2048,0.022728888524903193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,512,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,5120,0.01549600064754486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,8192,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,32,0.0031644445326593188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,512,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,3584,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,3072,0.0074515557951397365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,10240,0.027958220905727808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,768,8192,0.0832666688495212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,16384,0.022658665974934895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,1024,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,512,0.005427555491526921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,3072,0.05132800009515551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,65536,0.1257351107067532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,12288,0.020414221617910597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,65536,0.07431911097632514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,2560,0.028763555818133887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,256,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,4096,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,12288,0.14652089277903238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,512,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,256,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,12288,0.1523804399702284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,32,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,3584,0.10372533400853474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,12288,0.37690311008029515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,128,0.017515555024147034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,16384,0.16232177946302626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,1024,0.04064888755480448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,51200,0.27671554353502065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,3584,0.06910844643910725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,32,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,2560,4.070334116617839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,12288,4.288631863064237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,2048,0.012652444342772165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,1536,0.006630222416586346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,1536,0.0237137774626414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,256,0.0052879999081293745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2048,8192,0.01367288910680347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,8192,0.03089955449104309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,51200,0.0520648890071445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,3584,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,512,0.01865244408448537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,2560,0.010140444669458602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,4096,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,12288,0.030569778548346624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,1536,0.003564444267087512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,3072,0.09734400113423665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,1536,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,65536,0.12615999910566542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,3584,0.10814755492740208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,65536,1.0477298100789387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,4096,0.032368001010682844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,51200,0.0206977774699529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,65536,0.3073795636494954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,768,0.013065777719020844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,65536,0.28598933749728733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,3584,0.06052533123228285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,128,0.04857066604826185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,64,0.0041617775956789655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,32,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,768,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,12288,0.020791111720932853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,1536,0.009766221874290043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,8192,0.039389332135518394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,65536,0.07738044526841906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,12288,0.031090666850407917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,128,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,5120,0.011164444188276926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,8192,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,10240,0.11577510833740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,7168,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,4096,0.017834666702482436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,512,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,3072,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,2560,0.012993777791659037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,65536,0.04998311069276598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,6144,0.035029331843058266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,2560,0.013749333719412485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,8192,0.1359671089384291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,1024,0.006077333456940121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,256,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,65536,0.2079413334528605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,1536,1.9856141408284504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,10240,0.16166488329569498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,65536,0.3950800100962321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,256,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,7168,0.08149866925345527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,65536,0.15149244997236463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,16384,0.1824035512076484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,6144,0.3650080098046197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,3584,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,128,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,1536,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,256,0.012443555725945367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,1536,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,51200,0.09905778037177192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,51200,0.17653600374857584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,768,0.04606489009327359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,3072,0.009286222358544668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,3584,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,512,0.005478222337034013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,12288,0.02471911079353756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,16384,0.021747555997636583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,256,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,32,0.09021955728530884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,768,0.002448888909485605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,3072,0.025067556235525344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,6144,0.010256888965765635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,3584,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,3584,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,5120,0.03121688961982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,6144,0.00981155534585317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,16384,0.06079466475380791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,7168,0.028765334023369685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,16384,0.016677333248986136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,3584,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,3072,0.010656889114114972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,51200,1.0925928751627605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,3584,0.018820444742838543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,10240,0.010679999987284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,512,0.00666222224632899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,512,0.1553537713156806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,2048,0.056348443031311035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,3584,0.027383999692069158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,12288,0.022687999738587275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,3072,0.020215999748971727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,768,0.021312889125612047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,12288,0.06549777587254842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,12288,0.35388000806172687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,51200,0.06192533175150553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,4096,0.22814843389723036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,64,0.018966222802797954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,16384,0.3387795554267035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,6144,0.10332177744971381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,16384,0.17089333799150255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,7168,0.026537777649031744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,8192,0.15266044934590658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,64,0.007687999970383114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,7168,0.022993778189023335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,128,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,512,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,1536,0.018619555566045973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,512,0.01667644414636824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,12288,0.025958221819665697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,256,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,1024,0.012285333540704517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,3584,0.01684711045689053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,1536,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,2048,0.03071999881002638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,3072,0.011128889189826118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,512,0.0028248888750871024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,3584,0.08804888857735528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,65536,0.028840889533360798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,2560,0.03349155518743727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,4096,0.03765689002143012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,8192,0.02426311042573717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,2048,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,2048,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,1536,0.10469777716530694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,51200,0.024870221813519795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,10240,0.03171822097566392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,10240,0.30005423227945965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,10240,0.10211111439598931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,65536,0.09718222088283962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,6144,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,128,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,1024,0.005971555494599872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,51200,0.03542133172353109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,4096,0.013895110951529609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,1024,0.02427911096149021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,768,0.017699556218253244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,1024,0.03599377804332309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,7168,0.02758755617671543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,768,0.009347555538018545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,6144,0.20545244216918945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,10240,0.03393422232733832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,10240,0.03141422073046366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,4096,0.022737777895397607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,65536,0.2690231005350749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,512,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,768,0.00281866660548581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,768,0.005381333331267039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,256,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,768,0.008025777836640676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,5120,0.035178667969173856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,2048,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,2048,0.11754577689700657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,64,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,5120,0.01034577770365609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,2560,0.04857688811090258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,51200,0.66903289159139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,6144,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,4096,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,2560,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,8192,0.015174221661355762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,65536,2.4599102867974176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,12288,0.1008639997906155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,5120,0.8323884540134006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,64,0.008972444468074376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,1024,0.2402142153845893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,64,0.005533333453867171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,1536,0.018574222922325134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,64,0.001639111070997185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,3584,0.009461333354314169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,2560,0.020946666598320007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,12288,0.024255111813545227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,3072,0.031481779283947416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,8192,0.04608622193336487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,10240,0.05622577667236328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,51200,0.14646311601003012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,5120,0.0950488911734687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,1024,0.04034666551484002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,8192,0.01960444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,12288,0.08685955736372207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,16384,0.08136266469955444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,6144,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,1024,0.004476444588767158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,512,0.019331556227472093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,128,0.3483733336130778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,5120,0.020635555187861126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,51200,0.15454310841030544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,12288,0.11728178130255805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,64,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,16384,0.016768889294730294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,51200,0.05002311203214857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,64,0.006653333289755716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,32,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,12288,0.04370933439996508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,8192,0.3075813452402751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,16384,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,2048,0.03867910967932807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,6144,0.010737777584128909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,256,0.014519110321998596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,51200,0.8089440133836535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,6144,0.009076444639099969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,6144,0.02332533399264018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,6144,0.1446239948272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,2048,0.009884444375832876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,256,0.027608889672491286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,51200,0.1028755572107103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,1536,0.03868088788456387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,32,0.001855111163523462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,6144,0.10775199863645767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,51200,0.20138222641415068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,1536,0.3287715646955702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,3072,0.3264479902055529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,768,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,7168,0.11361332734425862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,2048,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,65536,0.13253955046335855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,16384,0.1389840046564738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,12288,0.013755555782053204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,4096,0.03316088848643833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,51200,0.11019555727640788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,8192,0.016680889659457736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,12288,0.03717689050568475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,12288,0.02302933401531643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,128,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,256,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,3584,0.06841688685946994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,51200,0.10059999757342869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,3072,0.05998133288489448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,10240,0.03192977772818671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,2048,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,51200,0.34375821219550234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,8192,0.030212445391549006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,768,0.00509688879052798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,4096,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,5120,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,2048,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,64,0.011053333679835001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,4096,0.1188888947168986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,12288,0.06872177786297269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,3072,0.021299555897712708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,10240,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,3584,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,16384,0.01882133384545644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,768,2048,0.0295688874191708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,12288,0.07362577650282118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,1536,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,6144,0.009595555563767752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,3584,0.005680888891220093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,2048,0.02882044514020284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,768,0.2534373336368137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,51200,0.06565866867701213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,64,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,65536,0.312919987572564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,65536,0.06331200069851346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,10240,0.22999821768866646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,3072,0.007672888537247975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,65536,0.31883109940422905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,6144,0.00942666663063897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,65536,1.8471688164605036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,1536,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,4096,0.17131288846333823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,3072,0.06828088892830743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,4096,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,8192,0.1717217763264974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,1536,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,3072,0.019036443697081674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,128,0.011598221957683563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,7168,0.05339289042684767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,1536,0.06349244382646349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,1024,0.008834666675991481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,7168,0.011969777444998423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,65536,512,0.2208568917380439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,16384,0.035016000270843506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,6144,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,1536,0.026958222190539043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,10240,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,5120,0.012210666305489011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,10240,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,65536,0.9216701719495984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,128,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,512,0.012128888732857175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,65536,0.0808417797088623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,3584,0.0410462220509847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,256,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,6144,0.13560088475545248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,768,0.014519999424616495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,7168,0.012075555821259817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,768,16384,0.04763111140992907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,8192,0.025324443976084392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,1536,0.00646311127477222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,51200,0.07526222202512953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,2048,0.09578666422102187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,64,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,1024,0.046020444896486074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,5120,0.01184177812602785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,8192,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,2048,0.14916443824768066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,65536,0.407895114686754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,10240,0.024979554944568213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,6144,0.031175110075208876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,5120,0.032255109813478254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,2048,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,8192,0.10782044463687473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,128,0.0151555554734336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,512,0.010960000256697336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,1024,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,5120,0.7137466536627876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,10240,0.02947910957866245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,768,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,6144,0.034232000509897866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,1536,0.011355555719799466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,256,0.023355555203225877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,16384,0.01649777756796943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,3584,0.023910222782029047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,51200,16.942138671875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,2560,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,4096,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,16384,0.07663910918765597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,512,0.004007110993067424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,65536,0.08606311347749497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,6144,0.00942666663063897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,10240,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,16384,0.04765599966049194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,10240,0.011378666592968835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,2048,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,3584,0.03275200062327915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,8192,0.08122133546405368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,16384,0.042429334587521024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,16384,0.040427555640538536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,1024,0.01754488878779941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,64,0.004249777644872665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,1536,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,51200,0.15918844276004368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,1024,0.015435554915004306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,1024,0.01603200038274129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,8192,0.7850746578640408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,16384,0.028047998746236164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,2560,0.05082044336530897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,6144,0.04992622137069702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,3072,0.05634399917390612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,512,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,128,0.004921777794758479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,3072,0.023253333237436082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,65536,0.6378613577948676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,51200,0.043819556633631386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,5120,0.020689777202076383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,7168,0.03515911102294922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,65536,0.560624016655816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,32,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,1536,0.2735697693294949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,65536,0.867103099822998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,2048,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,16384,0.023009777069091797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,1536,0.03966933488845825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,1536,0.02754488918516371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,5120,0.010656000011497073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,7168,0.02331111166212294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,7168,0.3441937764485677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,16384,0.04244977898067898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,10240,0.025748444928063288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,12288,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,512,0.007078222102589077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,512,0.02033155494266086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,65536,0.08338844113879734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,16384,0.1897937721676297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,10240,0.027036445008383855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,64,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,1536,0.08109244373109606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,10240,0.04968711071544223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,51200,0.5286222034030491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,3072,0.062036441432105176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,512,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,7168,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,768,0.025438222620222304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,3072,0.022759111391173467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,3072,0.004262222184075249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,1024,0.018410666121376883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,3072,0.14986844857533774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,2048,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,1536,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,2048,0.5894008742438422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,128,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,2560,0.0863342218928867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,51200,1.810543908013238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,12288,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,16384,4.243679894341363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,32,0.00907022257645925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,51200,0.17139644092983672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,51200,0.10296889146169026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,8192,0.04539377821816338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,512,0.02421244482199351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,3072,0.011789333489206104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,4096,0.03787111242612203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,5120,0.09651288721296523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,1536,0.07213155428568523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,3584,0.06320355335871379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,16384,0.040756444136301674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,768,0.018065777089860704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,3072,0.02161066730817159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,1024,0.032791998651292585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,256,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,65536,0.058992001745435924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,6144,0.02436888880199856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,3584,0.01926666663752662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,3072,0.02959644463327196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,512,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,256,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,2560,0.1028124425146315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,8192,0.012771555946932899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,6144,0.016597333881590102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,16384,0.05539733171463013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,3072,0.05872178077697754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,12288,0.03481955660714044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,51200,0.8709617720709907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,512,0.025253333979182776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,10240,0.04218577676349216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,4096,0.014935111006100973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,16384,0.019016888406541612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,16384,0.032799998919169106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,512,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,256,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,128,65536,0.14422399467892116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,768,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,7168,0.07134933604134454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,65536,0.16264443927341038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,5120,0.038550221257739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,12288,0.11158043808407253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,512,0.005294222384691238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,4096,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,51200,0.25812355677286786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,3584,0.010963555839326648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,768,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,2560,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,512,0.020280000236299302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,3072,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,4096,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,6144,1.1659751468234592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,3072,0.04382133483886719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,10240,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,6144,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,768,0.013204444613721637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,4096,0.016508445143699646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,1536,0.1099697748819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,10240,0.11938667297363281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,256,0.055813332398732506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,512,0.11687733067406549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,2048,0.020283554991086323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,12288,0.025098666548728943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,51200,0.06618577904171415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,5120,0.08993244171142578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,3584,0.013623110949993134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,16384,0.050352887974845044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,7168,0.06870133346981473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,7168,0.04334577918052673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,1536,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,16384,0.041490667396121554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,3072,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,5120,0.09367999765608047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,5120,0.008440888590282863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,8192,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,4096,0.04649688800175985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,12288,1.049141354031033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,51200,4.8780801561143665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,128,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,51200,0.12146843804253472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,128,0.03490399983194139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,8192,0.18113956186506483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,2048,0.02405422263675266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,10240,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,2560,0.06478399700588651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,512,0.014021333720948962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,2048,0.05261333452330696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,8192,0.013245333400037555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,12288,0.20511199368370903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,128,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,10240,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,512,0.014522666732470194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,64,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,12288,10240,0.30040889316134983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,65536,0.4128257698482937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,128,0.030533333619435627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,128,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,2048,0.020478222105238173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,64,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,128,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,512,0.027836445305082533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,51200,0.5153857866923014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,512,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,8192,0.005197333378924264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,32,0.014139556222491793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,51200,0.2600959936777751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,8192,0.03727555606100295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,1024,0.023898666103680927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,768,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,12288,0.03716177741686503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,32,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,3584,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,10240,0.10984000232484604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,5120,0.02723377777470483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,1536,0.011347555451922946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,768,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,8192,0.04703466759787666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,1536,0.06024622254901462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,1024,0.015787555111779105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,4096,0.014429334137174817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,4096,0.06508977545632257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,12288,0.03571822245915731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,65536,0.07277866866853502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,1536,0.07698577642440796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,4096,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,4096,0.032893333170149065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,256,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,8192,3.1895955403645835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,512,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,4096,0.010644444160991244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,12288,3.109423107571072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,64,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,1536,0.11044088999430339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,256,0.013741333451535968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,1536,0.00958577791849772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,1024,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,12288,0.06520177920659383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,128,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,1024,0.010176888770527309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,6144,0.02162844439347585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,7168,0.032722665203942194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,51200,0.11869600084092881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,32,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,32,0.5647155443827311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,4096,0.00943288869327969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,7168,0.10566311412387425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,16384,0.038097778956095375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,10240,0.14761066436767578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,256,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,8192,0.051184889343049794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,51200,0.09677777687708537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,4096,0.0782346659236484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,3072,0.010127999716334874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,256,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,7168,0.39078932338290745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,1024,0.023923556009928387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,2560,0.01871288816134135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,3584,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,1536,0.13780532942877874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,1024,0.08626844485600789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,3584,0.00453599997692638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,65536,0.36211734347873265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,6144,0.027112000518374976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,51200,0.4306488831837972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,512,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,6144,0.03439022103945414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,3584,0.18021867010328504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,16384,3.3084000481499563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,2048,0.028522666957643297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,256,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,3072,0.009137777818573846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,2560,0.01793333391348521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,5120,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,3072,0.16457777553134495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,768,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,10240,0.04549244377348158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,128,0.015011555618709989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,64,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,128,0.0022622222701708474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,8192,0.017103110750516255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,3584,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,3072,0.023357333408461675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,256,0.022566222482257422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,256,0.021744888689782884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,2048,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,10240,0.06314400169584486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,768,0.0028044444819291434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,10240,0.018210666047202218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,8192,0.26617956161499023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,1024,0.1328764491611057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,5120,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,10240,0.0583546691470676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,2560,0.045167999135123364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,16384,0.02200711104604933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,2048,0.025475554996066626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,7168,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,16384,0.029317332638634577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,128,0.004067555483844546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,51200,0.5042675336201986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,6144,0.12785777780744764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,128,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,1024,0.006204444501135085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,256,0.01518133282661438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,1536,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,12288,0.09502577781677246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,8192,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,8192,0.347196446524726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,1024,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,256,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,16384,0.020266667008399963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,2560,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,3072,0.008171555896600088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,512,0.028759999407662287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,6144,0.5949138005574545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,16384,0.14772266811794704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,1024,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,5120,0.016339555382728577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,65536,1.1597937477959526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,128,0.017314665847354464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,65536,0.5356399748060439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,7168,0.10883466402689616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,256,0.028991109795040552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,5120,0.01824355622132619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,1536,0.00655644428398874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,6144,0.011759999725553723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,768,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,128,0.01994400057527754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,65536,0.09206489059660171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,128,0.02496444516711765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,16384,0.5334169069925944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,16384,0.6201528973049587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,2048,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,32,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,65536,0.06530044476191203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,6144,0.018938667244381376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,16384,0.037575112448798284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,7168,0.08207555611928304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,12288,0.012805333567990197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,2560,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,4096,0.08587466345893012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,6144,0.06616089079115126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,256,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,3072,0.09843466679255168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,16384,0.0709199971622891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,2560,0.037054220835367836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,10240,0.09198577536476983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,8192,0.010394666757848527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,768,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,51200,0.37659377521938747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,16384,0.08806488911310832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,256,0.020205333828926086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,768,0.026872888207435608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,65536,0.22055110666486952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,1536,0.06833155287636651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,2560,0.018717333674430847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,1536,0.004559999952713649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,3584,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,3072,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,16384,0.023066666391160753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,6144,0.030613332986831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,64,0.01848444508181678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,128,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,5120,0.0382008883688185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,12288,0.02976977825164795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,768,0.005653333332803514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,3584,0.1402719948026869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,5120,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,1536,0.17259377903408477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,3072,0.018838221828142803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,3072,0.057349330849117704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,12288,0.03763733307520548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,4096,0.008045333127180735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,2048,0.006729777902364731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,1024,0.020076443751653034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,65536,2.7638710869683156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,3584,0.02231377859910329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,6144,0.057849778069390186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,1536,0.0035626664757728577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,10240,0.012367111113336353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,256,0.10468000173568726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,12288,0.10326666964424981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,1536,0.05773599942525228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,16384,0.02718577782313029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,8192,0.0218551109234492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,128,0.0025475554996066624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,128,0.009361777868535783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,6144,0.04627466532919142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,7168,0.02147644427087572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,5120,0.3751502301957872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,1536,0.025332444243960913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,256,0.03496888942188687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,12288,0.09720000293519761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,1536,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,2048,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,64,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,8192,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,768,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,8192,0.03915999995337592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,4096,0.048947556151284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,1024,0.007038222418891058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,4096,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,65536,0.19599821832444933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,7168,0.018952000472280715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,7168,0.03750666644838121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,1024,0.01568266749382019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,8192,0.010783111055692038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,51200,0.05395999881956312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,512,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,256,0.017508443858888414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,3072,0.02971200148264567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,12288,0.04323022233115303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,1536,0.007439999944633908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,2560,0.014195554786258273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,6144,0.012481778032249875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,512,0.01551644504070282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,65536,0.46135732862684464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,16384,1.7151919470893011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,6144,0.004918222212129169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,256,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,1536,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,2048,0.017635555730925668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,4096,0.026406221919589575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,6144,0.033703111939960055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,16384,0.01891911029815674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,51200,0.1091102229224311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,6144,0.018068444397714403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,2048,0.00904266701804267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,7168,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,4096,0.009191110730171204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,64,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,8192,0.016640888320075143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,32,0.0069013333155049225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,3072,0.009136888715955947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,51200,0.014110222458839417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,3584,0.00849333322710461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,2048,0.006689777804745569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,10240,0.03000088863902622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,5120,0.025054223007626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,768,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,4096,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,65536,0.07346844673156738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,16384,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,256,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,12288,0.08284088638093737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,1024,0.01925066610177358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,3584,0.23062043719821504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,512,0.025400888588693406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,2048,0.014030221435758801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,2048,0.012647111382749347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,128,0.07254578007592095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,256,0.18865511152479383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,4096,0.13532266351911756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2560,6144,0.030074665943781536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,65536,0.0686746637026469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,5120,0.06828888919618395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,512,0.03547733359866672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,3072,0.05501866671774122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,1536,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,12288,0.03874666823281182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,10240,0.016057777735922072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,128,0.009378666679064432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,51200,0.04342666599485609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,2560,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,51200,0.016534222496880423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,3072,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,6144,0.025294222765498694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,3072,0.02614044480853611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,5120,0.020964443683624268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,12288,0.06235288911395603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,1024,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,8192,2.8768872155083547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,2560,0.03563200102912055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,2048,0.026198221577538386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,512,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,16384,0.06936799817615084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,3072,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,3584,0.02790488799413045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,8192,2.33463372124566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,1024,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,2560,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,32,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,8192,0.06986755794949003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,2048,0.055452442831463285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,51200,1.0555564032660592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,6144,0.012789333032237159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,3584,0.03763733307520548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,768,0.013980444934633044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,768,0.02404800057411194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,7168,0.20356444517771402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,6144,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,10240,0.03978844483693441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,6144,0.07491377989451091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,51200,0.1508906682332357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,10240,0.013798221945762634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,65536,0.028479999966091577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,1536,0.039099554220835366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,7168,0.05423022309939066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,5120,0.05999200211630928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,3072,0.10951199796464707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,128,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,4096,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,4096,0.013853333062595792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,65536,12288,1.9482027689615886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,512,0.0041751112374994485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,256,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,32,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,256,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,8192,0.04136888848410712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,2048,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,7168,0.021792888641357422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,3584,0.022988445229000513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,3072,0.025427555044492085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,10240,0.018005333013004728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,128,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,16384,0.6754933463202583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,1024,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,2560,0.1090808908144633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,12288,0.033915556139416166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,3072,0.36319377687242294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,4096,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,12288,0.01901422275437249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,128,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,5120,0.10524266958236694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,8192,0.011354666617181567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,12288,0.07237155569924249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,7168,0.07233244180679321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,5120,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,32,0.0019084444890419643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,8192,0.02479377720091078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,2048,0.030164443784289893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,12288,0.03598489032851325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,2560,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,64,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,3584,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,3584,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,768,0.018366222580273945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,12288,0.07012622223960029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,512,0.004511111312442356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,1024,0.16888622442881265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,16384,0.06536177794138591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,2560,0.01832355558872223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,256,0.017539555827776592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,10240,0.15306844976213244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,1536,0.01937866707642873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,8192,0.02550044490231408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,32,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,4096,0.07019822465048896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,512,0.005639111002286275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,3584,0.019851555426915485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,1024,0.10053422053654988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,4096,0.03897600041495429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,128,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,256,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,12288,0.005020444591840108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,16384,0.0236471113231447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,3584,0.02444355520937178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,51200,0.13456800248887804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,10240,0.040362666050593056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,768,0.016241777274343703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,768,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,64,0.14382310708363852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,3584,0.011023110813564725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,7168,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,4096,0.009892444643709395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,256,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,768,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,3584,0.06139466497633192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,8192,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,256,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,3584,0.03334844443533156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,51200,0.047232889466815524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,4096,0.02810488806830512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,3072,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,1024,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,32,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,8192,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,51200,4.5492329067654085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,51200,1.5872746573554144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,12288,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,32,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,2560,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,1536,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,512,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,7168,0.11390222443474664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,7168,0.014594667487674289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,512,0.012334222594896952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,3072,0.07128710879219903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,512,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,1024,0.19557689295874703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,3072,0.021500445074505273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,65536,0.09884177976184422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,3072,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,16384,0.041526221566730075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,4096,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,65536,0.05789866712358263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,1024,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,1536,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,65536,0.06448800033993192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,12288,0.03253511256641812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,1024,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,7168,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,3584,0.03286755416128371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,7168,0.5648915502760146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,4096,0.011365333365069495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,4096,0.01256444470749961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,8192,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,51200,0.1701822280883789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,12288,0.18715200159284803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,12288,0.4262559943728977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,4096,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,3584,0.19438933001624215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,12288,0.031023999055226643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,64,0.005979555762476391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,2560,0.06602488623725043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,2560,0.00489155575633049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,768,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,5120,0.042615112331178456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,1536,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,3072,0.06902133093939887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,128,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,3584,0.050922665331098765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,7168,0.339224894841512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,12288,0.026744888888465032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,7168,0.012049777640236748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,128,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,8192,0.09643822246127659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,51200,0.011674666570292579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,6144,0.032299555010265775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,128,0.00960533320903778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,65536,1.6950551138983831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,8192,0.03974577784538269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,1536,0.0864844454659356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,8192,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,1024,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,1024,0.009932444327407414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,3072,0.055734223789638944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,128,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,10240,0.09324355257882012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,3584,0.024123556084103052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,16384,0.4139271047380235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,16384,0.017622222503026325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,512,0.013772444592581855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,5120,51200,0.5838133494059244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,16384,0.021033777130974665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,768,0.009010666774378883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,16384,2.0376773410373263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,5120,0.02475111186504364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,3072,0.07032444741990831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,2048,0.15539288520812988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,5120,0.06217688984341092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,64,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,1024,0.02019644445843167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,51200,0.12359821796417236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,5120,0.017820444371965196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,256,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,2560,0.04044888748062982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,512,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,3584,0.05480622251828512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,32,0.04792444573508369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,3584,0.007679111427730984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,128,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,2560,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,2048,0.11087110969755386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,1024,0.017829333742459614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,3072,0.017504000001483493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,8192,0.016102222932709586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,256,0.01997777819633484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,12288,0.18522400326199004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,8192,0.062090668413374156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,4096,0.015821332732836407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,6144,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,8192,0.018025777406162687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,16384,0.025751999682850305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,4096,0.025843555728594463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,512,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,3584,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,12288,0.01645511057641771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,8192,0.03717155588997735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,2048,0.011552000211344825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,2048,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,12288,0.03190044562021891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,2560,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,8192,0.07890222469965617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,65536,0.2257004313998752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,768,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,10240,0.015215999550289579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,8192,0.0821395582622952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,65536,0.1814275582631429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,128,0.01551999979548984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,512,0.005383111122581694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,5120,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,8192,0.018132444885041978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,256,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,51200,1.747102207607693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,768,0.05354488889376322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,10240,0.013080889152155982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,12288,0.20675199561648896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,512,0.006219555520349079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,6144,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,2560,0.18650133079952666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,12288,0.17805866400400797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,768,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,1536,0.03436177637841966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,16384,0.043863998519049756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,2048,0.028823110792371962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,2560,0.01461955573823717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,3584,0.03601510988341437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,2048,0.01384888920519087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,5120,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,51200,0.11280977725982666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,65536,0.634875562455919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,10240,0.233604457643297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,5120,0.02160533269246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,7168,0.04319644305441114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,512,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,3072,0.010279999838935005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,8192,0.8899129231770834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,6144,0.035841776265038386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,512,0.10379555490281846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,10240,0.019536889261669584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,1536,0.05647110939025879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,3072,0.01946311030122969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,768,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,1536,0.016920000314712524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,10240,0.08733066585328843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,3072,0.1276879972881741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,1024,0.013796444568369122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,8192,0.06441599792904325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,5120,0.017304889029926725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,12288,0.12622666358947754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,3584,0.052164445320765175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,2560,0.01645866698688931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,12288,0.014097778333557976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,4096,0.19627111487918428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,2048,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,10240,0.08328177531560262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,2048,0.005520888914664586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,51200,0.07169244686762492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,64,0.0018008889423476325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,128,0.03144355614980062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,6144,0.01737155516942342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,4096,0.01664444473054674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,32,0.005422222117582957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,6144,0.4133102099100749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,12288,0.016022221909628976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,256,0.0026524443593290118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,512,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,4096,0.020112888680564035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,256,0.016137777103318107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,3072,0.048511998520957104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,6144,0.030488888422648113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,51200,1.2702257368299696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,10240,0.028734223710166082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,64,0.007505777809354994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,256,0.004208000169859992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,51200,0.11601066589355469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,5120,0.01647555496957567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,7168,0.024892444411913555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,16384,0.023983110984166462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,16384,0.03428444597456191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,512,0.014141332772043018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,51200,0.8937991460164388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,1024,0.00776533368561003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,7168,0.04218577676349216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,6144,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,1024,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,768,0.0185253338681327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,6144,0.05492622322506375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,16384,0.4150373405880398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,2560,0.15928533342149523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,2048,0.016549333930015564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,128,0.0052577778697013855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,4096,0.023392000132136877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,7168,0.1755573352177938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,3584,0.02990844514634874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,65536,0.2126071055730184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,65536,0.05821688969930013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,65536,0.5293253262837728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,5120,0.051664001411861844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,256,3072,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,128,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,4096,0.012220444778601328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,64,0.0299893319606781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,4096,0.017831999394628737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,10240,0.16031821568806967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,6144,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,256,0.019343111250135634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,128,0.01107022249036365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,8192,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,65536,0.07144800159666273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,1024,0.006671999891599019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,10240,0.023397333092159692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,16384,0.26247111956278485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,12288,0.015299555328157214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,512,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,3072,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,1024,0.04818133181995816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,768,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,10240,0.1732382244533963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,64,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,51200,4096,0.6037315792507595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,12288,0.05052977800369263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,8192,0.04450666573312548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,5120,0.008406221866607666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,8192,0.050273779365751475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,8192,0.0351875556839837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,4096,3072,0.017133333616786532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,6144,0.045102222098244556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,1536,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,51200,4.105352189805773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,65536,1.50091552734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,3584,0.10735288593504165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,768,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,12288,0.035368889570236206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,2048,0.019873778025309246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,7168,0.32348622216118705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,32,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,5120,0.03762844536039565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,6144,0.010899555351999072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,8192,0.018387556076049805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,2560,0.03167466653717889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,1536,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,5120,0.037272890408833824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,5120,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,3072,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,768,0.00870488915178511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,4096,0.016314667132165697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,1024,0.016233777006467182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,7168,0.09640444649590386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,1536,0.013366222381591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,768,0.0026684444811609057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,2560,0.018162666095627677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,3584,0.05216799841986763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,12288,0.0435804459783766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,3072,0.06511822011735705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,1536,0.02830933199988471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,32,0.0083155557513237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,16384,2.422236548529731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,8192,0.07786399788326688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,1024,0.02199555602338579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,8192,0.2553306685553657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,256,0.026735110415352717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,512,0.0031475555151700974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,10240,0.012755555411179861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,12288,0.04831377665201823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,2048,0.018704000446531508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,2560,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,768,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,12288,0.028892444239722356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,512,0.005671111245950063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,51200,0.037971556186676025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,1024,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,128,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,10240,0.23232444127400717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,1536,0.04463111029730903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,7168,0.09630222453011407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,2560,0.07136711147096422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,7168,0.01033688916100396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,64,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,3072,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,2048,0.016622222132152982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,5120,0.014226666755146451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,1536,0.0064791109826829694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,7168,0.039332442813449435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,5120,0.05264444483651055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,256,0.010590222146775989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,8192,0.014922666880819531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,51200,0.05316799879074097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,64,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,1536,0.025919111238585577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,4096,0.0836648874812656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,12288,0.05085066623157925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,5120,0.0083155557513237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,1536,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,2560,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,1536,0.0328800015979343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,768,0.008956444760163626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,768,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,51200,0.13365333610110813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,3584,0.008170666793982187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,6144,0.11247289180755615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,512,0.02199111051029629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,12288,0.012711111042234631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,256,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,3584,0.26646044519212514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,768,0.004250666747490565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,1024,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,3584,0.010819555156760745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,3072,0.037654221057891846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,12288,0.04448444313473172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,12288,0.04881333311398824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,51200,0.5179742177327474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,768,0.028039998478359644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,2048,0.010255110760529837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,10240,0.016513778103722464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,256,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,768,0.023571555813153584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,256,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,7168,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,7168,0.012206222448084088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,128,0.020987555384635925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,64,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,10240,0.04126666651831733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,1536,0.02869422237078349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,65536,12288,0.18832622634039986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,2048,0.09926311175028484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,12288,0.015278221832381355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,7168,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,768,0.010989333192507425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,7168,0.03936444388495551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,768,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,10240,0.04228711128234863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,8192,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,128,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,12288,0.04313777883847555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,8192,0.04102755586306254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,256,0.017519111434618633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,2048,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,2048,0.019338667392730713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,2560,0.045759111642837524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,65536,0.06126399834950765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,5120,0.03143289022975498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,7168,0.012487110992272695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,2048,0.030693332354227703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,1024,0.013731555806265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,3584,0.008365333080291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,16384,0.04330755604637993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,8192,0.09942222303814358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,51200,0.11587288644578721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,128,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,8192,0.033859554264280535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,1024,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,1024,0.01346577786737018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,12288,0.013148444394270578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,12288,0.004893333133724001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,768,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,768,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,8192,0.049606220589743726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,128,0.006200888918505774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,51200,0.2203662263022529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,51200,0.49516089757283527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,51200,0.8037555482652453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,65536,0.06974311007393731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,256,0.013915555344687568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,3072,0.03849955399831136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,65536,0.061234666241539844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,256,0.010277333358923594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,65536,0.33463554912143284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,1024,0.003343111111058129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,2048,0.018256889449225534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,256,0.012832000023788877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,512,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,256,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,51200,0.10070133209228516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,5120,0.018174222773975797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,2560,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,128,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,7168,0.022446221775478784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,7168,0.023947555157873366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,3072,0.4140924347771539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,6144,0.04432266619470385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,512,0.01221866657336553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,2560,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,4096,0.019634667370054457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,16384,0.017589333984586928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,768,0.032955557107925415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,10240,0.1642097764545017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,51200,0.946422259012858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,2048,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,8192,0.024775111012988623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,6144,0.03622577918900384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,65536,0.2693839867909749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,51200,0.047263112333085805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,10240,0.03014577759636773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,1536,0.15209955639309353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,32,0.00291377781993813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,51200,0.049560887946022876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,51200,0.037639111280441284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,2560,0.007721777591440413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,12288,0.028485334581798975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,2048,0.013863110707865821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,8192,512,0.035465776920318604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,10240,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,2560,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,65536,0.24306045638190377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,3072,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,2048,0.01644888851377699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,5120,0.02394666605525547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,4096,0.01073066641887029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,7168,1.8103129069010417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,7168,0.009905777871608734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,1024,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,2048,0.04623466730117798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,64,0.008269333177142674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,51200,0.0502106679810418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,7168,0.03553244471549988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,5120,0.038718223571777344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,1024,0.05371910995907254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,8192,0.02924799919128418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,128,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,16384,0.19984000258975557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,768,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,10240,0.24853600396050346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,2048,0.004512000001139111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,64,0.04354044463899401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,12288,0.036658667855792575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,1024,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,3584,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,8192,0.043042666382259794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,3584,0.02365422248840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,3072,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,51200,0.11560444037119548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,8192,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,1536,0.02824444572130839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,1536,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,2560,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,3584,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,1024,0.011906666888131036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,12288,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,16384,0.19075822830200195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,16384,0.14219378100501165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,32,0.014392889208263822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,4096,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,16384,0.07023555702633329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,64,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,16384,0.018252443936136033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,4096,0.016410667035314772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,12288,0.04784799946679009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,512,0.012700444294346703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,51200,0.05393955442640516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,512,0.030339555607901678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,4096,0.028510222832361858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,128,0.0032142222755485107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,12288,0.11699377165900336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,512,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,65536,0.37342132462395566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,5120,0.05661333269543118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,51200,0.23377866215176055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,768,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,128,0.023556444380018447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,12288,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,51200,0.049878223074807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,32,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,3072,0.01885066595342424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,16384,0.03413155674934387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,128,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,2560,0.011672000090281168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,3584,0.03357244531313578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,16384,0.24785423278808594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,3072,0.03188622328970168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,128,0.011704000333944956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,768,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,1024,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,4096,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,1536,0.02865333358446757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,1536,0.003452444449067116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,8192,0.2907235622406006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,256,0.007240888973077138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,65536,0.14864444732666016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,1024,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,3584,0.029532445801628962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,4096,0.018950222267044913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,1024,0.014101333088344999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,6144,0.047579556703567505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,4096,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,128,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,3584,0.009815110928482479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,128,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,8192,0.015998222761683993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,3584,0.014892444014549255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,10240,0.02703022294574314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,3072,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,256,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,64,0.016826666063732572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,128,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,256,0.010105777945783403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,2048,0.26081956757439506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,16384,0.06779822376039293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,65536,0.09780622190899319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,2048,0.02735377848148346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,65536,1.0220746994018555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,1536,0.02312444481584761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,3072,0.01346933344999949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,1024,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,65536,1.3817947175767686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,8192,1.3706435097588434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,128,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,51200,0.20879111025068495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,65536,0.042544888125525586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,3072,0.015138667490747241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,1536,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,2048,0.01034755590889189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,51200,0.6524426672193739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,65536,0.15161421563890246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,512,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,6144,0.2051262193255954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,6144,0.017596445149845548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,64,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,1536,0.007242666350470648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,3584,0.17280533578660753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,3072,0.10746755864885117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,4096,0.024334222078323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,3072,0.07489599784215291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,256,0.006670222100284364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,3584,0.023010666171709698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,3072,0.03862399856249491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,7168,0.008440000315507254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,256,0.012832889126406776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,1024,0.030622220701641504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,2048,0.12636533048417833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,65536,0.09572444359461467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,12288,0.031209776798884075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,5120,0.029713776376512315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,7168,0.030320889419979517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,1024,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,2048,0.0472942226462894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,10240,0.011755555868148804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,32,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,512,0.006199999815887875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,128,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,10240,0.21528177791171602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,7168,0.013155555559529198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,16384,0.03806844353675842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,10240,0.2108551131354438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,512,0.0162808895111084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,10240,0.20985066890716553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,5120,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,3072,0.08156800270080566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,1536,0.02828888760672675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,12288,0.03910488883654276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,768,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,4096,0.051719999975628324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,12288,0.031762666172451444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,768,0.008222222328186035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,1536,0.05801244576772054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,12288,0.10038489103317261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,10240,2.124433729383681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,2048,0.028864890336990356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,768,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,5120,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,1024,0.06159733401404487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,8192,0.010949333508809408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,1536,0.009920888476901585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,64,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,2560,0.028933333026038274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,10240,0.08518133560816447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,1024,0.012090666426552666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,6144,0.007691555553012424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,12288,2.800274742974175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,8192,0.033018665181265935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,5120,2560,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,768,0.012721777790122561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,1024,0.036328001154793635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,128,0.03359466791152954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,3072,0.010505778094132742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,64,0.012425777812798819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,6144,0.009379555781682333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,8192,0.025779555241266888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,512,0.0030577776746617425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,3584,0.008061333662933773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,3072,0.546013355255127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,6144,0.04705511199103462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,65536,0.06817155414157443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,768,0.02863911125395033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,512,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,256,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,1024,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,8192,0.029567999972237483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,1536,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,16384,0.12320533063676621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,10240,0.19041866726345488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,6144,0.06183377901713053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,128,0.009383111364311641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,3584,0.00925155563486947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,2048,0.010774222513039907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,3072,0.4030151102277968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,2560,0.10715022352006699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,16384,0.05584177705976698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,4096,0.031593776411480375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,2048,0.031868444548712835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,8192,0.8870284292432996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,3584,0.10513599713643391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,16384,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,2048,0.021922666165563796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,64,0.015301333533393012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,5120,0.030210667186313208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,1024,0.01885066595342424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,5120,0.049801776806513466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,768,0.008365333080291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,1536,0.016221332881185744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,10240,0.016016888949606154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,768,0.021002666817771062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,65536,0.12656266159481472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,5120,0.012326222327020435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,512,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,10240,0.019557333654827543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,3072,0.008378666308191087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,65536,0.07491555478837755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,4096,0.053118222289615206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,5120,0.017909333109855652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,1024,0.016928889685206942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,2048,0.027078222897317674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,768,0.006305777778228124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,16384,0.03333955672052171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,1024,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,65536,0.1319626702202691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,256,0.029122667180167303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,64,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,128,0.5331439971923828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,6144,0.09214844306310017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,10240,0.01577333278126187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,7168,0.056997332308027476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,3072,0.026001777913835313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,8192,0.012449777788586087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,768,0.012635555532243518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,768,0.018378666705555387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,4096,0.010105777945783403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,512,0.004554666578769684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,8192,0.09360622035132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,3584,0.024950222836600408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,16384,0.00812888890504837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,2048,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,2560,0.03730488816897074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,6144,0.06396977769003974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,7168,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,128,0.002510222295920054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,1536,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,768,0.012469333079126147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,2560,0.006812444577614467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,2048,0.015949333707491558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,768,0.02695377833313412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,512,0.07431555456585355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,16384,0.03971822063128153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,3072,0.6546328862508138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,8192,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,2560,0.04858577913708157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,512,0.003980444537268745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,512,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,7168,0.04810933272043864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,2560,0.026522666215896606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,12288,0.4977502293056912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,10240,0.22765421867370605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,1536,0.010392000277837118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,8192,0.04409155580732557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,768,0.017874666386180453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,128,0.004226666771703296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,10240,0.028232001596026953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,8192,0.04971288972430759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,1024,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,12288,0.0162773331006368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,65536,0.46368000242445206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,7168,0.049096888966030545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,2560,0.0823777781592475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,16384,0.5133191214667426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,8192,1.6491226620144313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,51200,51200,4.053177727593316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,1024,0.0196079992585712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,512,0.003867555409669876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,2560,0.011649777491887411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,2560,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,256,0.010394666757848527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,2560,0.04797866609361437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,256,0.026735999517970618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,128,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,32,0.005019555489222209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,6144,0.02574311031235589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,10240,0.6837075551350912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,256,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,10240,0.04346311092376709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,768,0.026271111435360376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,1024,0.020741333564122517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,65536,11.723346286349825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,6144,0.026874666412671406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,6144,0.024596444434589807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,768,1.5305724673801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,1536,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,4096,0.011130666567219628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,1536,0.0525857773092058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,768,0.04811288913091024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,3072,0.03843466771973504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,2048,0.1078337762090895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,3072,0.025205334027608235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,64,0.008438222110271454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,65536,0.07770488659540813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,256,0.021351999706692163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,4096,0.014410666293568082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,8192,0.03568889035118951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,64,0.010289777484205034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,8192,0.04143288731575012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,2560,0.01551644504070282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,4096,0.019038221902317472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,7168,0.0797591143184238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,6144,0.09086133374108209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,2048,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,51200,0.35594577259487575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,8192,0.013974222871992322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,1024,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,5120,0.11506666077507867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,10240,0.11708266205257839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,512,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,1024,0.019688000281651814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,10240,2.336338678995768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,16384,0.15675022866990831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,1024,0.01016266644001007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,64,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,10240,0.020218667056825426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,512,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,2560,0.01051644484202067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,3072,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,3072,0.007522666619883643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,10240,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,5120,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,65536,0.20683644877539742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,8192,0.037583112716674805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,4096,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,7168,0.12180177370707194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,32,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,256,0.00959644466638565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,3072,0.012799111505349478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,768,0.005968889014588461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,4096,0.0355004436439938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,3584,0.008741333252853816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,1536,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,768,0.005085333353943295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,256,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,5120,0.32502489619784886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,12288,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,256,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,1536,0.006331555545330048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,3072,0.032913777563307024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,16384,0.5262453291151259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,4096,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,128,0.0019271110908852683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,1536,0.01383733335468504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,10240,12288,0.04525688952869839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,7168,0.013166222307417126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,128,0.012493333054913415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,3072,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,1024,0.0897573365105523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,7168,0.044587555858823985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,4096,0.010411555568377176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,512,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,2048,0.017581333716710407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,51200,0.05156266689300537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,2048,0.0075786668393347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,5120,0.030488000975714788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,768,0.01869422197341919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,64,0.002480000050531493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,7168,0.9880275726318359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,51200,0.06881066825654772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,3072,0.06447377469804552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,7168,0.7969013320075141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,10240,0.1025235586696201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,128,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,2560,0.03294310967127482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,768,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,7168,0.013271110753218332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,8192,0.015863999724388123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,1024,0.028155555327733357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,6144,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,5120,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,8192,0.030810667408837214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,6144,0.07064533233642578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,16384,0.036561777194341026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,3072,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,12288,0.015205333630243937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,6144,0.011794666449228922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,64,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,10240,0.23042665587531197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,2560,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,51200,0.22488623195224336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,768,0.004826666580306159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,128,0.014939554863505893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,16384,0.03688177797529433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,5120,0.22944890128241646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,10240,0.019659555620617337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,6144,0.01517955627706316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,2048,0.0243493335114585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,6144,0.022728888524903193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,512,0.028416888581381902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,512,0.025900445050663416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,512,0.0028444443725877335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,1024,0.03641066617435879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,256,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,51200,2.5391199323866105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,768,0.1895600027508206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,51200,0.051260444853040904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,6144,0.9746959474351672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,7168,0.01350488927629259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,10240,0.025076443950335186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,2560,0.012095111111799875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,65536,0.025102222959200542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,1024,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,3584,0.02168088820245531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,2048,0.15381156073676214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,8192,0.04019022319051955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,16384,0.016573333077960543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,51200,0.0603297750155131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,51200,0.05750222338570488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,32,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,16384,0.8744124306572808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,12288,0.12071377701229519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,64,0.007471111085679796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,3072,0.008828444613350762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,6144,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,10240,1.8554000854492188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,2560,3.2479608323838978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,128,0.02162488963868883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,6144,1.5202631420559352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,7168,0.1389546659257677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,128,0.04827910992834303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,4096,0.009408888717492422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,7168,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,128,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,3584,0.03187733226352268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,7168,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,3584,0.06846044460932414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,1024,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,768,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,768,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,65536,5.0602976481119795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,12288,0.04811911119355095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,32,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,2560,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,512,0.021312000022994146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,768,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,1024,0.026719111535284255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,2560,0.019573332534896005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,16384,0.02554399934079912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,5120,0.020260444945759244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,64,0.0018213333355055915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,2048,0.004387555436955558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,1536,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,1536,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,1536,0.0262080000506507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,2048,0.010415111151006486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,6144,2.362330754597982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,2048,0.06991733445061578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,2048,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,65536,0.4652782016330295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,12288,0.03256888853179084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,6144,0.11011822356118096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,65536,0.12793955538007948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,768,0.014717333846622042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,10240,0.013384888569513956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,768,0.01923377811908722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,128,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,12288,0.04182844360669454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,51200,2.3374000125461154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,16384,0.13691111405690512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,8192,0.01759377784199185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,1024,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,512,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,3584,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,1536,0.025398222936524287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,768,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,128,0.009485333330101436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,4096,0.0131217779384719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,3584,0.01657777859105004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,1536,0.025219556358125474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,51200,0.05352444450060526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,1536,0.03732444511519538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,10240,0.012124444047609964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,128,768,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,2048,0.040602667464150324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,1024,0.0041368889311949415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,4096,0.030250665214326646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,256,0.018202667435010273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,3584,0.011683555940786997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,5120,0.043486220969094165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,256,0.022112000319692824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,256,0.010601777997281818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,1024,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,6144,0.9799270629882812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,32,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,2048,0.013750221994188098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,7168,0.04738044407632616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,2048,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,768,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,12288,0.10377333561579387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,1024,0.055870221720801465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,6144,0.00815466625822915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,5120,0.008349333372380998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,1024,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,7168,0.021592888567182753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,5120,0.08884533246358235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,768,0.016968000266287062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,4096,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,32,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,16384,0.021165332860416834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,4096,0.007376000285148621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,128,0.008214222060309516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,256,0.005386666705211003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,10240,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,3584,0.010013333625263637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,2560,0.009139555195967356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,2560,0.012257777982287936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,1536,0.09650933080249363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,8192,0.01625777781009674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,512,0.05124355687035454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,512,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,2560,0.007239110767841339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,51200,0.05268088976542155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,1536,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,128,0.021721776988771226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,512,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,2560,0.00903200027015474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,256,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,51200,0.20785510540008545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,2560,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,128,0.015562666787041558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,5120,0.058033777607811816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,51200,0.2586497730678982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,32,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,64,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,768,0.06557866599824694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,512,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,7168,0.010120000276300643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,256,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,4096,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,4096,0.09753511349360149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,3584,0.03108711043993632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,2048,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,12288,0.051923556460274585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,128,0.006500444478458828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,16384,0.4723822275797526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,4096,0.03302933441268073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,16384,0.01585155559910668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,2048,0.11261778407626682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,7168,2.282813390096029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,16384,0.029670221938027277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,16384,2.8829263051350913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,8192,0.4726568857828776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,4096,0.046520888805389404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,2048,0.007679999702506595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,10240,0.20703466733296713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,7168,0.03581422236230638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,6144,0.0812551114294264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,3072,0.02666577696800232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,1536,0.021253334151373968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,5120,0.039080000585979886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,2048,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,4096,0.07579200135336982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,2048,0.011031111081441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,8192,0.012990222209029727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,2560,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,4096,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,256,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,128,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,3584,0.32544978459676105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,2048,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,5120,0.01076533314254549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,1024,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,1024,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,128,0.07842755317687988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,16384,0.022657778528001573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,65536,0.0868942207760281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,1024,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,10240,0.014714666538768344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,2048,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,1024,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,256,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,256,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,3072,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,256,0.005218666460778978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,3584,0.022302221920755174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,3072,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,3072,0.027388445205158655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,768,0.006215999937719769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,6144,0.012748444245921241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,512,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,128,0.005836444596449534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,128,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,16384,0.7727466689215766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,2048,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,4096,0.13115999433729383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,2560,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,2048,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,51200,0.08411288923687404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,1536,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,65536,128,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,1536,0.025588444537586633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,64,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,65536,0.1379964484108819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,4096,0.011339555184046427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,10240,0.06610488891601562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,51200,0.05202933483653598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,5120,0.03848888807826572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,512,0.002534222271707323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,2048,0.025388444463411968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,768,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,128,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,3072,0.03559111224280463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,12288,0.23147998915778267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,2048,0.19916533099280465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,3072,0.028005331754684448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,12288,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,4096,0.02360533343421088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,8192,0.04594844579696655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,10240,0.9535520341661242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,128,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,1024,0.01460088955031501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,128,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,3072,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,256,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,5120,0.012776888906955719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,768,0.029449777470694646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,768,0.014771555860837301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,2560,0.027161777019500732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,128,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,512,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,65536,1.675797356499566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,256,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,3584,0.01075111081202825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,3584,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,32,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,8192,0.05699555741416084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,12288,0.037588444021013044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,8192,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,2048,0.031268444326188825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,256,0.0174026671383116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,65536,0.012048888537618848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,8192,10240,0.0776968863275316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,2048,0.018292443619834054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,768,0.22614667150709364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,3584,0.020679111282030743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,5120,0.036130666732788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,2560,0.006533333410819371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,768,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,128,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,768,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,2048,0.023452444208992854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,128,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,5120,0.03914755582809448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,12288,0.050548444191614784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,1024,0.01349955548842748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,12288,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,65536,2.975456025865343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,65536,0.09204710854424371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,1536,0.010459555519951714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,5120,0.03334044416745504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,4096,0.019574221637513902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,4096,0.020412445068359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,8192,0.18704444832272002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,768,0.22059999571906197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,768,0.00942844483587477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,1024,0.010305778019958073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,512,5120,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,256,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,32,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,10240,0.06296800242529975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,5120,0.0597626699341668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,5120,0.018290667070282828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,10240,0.0473599996831682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,768,0.01829866733815935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,7168,0.019883554842736986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,128,0.09296888775295681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,12288,0.24317065874735513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,768,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,64,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,12288,0.03440622157520718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,4096,0.08606221940782334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,65536,0.12729510996076795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,768,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,32,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,768,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,65536,1.2265991634792752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,51200,0.18271556165483263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,64,0.009806222385830348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,512,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,2560,1.1902027130126953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,7168,0.020294222566816542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,65536,3072,0.18167289098103842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,512,0.007751999629868402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,512,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,16384,0.27794310781690806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,512,0.06217422087987264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,51200,135.77779134114584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,256,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,3072,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,10240,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,2048,0.01997511088848114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,65536,0.0914613339636061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,512,0.0866960022184584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,4096,0.012677333421177335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,5120,0.016141333513789706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,4096,5120,0.16102222601572672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,2560,0.011729777687125735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,512,0.010423111418883005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1024,256,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,128,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,12288,0.08790310886171128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,6144,0.029817776547537908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,256,0.005883555445406172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,4096,0.12427288956112331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,128,0.001963555605875121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,3584,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,51200,2.8734825981987844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,512,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,10240,0.5716817643907335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,4096,0.0961937771903144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,256,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,128,0.05622400177849663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,3584,0.047998223039839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,51200,0.04502933224042257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,32,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,3584,0.009626666704813639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,8192,0.025779555241266888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,51200,1.1255662706163194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,16384,0.029137776957617864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,12288,0.02047644390000237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,64,0.02273955610063341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,7168,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,64,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,10240,0.07179111242294312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,768,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,32,0.03234577841228909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,2048,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,768,0.01627200014061398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,3072,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,128,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,1024,0.02533333334657881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,8192,0.014349333114094205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,6144,0.018903111418088276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,12288,0.06156533294253879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,2048,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,3072,0.07038933038711548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,16384,0.017303999927308824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,51200,0.15991555319892034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,6144,0.011148444480366178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,8192,0.0825128886434767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,2560,0.009000889129108852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,256,0.022759111391173467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,65536,0.010703999963071613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,5120,0.15451378292507595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,5120,0.07851733101738824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,8192,0.016255999604860943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,256,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,1024,0.008004444340864817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,16384,0.13576000266604954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,4096,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,51200,0.2680444452497694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,4096,0.09061421950658162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,8192,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,2048,0.049977779388427734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,3584,0.01740622189309862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,7168,0.011576889289749993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,4096,0.012703110774358114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,2560,0.016726222303178575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,12288,0.01736177835199568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,1024,0.021756443712446425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,1536,0.057012445396847196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,1536,0.06790044572618273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,65536,0.061623109711541064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,8192,0.32314933670891655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,4096,0.11423999733395046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,65536,0.27415643797980416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,65536,0.5833537843492296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,2560,0.012273777690198688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,16384,0.13387110498216417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,6144,0.12149777677324082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,2560,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,7168,0.02363555630048116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,3072,0.03689510954750909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,512,0.2624453438652886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,5120,0.008947555389669206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,512,0.0032506665835777917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,51200,0.0947386688656277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,65536,0.8140293227301704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,64,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,512,0.025123554799291823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,51200,2048,0.21265599462721083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,10240,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,1024,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,64,0.020284444093704224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,12288,0.014358222484588623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,768,0.07267022132873535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,10240,0.0447306666109297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,5120,0.1343911091486613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,32,0.01888533267709944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,5120,0.015319999721315173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,2048,0.0038737778862317405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,2048,0.0377493335141076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,512,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,512,0.021896888812383015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,2048,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,12288,0.013112000293201871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,16384,0.039641777674357094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2560,128,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,2560,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,2560,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,32,0.002757333426011933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,16384,0.03944799966282315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,4096,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,1024,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,6144,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,512,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,128,6144,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,1536,0.016306666864289176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,4096,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,4096,0.04524888926082187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,128,0.039649777942233615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,7168,0.031926221317715116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,3584,0.022456889351209004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,5120,0.014056000444624158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,32,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,768,0.1755004458957248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,4096,0.023616888456874426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,65536,0.061942219734191895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,256,0.01517955627706316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,16384,1.2642409006754558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,51200,0.07086133294635348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,5120,0.008540444076061249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,128,0.03130133284462823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,2048,0.08919999996821086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,128,0.012631999949614206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,768,0.015511999527613321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,3584,0.12743200196160212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,2560,0.04144800040456984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,8192,0.027448889282014634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,10240,0.05923022164238823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,32,0.0444471107588874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,7168,0.26647644572787815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,8192,0.09572088718414307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,32,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,10240,0.020015110572179157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,8192,0.009023110899660323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,1536,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,2048,0.007136888802051544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,12288,0.018976888722843595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,12288,0.015708444847000968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,8192,0.052638222773869835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,4096,0.05441955394215054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,6144,0.10537422365612453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,3584,0.0222080002228419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,128,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,8192,1.3434018029106989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,256,0.008965333302815756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,32,0.014090667168299357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,4096,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,51200,0.07937510808308919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,512,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,2048,0.46235911051432294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,2560,0.04427999920315213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,3584,0.0651715530289544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,12288,0.1928373310301039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,8192,0.08183022340138753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,1536,0.018485332528750103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,3072,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,5120,0.04038044479158189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,12288,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,1536,0.022163555026054382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,4096,3072,0.008893333375453949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,10240,0.01947022146648831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,256,0.08449866374333699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,1536,0.279167996512519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,51200,5.780185699462891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,512,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,8192,0.05589422252443102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,3584,0.004264889078007804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,51200,0.3718622260623508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,1536,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,6144,0.16235110494825575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,51200,0.7118559943305122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,1536,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,4096,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,51200,0.20993155903286406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,8192,0.018756444255510967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,2048,0.04893689023123848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,32,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,1536,0.06059466467963325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,1536,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,8192,0.1205022202597724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,768,0.026404443714353774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,1536,0.00620977787507905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,65536,0.07490933603710599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,65536,0.08708266417185466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,16384,0.026752889156341553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,65536,0.0633102191819085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,512,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,3584,0.008066666622956594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,51200,0.7085999912685819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,16384,0.04243910974926419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,8192,0.011644444531864591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,768,0.009706666900051964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,5120,0.030123554997973975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,3584,0.08767200178570217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,6144,0.02808177802297804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,3584,0.03620355659061008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,1024,0.018604444132910836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,6144,512,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,1024,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,12288,0.4947688844468858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,12288,0.0233288887474272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,3584,0.026903111073705886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,3584,0.031341334184010826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,2048,0.023735110958417256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,65536,0.016798221402698092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,6144,0.010041777458455827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,6144,0.03849600089920892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,1024,0.12677866882748073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,6144,0.01165244479974111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,51200,0.07006044520272149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,2048,0.01127022256453832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,10240,0.06937244203355578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,12288,0.008728000024954477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,512,0.012301333248615265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,3072,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,768,0.018918222851223417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,16384,0.36100443204243976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,51200,0.04859111044141981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,51200,3.2093368106418185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,3584,0.026723555392689172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,512,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,5120,0.01907111042075687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,51200,2.4758497873942056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,10240,0.03836266530884637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,10240,0.10059644116295709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,2048,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,3584,0.9553306367662219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,16384,0.05004622207747566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,768,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,256,0.010091555615266165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,8192,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,512,0.00406666679514779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,2048,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,3584,0.020289777053727042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,1024,0.003145777723855443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,512,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,256,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,128,0.006542222367392645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,3584,0.023746665981080797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,12288,0.004850666556093427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,4096,0.033377776543299355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,51200,0.17871910995907256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,768,0.019691555036438834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,65536,0.3236453268263075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,8192,0.7273066838582357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,1536,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,2560,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,4096,0.020614221692085266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,4096,0.040390223264694214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,1536,0.01833511061138577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,1536,0.04829688866933187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,4096,0.01684266659948561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,10240,0.018163555198245578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,5120,0.35773687892489964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,768,0.18310311105516222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,8192,0.01517599986659156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,3072,0.007521777517265744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,6144,0.32317688730027944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,256,0.015503111812803479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,5120,0.22189422448476157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,3584,0.01824355622132619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,32,0.002557333351837264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,12288,0.017293334007263184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,768,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,16384,0.10990044143464829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,1024,0.013591999808947245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,3584,0.016351110405392118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,2048,0.012464000119103326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,512,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,1536,0.029147555430730183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,2560,0.01638488968213399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,512,0.019652444455358718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,51200,0.04859288864665561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,7168,0.018020444446139865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,1536,0.03253600001335144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,64,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,128,0.0026435556097163092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,1024,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,7168,0.020571556356218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,12288,0.08423644304275513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,51200,2.9542407989501953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,768,0.020120888948440552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,256,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,1536,0.005980444451173146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,512,0.003160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,4096,0.01974933346112569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,1536,0.009209777745935652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,2560,0.03170488940344916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,512,0.0614302224583096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,16384,0.08589066399468316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,7168,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,128,0.033562666840023465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,2560,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,8192,0.01461777753300137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,8192,0.014319999350441827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,32,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,10240,0.04993066522810194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,16384,0.0199617776605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,5120,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,128,0.020229332976871066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,1024,0.004590222405062782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,3072,0.07119288709428576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,8192,512,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,3072,0.031982223192850746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,12288,0.017872888180944655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,10240,0.2875422106848823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,6144,0.09635733233557807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,128,0.008354666332403818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,768,0.019519999623298645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,1024,0.024734222226672705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,7168,0.17737689283159044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,128,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,10240,0.09027733405431111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,768,0.12048266993628608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,512,0.0038684445122877755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,32,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,128,0.006000888844331105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,6144,0.013815999858909182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,2048,0.01615111033121745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,3584,0.007882666256692674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,10240,128,0.02439022229777442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,10240,0.015410666664441427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,256,0.021311110920376245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,4096,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,5120,0.02780888809098138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,12288,0.036666668123669095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,7168,0.011718221836619906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,3584,0.06220089064704048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,51200,2.8197494082980685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,768,0.01034222212102678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,1536,0.008597333398130205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,12288,0.020607999629444547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,10240,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,512,0.0032506665835777917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,3072,0.029167999823888142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,3584,0.015340444114473132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,1536,0.010955555571450127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,16384,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,12288,0.027867555618286133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,1536,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,10240,0.13189155525631374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,64,0.05387555559476217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,3584,0.04425422350565592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,64,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,7168,0.010710222025712332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,51200,0.14077511098649767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,128,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,2560,0.026202667090627883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,256,0.01667022208372752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,3072,0.19302845001220703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,4096,0.012128888732857175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,64,0.12090399530198838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,256,0.016362667083740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,256,0.05006577902370029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,3584,0.007816889219813878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,7168,0.10107377502653335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,4096,0.4421866734822591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,3072,0.015867556134859722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,16384,0.050556444459491305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,16384,0.12722222010294595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,2560,0.007117333511511485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,65536,0.13418844011094835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,4096,0.07032444741990831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,1536,0.01536177761024899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,8192,0.0439644455909729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,3584,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,128,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,7168,0.12959199481540257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,65536,0.07231110996670194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,4096,0.034704890516069196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,51200,0.12030933962927924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,256,0.01146933353609509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,8192,0.04696977800793118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,1024,0.014856888188256158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,4096,0.01239466667175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,5120,1.5597395367092555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,10240,0.01681066718366411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,5120,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,65536,0.06800355513890584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,16384,0.04324177900950114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,4096,0.01127644462717904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,128,0.031722668144438006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,7168,0.5313564406500922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,16384,0.01796355512407091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,51200,0.15967378351423475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,4096,0.019871999820073444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,64,0.00566133318675889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,7168,0.02659911082850562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,12288,0.07241511344909668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,6144,0.12457956208123101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,2560,0.047298666503694325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,5120,0.09071911043590969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,128,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,5120,0.013439111411571503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,12288,0.21478754944271514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,8192,1.0608675214979384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,1536,0.03576533330811395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,5120,0.010327999790509542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,16384,0.019903999235894945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,512,0.013433777623706393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,5120,0.027638221780459087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,65536,2.2743892669677734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,256,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,2560,0.03737955623202854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,2048,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,256,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,10240,0.014642667439248828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,256,0.014828445182906257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,8192,0.010686222049925061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,256,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,3584,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,12288,0.10230400164922078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,1024,0.017690666847758822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,1024,0.005764444255166584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,51200,0.8458177778455945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,3584,0.053595556153191455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,10240,0.20536977714962432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,7168,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,6144,0.06893955336676703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,65536,0.07828266753090753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,5120,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,768,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,128,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,768,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,256,0.0022133334229389825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,16384,0.04427822099791633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,512,0.006321777900060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,65536,0.17261866728464761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,1536,0.05555377735031975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,51200,0.06210133102205065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,6144,0.09671199984020656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,12288,0.37204978201124406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,32,0.025584888127115037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,6144,0.08144711123572455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,4096,0.015817777978049386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,12288,0.004858666823969947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,12288,0.05676266882154676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,16384,0.028833776712417603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,5120,0.1255342165629069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,3072,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,5120,0.022152000003390845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,1024,0.03977155685424805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,6144,0.018211555149820115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,4096,0.04148088892300924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,2048,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,12288,0.017864000466134813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,512,0.013808888693650564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,5120,0.440180434121026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,8192,0.03892888956599765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,4096,0.004253333227501975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,16384,0.02265600032276577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,3584,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,1536,0.03422844409942627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,768,0.08248533142937554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,7168,0.05326844586266411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,10240,0.090756442811754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,6144,51200,0.06179022126727634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,10240,0.025729777084456548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,65536,1.2949440214369032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,128,0.029784000582165186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,10240,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,10240,0.042502221133973866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,8192,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,12288,0.031545778115590416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,12288,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,64,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,16384,0.07752444346745808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,8192,0.0151235560576121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,3584,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,6144,0.03012088934580485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,65536,0.07015199793709649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,3584,0.055367112159729004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,3584,0.02534755567709605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,5120,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,16384,0.2141946686638726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,256,0.005419555637571547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,16384,0.025184889634450276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,2560,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,4096,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,5120,0.18202932675679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,1024,0.05363377928733826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,3072,0.2433262136247423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,3072,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,51200,0.950560039944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,10240,0.04182044333881802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,3072,0.06541511085298327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,10240,0.2278817759619819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,65536,0.0879475540584988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,512,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,128,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,4096,0.22143822246127656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,10240,0.1898533370759752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,512,0.06157244576348198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,256,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,768,0.020095111595259774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,16384,0.26325421863132054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,1536,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,8192,0.09691111246744792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,6144,0.01000533335738712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,8192,0.08964266379674275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,7168,0.051085333029429116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,512,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,6144,0.012832889126406776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,2560,0.07635288768344455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,51200,25.129587809244793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,768,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,7168,0.06825955708821614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,3584,0.06004799736870659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,65536,0.14338666862911648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,1024,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,16384,0.07797244522306654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,2048,0.007692444655630324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,64,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,64,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,65536,0.14788266023000082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,32,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,1024,0.028385778268178303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,51200,0.06474400228924222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,2560,0.0069422221018208405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,128,0.0019697778754764134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,256,0.01184977756606208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,32,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,3072,0.35769422849019367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,1024,0.00452888881166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,8192,0.03830133212937249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,256,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,51200,0.11738667223188613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,5120,0.01387644476360745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,3072,0.008207110895050896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,3072,0.025016888976097107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,1024,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,65536,0.1970097753736708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,5120,0.008966222405433655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,1024,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,6144,0.016209777858522203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,12288,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,65536,12288,0.1860168907377455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,10240,0.06566044357087877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,8192,0.05385155479113261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,2560,0.007350222104125553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,1536,0.02163200080394745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,51200,0.427257776260376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,3072,0.12056267261505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,6144,0.014402666025691561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,2048,0.016556445095274184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,16384,0.281351990169949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,512,0.006723555425802867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,12288,0.01685066686736213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,1536,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,1024,0.008931555681758458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,3584,0.009190222455395592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,6144,0.03528799944453769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,2048,0.02010844482315911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,512,0.012826667063766055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,32,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,32,0.005078222188684675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,2560,0.007320889168315464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,5120,0.013620444469981723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,16384,0.023743111226293776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,1536,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,1024,0.028509332074059382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,51200,0.32872533798217773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,1536,0.03826933436923557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,7168,0.02637600070900387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,6144,0.1116773287455241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,2560,0.007537778053018782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,768,0.22202844089931914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,10240,0.05510666635301378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,512,0.005659555395444234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,128,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,128,0.016534222496880423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,16384,0.11206755373213027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,7168,0.15540711085001627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,256,0.018546667363908555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,512,0.0024746666765875286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,5120,0.13351199362013075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,3584,0.017501332693629794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,64,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,768,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,768,0.0037928890022966596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,51200,0.13738755385080972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,256,0.029150221082899306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,4096,0.027111111415757075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,4096,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,8192,2.3913243611653647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,8192,0.07093689176771376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,5120,0.6793466673956977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,8192,0.03360266817940606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,1536,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,4096,0.023401778605249193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,1024,0.02313066687848833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,128,0.0632186664475335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,2048,0.02832533253563775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,2048,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,3072,0.05728089147143894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,65536,1.0314231448703342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,2560,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,10240,0.07499200105667114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,2048,0.08540799882676865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,256,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,32,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,3584,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,512,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,12288,0.1899226639005873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,256,0.02198666665289137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,5120,0.04095110959476895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,64,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,7168,0.009723555710580613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,2560,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,2048,0.022691556149058874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,12288,0.07552711168924968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,6144,0.0229768885506524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,4096,0.02717777755525377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,32,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,64,0.01909244391653273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,8192,0.011037333144081963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,2560,0.09482933415306939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,768,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,8192,0.04439999990993076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,2048,0.01822933389080895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,64,0.004513777792453766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,12288,0.2679102155897352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,65536,5.543502383761936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,3072,0.03453777896033393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,10240,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,16384,0.024041778511471216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,1024,0.003552000141806073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,7168,0.03832533293300205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,65536,0.02000799940692054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,1536,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,6144,0.009162666896979014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,32,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,10240,0.3128604359096951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,3584,0.030863109562132094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,5120,0.24380000432332358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,65536,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,3072,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,32,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,256,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,10240,0.05112533436881172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,51200,0.4269084400600857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,128,0.009062222308582729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,16384,0.06227111154132419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,256,0.13370133770836726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,51200,0.1306426657570733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,12288,0.01982222166326311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,65536,0.13374666372934976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,4096,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,2048,0.022340445054901972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,1536,0.02916533417171902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,3584,0.03440266516473558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,2048,0.2662035624186198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,64,0.09214755561616685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,768,0.0026568888376156488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,256,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,16384,0.054007109668519765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,768,0.01791288952032725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,1536,0.03927466604444716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,10240,0.08279644118414985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,5120,0.0439208878411187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,5120,0.008035555481910706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,6144,0.009342222577995723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,64,0.030369778474171955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,3072,0.012028444144460889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,6144,0.012350222302807702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,1536,0.006730666591061487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,5120,0.01850488947497474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,1024,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,3072,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,4096,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,2048,0.0203137778573566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,64,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,768,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,2048,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,5120,0.051046222448349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,5120,0.049104889233907066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,128,0.01279022213485506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,1536,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,64,0.007811555431948767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,3072,0.016196444630622864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,5120,0.007257777783605788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,2560,0.015108444624476962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,1536,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,768,0.01726222203837501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,3072,0.009033777647548253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,1536,0.016524444023768108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,7168,0.10279022322760688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,7168,0.0199608885579639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,512,0.012042666474978128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,10240,1.0464497672186959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,768,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,3584,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,1024,0.009481777747472128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,512,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,512,0.15324800544314915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,128,0.006309333360857434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,8192,0.03852533300717672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,10240,0.22069777382744682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,2560,0.0074888889988263445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,51200,0.05925777885648939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,128,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,2560,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,512,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,256,0.0031395554542541504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,3072,0.007341333561473423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,6144,0.15259467230902776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,256,0.036159111393822566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,4096,0.034619556532965765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,64,0.02530488868554433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,512,0.044885334041383534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,768,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,128,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,51200,0.4045564333597819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,7168,0.030759112702475652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,12288,0.02334044377009074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,128,0.01975200076897939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,51200,0.6128746668497721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,12288,0.01330133361948861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,3072,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,512,0.025402666793929204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,2048,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,12288,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,6144,0.02672622270054287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,12288,0.2013875510957506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,256,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,2560,0.08511733346515232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,7168,0.2731120056576199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,1024,0.04728622237841288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,51200,0.1299688948525323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,1536,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,7168,0.08167288700739543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,32,0.027531555957264368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,65536,0.0504755543337928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,16384,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,128,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,64,0.001790222194459703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,1024,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,12288,0.43257511986626523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,7168,0.1542888879776001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,8192,0.03416444526778327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,1024,0.003543111185232798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,16384,0.011699555648697747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,64,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,64,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,3584,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,2048,0.007198221981525421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,51200,0.23130223486158583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,7168,0.022333333889643352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,10240,0.018546667363908555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,5120,0.019469334019554984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,3584,0.06485422452290852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,256,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,10240,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,7168,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,32,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,8192,0.030028445853127375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,32,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,7168,0.03558133376969232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,8192,0.027023111780484516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,12288,0.017860444055663217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,256,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,51200,0.11735288302103679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,2560,0.22946755091349283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,7168,1.0133155186971028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,4096,0.038655999634000994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,1536,0.26131823327806264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,10240,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,10240,0.0177848885456721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,128,0.0029057777590221832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,10240,0.06308888726764254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,2560,0.032609777318106756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,6144,0.011567999919255575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,3584,0.007321777443091075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,7168,3072,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,5120,0.014025777578353882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,51200,0.05031733380423652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,512,0.01458577811717987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,8192,0.20495111412472197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,6144,0.06638044118881226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,3584,0.03445422318246629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,4096,0.008902221918106079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,5120,0.02163822286658817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,5120,0.010191111101044549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,128,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,7168,0.07093689176771376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,1024,0.016523554921150208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,1024,0.07953333192401461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,3584,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,64,0.002461333241727617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,1536,0.019002666076024372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,8192,0.17629778385162354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,12288,0.23364355829026964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,6144,0.7094026671515571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,65536,0.21899467044406465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,1536,0.010962666736708747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,4096,0.00852177788813909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,1024,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,2560,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,512,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,2560,0.008787555827034844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,5120,0.2032559977637397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,128,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,6144,0.010029333333174387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,2560,0.06492888927459717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,7168,2.8159262339274087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,8192,0.07901955313152738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,512,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,16384,1.3159377839830186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,512,0.026698667142126296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,1024,0.07523288991716173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,16384,0.0470106667942471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,1536,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,6144,0.061787552303738065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1536,2560,0.018075555562973022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,10240,0.19570310910542807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,5120,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,768,0.14253955417209202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,3072,0.00905777762333552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,1536,0.006339555399285422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,16384,0.1822942230436537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,8192,0.026874666412671406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,3072,0.014281777871979607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,3584,0.007723555796676212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,10240,0.055045333173539907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,5120,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,512,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,7168,0.010048000348938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,128,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,12288,0.1004106667306688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,128,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,7168,0.02124888863828447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,51200,0.06180533435609606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,65536,0.06345422400368585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,768,0.0782160030470954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,51200,0.2000746726989746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,512,0.017856000198258292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,128,0.016144000821643405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,3584,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,16384,0.03211822112401327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,2560,0.05443377627266777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,10240,0.18601066536373564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,16384,0.03443466623624166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,6144,0.19709155294630262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,16384,0.16084888246324328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,8192,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,256,0.031312889522976346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,1024,0.015460444821251763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,16384,0.716650644938151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,12288,0.019595555133289762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,256,0.023239110906918842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,3072,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,7168,0.08030488755967882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,3072,2.446438259548611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,4096,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,12288,0.1294462283452352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,5120,0.07564000288645427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,2048,0.015311110350820752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,4096,0.09455821911493938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,128,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,65536,1.4041528701782227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,256,0.13013244999779597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,3584,0.02033244404527876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,5120,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,4096,0.034272889296213783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,65536,0.06837955448362562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,6144,0.020403555697864957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,16384,0.09614577558305527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,1024,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,64,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,1536,0.02132266594303979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,768,0.01789155602455139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,128,0.006688000013430913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,65536,0.24666666984558105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,8192,0.10798577467600505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,1536,0.015799111790127225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,65536,12288,0.18915288978152803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,2560,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,65536,0.8275591002570258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,3072,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,4096,0.039184000757005476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,6144,0.1814168824089898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,256,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,16384,0.07202221949895223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,8192,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,5120,0.011047999891969891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,6144,0.37273600366380477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,6144,0.06657777892218696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,51200,0.12143377463022868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,12288,0.05514666769239637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,128,0.007374222079912822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,65536,0.1282622284359402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,65536,1.319350242614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,5120,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,3584,0.026709333062171936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,10240,0.027497778336207073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,65536,0.2721875508626302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,12288,0.018012444178263348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,256,0.02515911062558492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,2560,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,768,0.036778665251202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,3072,0.01997777819633484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,2048,0.01521511044767168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,2048,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,64,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,10240,4096,1.964212417602539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,3584,0.015695111619101632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,7168,0.31718044810824925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,51200,0.06572533316082425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,1536,0.022156443860795762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,6144,0.026486222942670185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,1536,0.027296889159414504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,3584,0.036433776219685875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,2560,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,1024,0.020747555626763236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,8192,0.22658843464321562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,16384,0.028435554769304063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,8192,0.09462755256228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,5120,0.043770667579438954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,64,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,1536,0.00638133328821924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,768,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,16384,0.07227466503779094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,1024,0.007028444773621029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,1024,0.01905066602759891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,7168,0.0136835558546914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,51200,0.13978132936689588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,2048,0.01016711112525728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,512,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,256,0.015281778242852954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,6144,0.16142933898501927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,12288,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,10240,0.0712346699502733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,64,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,1536,0.010681777364677854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,51200,0.6396044625176324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,128,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,3072,0.1284302208158705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,512,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,2048,0.038225776619381376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,51200,0.05298755566279093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,768,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,128,0.02136444383197361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,128,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,10240,0.021384000778198242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,2560,0.2066826687918769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,8192,0.5938213136461046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,768,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,2560,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,65536,0.1867226627137926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,1024,0.012727999852763282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,512,0.0042133331298828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,2048,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,64,0.0021013333979580137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,7168,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,7168,0.042908443344963916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,1536,0.048968000544442065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,8192,0.03158933255407546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,7168,4.172419653998481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,7168,0.01312355531586541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,1536,0.01834755473666721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,7168,0.010399999717871347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,6144,0.06397244665357801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,512,0.013918222652541267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,1536,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,51200,0.05032177766164144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,7168,0.01054133309258355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,65536,1.1088861889309354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,128,0.012400888734393649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,4096,0.038231998682022095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,6144,0.006775110960006714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,128,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,5120,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,3072,0.0359226663907369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,1536,0.04211733407444424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,32,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,8192,0.0217902229891883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,256,0.0306311117278205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,128,0.001989333372977045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,3584,0.03351022137535943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,5120,0.0064080001579390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,7168,1.1091786490546334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,16384,0.1010613309012519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,128,0.0026213334252436957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,128,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,8192,0.05305955476231045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,128,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,7168,0.024733333124054804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,256,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,6144,0.07788266738255818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,768,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,1024,0.013813333378897773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,65536,0.08641955587599014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,512,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,7168,0.040072888135910034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,4096,0.0461751123269399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,32,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,3072,0.05915821923149956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,2560,0.008081778056091731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,10240,0.02296888828277588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,5120,0.372487121158176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,1536,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,1024,0.00963288876745436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,7168,0.4069066577487522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,6144,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,512,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,1024,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,7168,0.4000355667538113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,2048,0.0850471125708686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,4096,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,7168,2.4989261627197266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,51200,0.1473591062757704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,2560,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,16384,0.03373422225316366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,1536,0.06075644493103027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,10240,2.170063018798828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,4096,0.046247111426459424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,128,0.0032888888898822996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,3584,1.7824043697781031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,3584,0.007690666450394525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,3584,0.022016000416543748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,2048,0.0181022220187717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,65536,6.603153652615017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,51200,0.06254666381412081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,16384,0.023860444625218708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,4096,0.022879110442267522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,128,0.005372444374693765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,2560,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,2560,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,6144,0.01204711116022534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,1536,0.011457777685589261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,256,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,2560,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,128,0.0021048889805873237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,7168,0.10337777932484944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,51200,0.5547182295057509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,3584,0.022143999735514324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,2048,0.045122666491402515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,51200,0.09516088830100165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,128,0.006786666810512543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,3072,0.012853333519564735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,7168,0.030004445049497817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,3584,0.06291733185450236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,2560,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,65536,0.7330542140536838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,1024,0.008261333737108443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,16384,0.80240265528361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,32,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,7168,0.03788444399833679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,8192,0.0577217804061042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,3072,0.03591466612286038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,1024,0.00423200014564726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,3072,0.5063369009229872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,8192,0.06855199734369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,3072,0.007714666426181793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,512,0.012797333300113678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,3584,0.11400443977779812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,8192,0.027422222826215956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,1024,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,6144,0.03667466508017646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,12288,0.20359378390842012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,32,0.02348888913790385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,512,0.014861333701345654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,8192,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,128,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,256,0.01846844454606374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,2560,0.005138666679461797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,8192,0.04561511013242933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,4096,0.02988977895842658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,8192,0.12375377284155951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,1536,0.14670222335391575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,3072,0.009753777749008602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,8192,0.05786222219467163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,8192,0.0053680001033677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,3584,0.06680444214079115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,16384,0.1098782221476237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,512,0.02385955552260081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,256,0.010258666343159145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,6144,0.08124622371461657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,128,0.002019555618365606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,1536,0.06986755794949003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,32,0.008303110798199972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,3584,0.06165866719351875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,768,0.061181333329942494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,6144,0.054324444797303945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,6144,0.05102577805519104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,1024,0.3201475673251682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,10240,0.03119644522666931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,768,0.013763555222087435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,51200,1.818521711561415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,16384,0.03087200058831109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,32,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,768,0.013100444442696042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,5120,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,1024,0.0036515556275844574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,128,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,64,0.01369955556260215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,16384,0.0843777789009942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,1024,0.062162664201524526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,10240,0.03158844510714213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,1024,0.0028560000161329904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,5120,0.009708444277445475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,128,0.013419555293189155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,1536,0.05464888943566216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,5120,0.009860444400045607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,768,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,3072,0.03604355454444885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,512,0.008628444539176093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,3584,0.004554666578769684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,51200,0.026657778355810378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,2560,0.027961777316199407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,128,0.0031439999325407874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,6144,0.04174400038189358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,3072,0.007153777612580194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,4096,0.009670221971140968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,2560,0.3301875591278076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,512,0.009713778065310584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,65536,0.0857973363664415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,128,0.014793778459231058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,4096,0.02253955602645874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,2048,0.2135359976026747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,6144,0.02237155536810557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,51200,0.48441600799560547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,7168,0.035764445861180626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,2048,0.026684444811609056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,128,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,768,0.11311111185285781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,65536,0.07032977872424655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,65536,1.7525919808281794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,512,0.042485333151287503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,128,0.012978666358523898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,1024,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,8192,0.02345866627163357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,2048,0.005756444401211209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,3072,0.3053342236412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,12288,0.05724889039993286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,2560,0.0743982195854187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,2560,0.010077333284748925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,65536,0.07224444548288982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,4096,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,6144,0.1274906661775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,64,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,256,0.01663644446267022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,1024,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,6144,0.05506044295099047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,512,0.0053102220926019884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,32,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,512,0.1221911112467448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,12288,0.013761777844693927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,16384,0.06431644492679171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,6144,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,768,0.01717866626050737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,12288,0.050310220983293324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,1024,0.058386663595835366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,5120,0.031234665049446955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,4096,0.0767786635292901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,256,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,1024,0.012135999898115793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,256,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,128,65536,0.12166666984558105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,128,0.01241422196229299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,4096,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,256,0.023772444989946153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,6144,0.12453244792090522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,3072,0.20766133732265898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,3072,0.005292444593376583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,1024,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,32,0.0024986666523747975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,3072,0.008192888564533657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,32,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,1536,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,7168,0.1226079993777805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,1536,0.003604444364706675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,3072,0.014558222558763293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,65536,0.7961760097079806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,2048,0.01646577815214793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,65536,0.5283902486165365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,16384,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,128,0.011534222298198275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,512,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,3584,0.008182221816645728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,512,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,65536,0.34168444739447695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,2560,0.036943111154768206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,128,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,7168,0.04014577799373203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,256,0.01772799922360314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,7168,0.32884444130791557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,2048,0.01572711103492313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,12288,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,128,0.00958933350112703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,32,0.019137778215938143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,3072,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,7168,0.03544888893763224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,10240,0.06281866629918416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,1536,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1024,64,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,6144,0.009480888644854227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,16384,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,51200,0.06262577904595269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,2560,0.004839999808205498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,5120,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,3072,0.0328382220533159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,7168,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,64,0.003142222141226133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,1536,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,1024,0.1167644394768609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,256,0.07924977938334148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,51200,0.10320978032218085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,7168,0.14868444866604275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,4096,0.3514462312062581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,128,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,128,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,12288,0.02195022172398037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,51200,0.01458577811717987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,65536,0.5070506731669108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,512,0.007344889144102733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,3584,0.008192888564533657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,8192,1.9765004052056208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,768,0.02367199957370758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,32,0.015106666419241162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,2560,0.024359110328886244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,12288,0.022264000442292955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,10240,0.025191111697090998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,16384,0.022713777091768052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,8192,0.023573334018389385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,256,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,2048,0.02637600070900387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,256,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,256,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,65536,0.052128000391854175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,4096,0.06819733646180895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,512,0.019666666785875957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,8192,0.5258373154534234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,3584,0.1163288884692722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,5120,0.3246577845679389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,3584,0.012988444831636218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,4096,0.012248888611793518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,32,8192,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,12288,0.054062220785352916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,128,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,768,0.017644445101420086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,10240,0.5519333415561253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,128,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,5120,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,1536,0.015136000182893542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,8192,0.18221155802408853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,51200,4.157877180311415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,3072,0.2493128776550293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,8192,0.14664444658491346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,1536,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,7168,0.18659822146097818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,768,0.03000266684426202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,6144,0.02269777821169959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,10240,0.017176000608338248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,65536,0.05884444713592529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,51200,0.5471706920199925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,512,0.012801777985360889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,3072,0.009438222481144799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,5120,0.036729776197009616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,768,0.02700000007947286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,4096,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,51200,0.11555911435021295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,3584,0.06070488691329956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,2560,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,1536,0.005526222288608551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,1024,0.023349333140585158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,128,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,128,0.015524443652894763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,7168,0.016780444317393832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,51200,0.2981697718302409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,512,0.011350221931934357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,32,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,128,0.008219555848174626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,6144,0.04754399922158983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,51200,0.12973244984944662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,512,0.004230222354332606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,512,0.00868088917599784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,7168,0.010099555883142684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,1024,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,256,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,32,0.0057706667317284485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,64,0.010716444088353051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,64,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,2560,0.007189333438873291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,1536,0.003834666477309333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,3072,0.4513386620415582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,8192,0.010423111418883005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,16384,0.018720888429217868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,3584,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,5120,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,51200,0.05100533366203308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,2560,0.05933511257171631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,51200,0.020043555233213637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,65536,1.281707551744249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,768,0.01349422252840466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,768,0.01425333321094513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,3584,0.04888622297181023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,128,0.009315555294354757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,51200,0.20527911186218262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,1024,0.2112871143552992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,3072,1.0910000271267362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,10240,0.029547555579079524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,1024,0.31414132648044163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,16384,0.09106400277879502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,2048,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,2560,0.019825778073734708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,5120,0.12885955969492593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,256,0.007526222202512953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,7168,1.8503119150797527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,4096,0.021576000584496394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,3584,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,8192,0.014984889162911309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,2048,0.02329066726896498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,1536,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,51200,0.2929333315955268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,2560,0.007556444240940942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,51200,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,5120,0.14585067166222468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,51200,0.06119022104475233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,3584,0.03566222058402167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,6144,0.02161688937081231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,65536,0.2977253331078423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,512,0.010910222099887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,2560,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,16384,1.25928529103597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,2560,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,2048,0.036714666419559054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,16384,3.1576480865478516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,12288,0.02942844563060337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,64,0.022486221459176805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,8192,0.0280693338976966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,4096,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,6144,0.04594755503866407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,768,0.034280889564090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,10240,0.021983999345037673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,65536,0.19677155547671846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,16384,0.02236088944805993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,2560,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,4096,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,16384,0.037272890408833824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,128,0.020455999506844413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,6144,0.011751111182901593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,1024,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,4096,0.016199999385409884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,3584,0.1067955560154385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,51200,0.10957777500152588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,12288,0.00849511143234041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,1536,0.14246666431427002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,2048,0.0239964442120658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,2560,0.020572443803151447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,6144,0.01848266687658098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,65536,0.13321067227257624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,6144,0.044609778457217746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,6144,0.03694222370783488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,7168,0.033258666594823204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,5120,0.05551200111707052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3584,32,0.0052373334765434265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,128,0.0033848887930313745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,768,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,5120,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,65536,0.07162844472461276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,8192,0.0362755556901296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,2048,0.00674044465025266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,4096,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,3584,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,1536,0.01905066602759891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,10240,0.09582400321960449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,51200,0.06473600202136569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,256,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,51200,0.11324532826741536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,3584,0.015127999915017022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,12288,0.8562995592753092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,256,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,4096,0.28876977496676975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,2048,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,1024,0.035896887381871544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,1024,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,65536,0.3709724479251438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,2560,0.2480471134185791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,768,0.0038355555799272326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,2048,0.353690677218967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,2560,0.03164800008138021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,2560,0.015044444137149386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,5120,0.011426666544543372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,51200,0.04829244481192695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,2048,0.14966755443149143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,768,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,3584,0.021195555726687115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,51200,0.131239997016059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,1024,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,65536,0.03457511133617825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,12288,0.0437520013915168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,6144,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,12288,0.10977066887749566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,5120,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,65536,0.9731884002685547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,3072,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,1536,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,10240,0.057143999470604785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,5120,0.027494221925735474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,512,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,64,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,51200,0.2452782260047065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,12288,0.14721511469946966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,5120,0.021262221866183813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,64,0.04856977860132853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,2560,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,64,0.002187555655837059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,12288,0.030216889248953924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,2560,0.004976888911591636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,12288,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,2048,0.00794488853878445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,12288,0.02591288917594486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,32,0.009654222263230218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,3072,0.013730666703648038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,65536,0.027694221999910142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,12288,0.033591111501057945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,3584,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,16384,0.3202515443166097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,8192,0.02920266654756334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,6144,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,32,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,512,0.03295466634962294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,512,0.01349688900841607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,512,0.03067822257677714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,7168,0.089629332224528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,2560,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,4096,0.026015111141734656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,3072,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,128,0.005294222384691238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,6144,0.5357857810126411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,3584,0.016902221573723685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,12288,0.04366488920317756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,51200,0.025943999489148457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,65536,0.06504799922307332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,6144,0.034663110971450806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,5120,0.008691555923885768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,768,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,16384,0.07033689154518975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,3584,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,5120,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,8192,0.017893332574102614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,128,0.016695999436908297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,7168,0.044575111733542554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,12288,0.06479555368423462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,2560,0.014317333698272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,3584,0.03292444348335266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,2560,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,2048,0.040592888991038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,10240,0.02829244401719835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,512,0.005981333139869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,512,0.01828799976242913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,3072,0.030581331915325586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,51200,2.6209778255886502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,6144,0.1572106679280599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,7168,0.021912889348136053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,768,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,128,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,3584,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,65536,0.29380443361070424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,4096,0.03433510992262099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,6144,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,8192,0.12278132968478733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,128,0.0021662221600612006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,768,0.019170666734377544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,2560,0.03994933433002896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,6144,0.023352000448438857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,12288,0.013418667018413544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,3072,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,2048,0.006680888848172293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,128,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,5120,0.11915111541748047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,128,0.013775111072593264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,8192,0.32347289721171063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,16384,0.021783999270863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,8192,0.02162044412559933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,12288,0.0660604437192281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,51200,0.05167644553714328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,512,0.00942666663063897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,51200,0.9965937932332357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,2560,0.05363377928733826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,16384,0.024619556135601465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,6144,0.010805333654085795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,5120,1024,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,3584,0.08131288819842868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,256,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,16384,0.06868800189759997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,65536,3.3737848069932728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,1024,0.14877243836720785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,10240,0.0527617765797509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,8192,0.07690578036838107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,2048,0.007328888608349695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,5120,0.09333155552546184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,51200,0.42344000604417587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,10240,0.10813244183858235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,65536,0.03618933426009284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,51200,1.2686755922105577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,16384,0.04457777738571167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,2048,0.024816888901922438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,1024,0.054323554039001465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,4096,0.012166221936543783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,3072,0.03131377696990967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,5120,0.05520711011356778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,256,0.03436355458365546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,12288,0.030757334497239854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,6144,0.14392622311909994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,2560,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,6144,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,65536,3.5676968892415366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,10240,0.04601866669125027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,128,0.016875555117925007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,12288,0.03208266695340475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,2048,0.024343111448817786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,6144,0.02461066676510705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,4096,0.017856000198258292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,65536,0.05998311440149943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,4096,0.015327110886573792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,5120,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,1536,0.003907555507289039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,1536,0.015783111254374187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,256,0.021176000436147053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,3072,0.011938666303952536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,65536,0.04992622137069702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,512,0.01973155637582143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,10240,0.07650399870342679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,65536,1.434463077121311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,5120,0.9385075039333768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,4096,0.03162666824128892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,2560,0.014000889327791003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,3584,0.025770665870772466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,65536,0.5606337653266059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,6144,0.011384000380833944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,3584,0.02680977847841051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,128,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,51200,19.903942532009548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,6144,0.044575111733542554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,768,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,7168,0.12898843818240696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,7168,0.024342222346199885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,512,0.01903555625014835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,6144,0.11059644487169053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,12288,0.02321333355373806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,16384,0.45715824762980145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,7168,0.19711111651526558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,12288,0.056536886427137584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,16384,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,4096,0.004983110974232356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,12288,0.028939555088678997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,768,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,1024,0.003491555651028951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,51200,0.13054845068189833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,10240,0.016567111015319824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,12288,0.4340764416588678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,1536,0.0041946665280395085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,51200,1.9397315979003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,5120,0.021535999245113794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,7168,0.6548559930589464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,51200,1.2672640482584636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,3584,0.20200444592369926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,16384,0.29819199774000377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,12288,0.5925866762797037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,256,0.002573333266708586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,16384,0.027970665031009253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,65536,0.06490844488143921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,32,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,51200,0.1863786644405789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,6144,0.22424888610839844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,3072,0.008236444658703273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,5120,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,2560,0.4554950926038954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,128,0.02932000160217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,6144,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,3584,0.09671022494633992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,65536,0.06964888837602404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,10240,0.014542222023010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,1536,0.02905155552758111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,3072,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,10240,0.056102222866482206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,768,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,6144,0.12842755847507054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,3072,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,512,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,3072,0.039664887719684176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,256,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,768,0.0680053300327725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,256,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,65536,0.49536175198025173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,5120,0.10025688674714829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,4096,0.07310755385292901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,768,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,512,0.04884355598025852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,128,0.017292444904645283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,8192,0.049626668294270836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,3072,0.039965334865782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,5120,0.016883555385801528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,3072,0.007691555553012424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,5120,0.04289866818322075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,51200,4.034807840983073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,10240,0.02279111080699497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,16384,0.02016266683737437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,8192,0.06120710902743869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,1024,0.04734399914741516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,7168,0.016504888733228047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,7168,0.0166284441947937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,7168,0.12196266651153564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,64,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,2560,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,128,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,2048,0.7167858017815484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,6144,0.016509332590632968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,16384,0.05807999769846598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,10240,0.01585244470172458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,6144,0.023028444912698533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,2560,0.01146933353609509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,12288,0.11440355247921413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,1536,0.03986488779385885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,7168,0.04436622063318888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2560,32,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,51200,0.11581067244211833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,512,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,10240,0.050071110328038536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,12288,0.10347911384370591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,10240,0.019346666004922655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,512,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,256,0.02051555613676707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,51200,0.30495111147562665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,16384,0.03439199924468994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,3584,0.0922071139017741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,51200,0.6062924596998427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,51200,2.371080822414822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,3072,0.14255021678076849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,768,0.03530222177505493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,7168,0.010979555547237396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,12288,0.034183998902638756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,1024,0.050795555114746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,51200,0.2946542104085286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,65536,0.070269332991706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,65536,0.3672400050693088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,3584,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,1536,0.010699555277824402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,512,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,2048,0.023659555448426142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,51200,0.24068355560302734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,7168,0.010298666854699453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,12288,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,12288,0.03626222080654568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,32,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,512,0.014495111174053617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,128,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,3072,0.017557332913080852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,3584,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,4096,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,4096,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,65536,0.037625776396857366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,128,0.03916800022125244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,2048,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,4096,0.00850844466023975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,1024,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,1536,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,256,0.010782221953074137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,5120,0.02747822304566701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,3584,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,65536,0.15868266423543295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1024,2560,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,512,0.002874666617976295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,64,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,51200,0.04276355438762241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,2048,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,3584,0.022356443934970435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,768,0.0032257777121331957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,128,0.013787555197874704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,12288,0.06260533465279473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,2560,0.0074275558193524676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,256,0.005521777603361342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,8192,0.014848889576064216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,6144,0.04845777816242642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,1024,0.016472889317406546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,768,0.02201066745652093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,512,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,5120,0.05157866742875841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,5120,0.013805333111021252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,1024,0.015838222371207345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,10240,0.10979644457499187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,51200,0.053420444329579674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1024,10240,0.040031999349594116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,1536,0.06048266755210029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,3072,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,2048,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,2048,0.12433689170413548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,6144,0.019314666589101154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,256,0.02586577832698822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,3072,0.055826663970947266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,16384,0.015841777125994366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,3584,0.037309332026375666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,128,0.031855109665128924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,1024,0.024004444479942322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,6144,0.012086222569147745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,65536,2.2061049143473306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,5120,0.31599733564588756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,1536,0.03202222122086419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,16384,1.9097173478868272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,6144,0.06530044476191203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,4096,0.05162577827771505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,1536,0.014075555735164218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,2560,0.010446222292052375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,65536,0.04800800151295132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,10240,0.08011555671691895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,512,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,1536,0.02994133366478814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1536,3584,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,10240,0.038254221280415855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,5120,0.025078222155570984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,256,0.02348888913790385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,2560,0.07126577695210774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,10240,0.09033333592944676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,10240,0.012672888735930124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,256,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,7168,0.16739822758568657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,768,0.03255911005867852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,1536,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,128,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,10240,0.031930665175120033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,16384,0.02328355610370636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,51200,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,5120,0.0706826647122701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,51200,0.2936035527123345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,3584,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,4096,0.040960888067881264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,2560,0.006736889067623351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,7168,0.036413331826527916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,16384,0.0522346662150489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,512,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,3584,0.011601777540312873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,128,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,3584,0.018752000398106046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,10240,0.08729066451390584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,3584,0.039339555634392634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,8192,0.03772088885307312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,6144,7.868587917751736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,5120,0.0339902208911048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,4096,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,8192,0.023367111881573994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,768,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,6144,0.00941066692272822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,4096,0.17369066344367134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,1536,0.01789155602455139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,6144,0.03016177813212077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,12288,0.030489779180950586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,16384,0.016561778055297006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,256,51200,0.0513608886135949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,7168,0.1442577838897705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,3584,0.08675377898746067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,3072,0.03899288839764065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,10240,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,2048,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,1024,0.1624364455540975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,3584,0.01148444414138794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,1536,0.01753777762254079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,768,0.021244444780879553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,2560,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,51200,0.09568799866570367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,65536,0.13024622864193386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,1024,0.0074924445814556545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,7168,0.032305777072906494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,8192,0.056311998102400035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,768,0.23581332630581328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,2048,0.012784889174832238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,7168,0.8748399946424695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,51200,0.6381262143452963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,768,0.0033208889265855155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,12288,0.018241778016090393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,7168,0.18970133198632133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,512,0.02421066661675771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,2048,0.036554667684766985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,3072,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,8192,0.029502222935358684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,2048,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,7168,0.06831822130415174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,128,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,7168,0.004707555390066571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,128,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,1536,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,128,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,65536,0.18391556209988066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,1024,0.004573333180612988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,16384,0.0458186666170756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,128,0.02756266627046797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,12288,0.07868088616265191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,512,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,1024,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,3072,0.025286222497622173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,64,0.008071111308203803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,10240,0.030192888445324365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,4096,0.16821955309973824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,51200,0.34349777963426376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,512,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,1024,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,65536,0.18943111101786295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,2560,0.015611555841233997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,5120,0.0211208893193139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,8192,0.025514667232831318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,128,0.006317333214812809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,6144,0.027266666293144226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,16384,0.19091378317938912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,512,0.00960711141427358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,3072,0.08744000064002143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,2560,0.38617245356241864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,12288,0.015481778317027621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,8192,0.20786489380730522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,16384,0.020601777566803824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,32,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,3072,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,6144,0.018336888816621568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,16384,0.05452177921930949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,1536,0.004502222355869082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,256,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,3584,0.018398221996095445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,4096,0.008566222257084316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,2048,0.017659554878870647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,512,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,12288,0.013747555514176687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,512,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,16384,0.11486666732364231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,7168,0.032949331733915545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,256,0.023424888650576275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,7168,0.1324062214957343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,512,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,32,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,2048,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,8192,0.014107555150985718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,65536,0.3693724473317464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,16384,0.018178666631380718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,3072,0.01276177747382058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,512,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,768,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,64,0.0022293333378103045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,3584,0.04974311259057787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,256,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,2560,0.04903644323348999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,256,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,256,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,512,0.02507466740078396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,12288,0.03365777929623922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,7168,0.044673777288860746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,65536,1.25746398501926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,256,0.017454221844673157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,256,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,256,0.0031191110610961914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,7168,0.12553600470225015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,7168,0.012992888689041138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,12288,0.0548986660109626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,3584,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,1024,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,16384,0.025573333104451496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,12288,0.024382222029897902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,128,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,4096,0.04008622301949395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,12288,0.06966311401791044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,12288,0.03885866536034478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,7168,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,3584,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,1024,0.0447324448161655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,1024,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,5120,0.027626666757795546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,32,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,8192,0.03120622369978163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,128,0.007093333535724216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,3072,0.009139555195967356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,3072,0.00905600024594201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,65536,0.2528853416442871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,10240,0.16527556048499212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,7168,0.014656888114081489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,16384,1.431878195868598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,51200,0.06379466586642794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,6144,0.04103822178310818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,65536,2.1510613759358725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,8192,0.880440870920817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,65536,3.505202611287435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,65536,0.008635555704434713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,512,0.00498933345079422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,8192,0.09148089090983073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,3584,0.22112622525956896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,768,8192,0.020236444142129686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,6144,1.0331172943115234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,128,0.008229333493444655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,2560,0.07158488697475858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,16384,0.05429511268933614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,512,0.0041573333243529005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,51200,0.11337155765957302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,3584,0.00849333322710461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,8192,0.08147378100289239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,2560,0.028415110376146104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,65536,0.2304684321085612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,4096,0.016885333591037326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,65536,0.699360900455051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,64,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,3584,0.07353422376844618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,2048,0.05052888724539015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,7168,0.040235555834240384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,3072,0.05087644524044461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,65536,2048,0.36063022083706325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,16384,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,7168,0.06437955300013225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,8192,0.22979378700256348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,12288,0.02260444396071964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,32,0.02110222147570716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,512,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,16384,0.06606400012969971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,2048,0.07695644431644015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,128,0.002036444428894255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,1536,0.1276737716462877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,3072,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,6144,0.18045243952009413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,6144,0.1726177798377143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,1024,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,3072,0.015488000379668342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,1024,0.02628622286849552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,1024,0.03240888979699876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,2048,0.032264000839657254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,768,0.006798222247097228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,1536,0.012054222325483957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,2048,0.010407999985747868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,10240,0.01681066718366411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,65536,0.06843022505442302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,8192,0.012056888805495368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,1536,0.003855111284388436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,51200,0.11710133817460801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,64,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,1024,0.04604711135228475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,6144,0.04639200038380093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,10240,0.1030506690343221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,8192,0.06327466832266913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,5120,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,128,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,3584,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,256,0.07425777779685126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,1536,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,64,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,5120,0.059847997294531934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,12288,0.03139999839994643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,6144,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,512,0.024645333488782246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,4096,0.07852888769573636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,768,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,32,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,3584,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,1536,0.004126222183307012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,2048,0.004521777646409141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,64,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,2560,0.6179262267218696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,768,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,3584,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,768,0.0662133362558153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,8192,0.025384889708624944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,65536,0.30016711023118764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,64,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,65536,0.17994222376081678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,8192,0.2846382194095188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,8192,0.018887110882335238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,3584,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,7168,0.04731733269161648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,128,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,10240,0.06168533033794827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,8192,0.019197333190176222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,3584,0.04188622368706597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,6144,0.011400889191362592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,1024,0.019965334071053397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,768,0.01959377858373854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,12288,0.09419466389550103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,51200,0.6262035369873047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,3072,0.016712889075279236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,5120,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,2560,0.6203191015455458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,65536,0.08345066838794285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,768,0.02966133422321743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,5120,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,5120,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,12288,0.07972800069385104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,128,0.029119110769695703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,128,0.04012889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,128,0.8247511121961805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,6144,0.009135999613338048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,1536,0.028735111157099407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,65536,0.06633244620429145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,3584,0.0107351111041175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,10240,0.08732266558541192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,3584,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,2048,0.00564177789621883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,32,0.013210666676362356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,32,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,5120,0.009128000173303816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,1024,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,1024,0.004129777765936322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,3072,0.07171199719111125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,65536,0.08470044533411662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,128,0.002019555618365606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,7168,0.03074666526582506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,2048,0.023704888092146978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,512,0.0028293333533737394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,51200,0.046181334389580615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,5120,0.01625688870747884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,65536,0.04652355445755852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,128,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,7168,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,1024,0.018536888890796237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,10240,0.025066667132907446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,51200,0.5357235802544488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,3584,0.036192887359195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,1024,0.008392888638708327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,512,0.0028106667515304354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,3584,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,3584,0.009206222163306342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,2048,0.011953777737087674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,6144,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,32,0.00832799987660514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,3072,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,4096,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,256,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,65536,0.06014844444062975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,65536,0.16951821910010445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,128,0.007458666960398356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,32,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,64,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,65536,0.14766400390201145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,2560,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,16384,0.18465067280663383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,2560,0.016872000363137987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,10240,0.09540710846583049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,512,0.02570933269129859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,6144,0.032756444480684065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,65536,4.333764394124349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,6144,0.14133689138624403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,16384,0.448928885989719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,1024,0.0222382214334276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,256,0.025644444757037695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,4096,0.3654782242245144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,768,0.030913776821560327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,4096,0.011583999627166323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,10240,0.019877332780096266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,2048,0.01001422190003925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,8192,8192,0.052301334010230176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,768,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,1024,0.010943999720944298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,3072,0.018368000785509746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,65536,0.23292888535393608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3072,1024,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,3584,0.03670933180385166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,32,0.0020924444413847392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,7168,0.022613333331214056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,1536,0.01408266690042284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,65536,0.0664551125632392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,1024,0.0251217782497406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,10240,0.01261333293384976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,2560,0.01056177748574151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,12288,0.11323377821180557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,2560,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,1536,0.006822222222884496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,2560,0.012099555797047086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,5120,0.03418755531311035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,1536,0.06789066394170125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,65536,12288,0.22262401050991484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,65536,0.13648000028398302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,512,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,2048,0.025997334056430395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,768,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,16384,0.3447431193457709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,2048,0.03128533230887519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,5120,0.00998222248421775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,8192,0.08449600140253703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,64,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,768,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,4096,0.009189333352777693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,3072,0.011072888970375061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,5120,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,32,0.14167022705078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,3584,0.019834667444229126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,256,0.014828445182906257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,65536,0.10751110977596706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,16384,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,4096,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,2048,0.025057777762413025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,3584,0.047988444566726685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,65536,0.6907146771748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,256,0.0053093334039052325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,7168,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,5120,0.026985777748955622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,2048,0.02647466626432207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,51200,3.1449146270751953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,32,0.26904087596469456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,12288,0.08010577493243747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,256,0.03660533494419522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,512,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,256,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,12288,1.2369084888034396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,8192,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,3584,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,51200,0.05407466822200351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,51200,0.36735733350118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,3584,0.26740800009833443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,10240,0.2598782115512424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,64,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,512,0.0063982220987478895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,128,0.06694044669469197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,5120,0.13962577448950872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,7168,0.025807110799683466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,4096,0.09532266855239868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,7168,0.04922755559285482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,7168,0.02551288902759552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,3584,0.04389066828621758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,1536,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,768,0.011399110986126794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,16384,0.05782310830222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,512,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,8192,1.4184284210205078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,8192,0.010448888772063784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,12288,0.06712177726957533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,2560,0.011337777806652917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,12288,0.06654844681421916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,8192,0.09206933445400661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,7168,0.014985778265529208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,256,0.03185688787036472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,256,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,3584,0.011320888996124268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,16384,0.12757689423031277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,256,0.020558221472634208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,8192,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,1024,0.03016266557905409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,51200,0.056552886962890625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,32,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,16384,0.04487288991610209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,8192,0.010689777632554373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,1536,0.07050489054785834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,5120,0.007053333024183909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,1536,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,5120,0.024316444993019104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,3072,0.027250667413075764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,5120,0.011561777856614856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,512,0.07737066348393758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,128,0.014440889159838358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,51200,0.05395644572046068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,16384,0.10269333256615533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,5120,0.13628355662027994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,6144,0.05804977814356486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,512,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,1024,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,3584,0.019675556156370375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,4096,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,2560,0.007083555890454187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,2560,0.032652444309658475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,65536,0.13316088252597386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,256,0.009749333063761393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,256,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,51200,0.049235555860731334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,10240,0.02524799936347538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,512,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,256,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,10240,0.08414222134484185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,1536,0.11202222771114773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,5120,0.027055111196306016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,1024,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,256,0.01552888916598426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,512,0.055438223812315196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,2048,0.03810844487614102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,6144,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,32,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,1536,0.04344888859324985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,3072,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,16384,0.021073778470357258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,10240,0.0447671115398407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,6144,0.024870221813519795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,1024,0.024596444434589807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,1536,0.36173688040839297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,51200,0.9365812937418619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,1536,0.03859644465976291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,4096,0.030993776188956365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,2048,0.01330488920211792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,256,0.023652444283167522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,2560,0.016519111063745286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,16384,0.021127111381954614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,3584,0.018952889574898612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,16384,0.01924711134698656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,12288,0.099016891585456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,5120,0.0373288889726003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,8192,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,6144,0.033260444800059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,3584,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,2048,0.013659555878904132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,128,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,256,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,2048,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,1536,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,10240,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,5120,0.09373422463734944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,256,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,16384,0.10199999809265137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,10240,0.43076711230807835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,3584,0.014648000399271647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,5120,0.013160000244776407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,4096,0.020278222031063504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,2560,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,5120,0.05335288908746508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,5120,0.015796444482273526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,12288,2.913058598836263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,4096,0.042596442831887134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,128,0.01973511113060845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,1024,0.013040888640615674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,7168,0.06999288664923774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,12288,0.027148443791601393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,16384,0.1419866614871555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,65536,0.174288895395067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,7168,0.0451164444287618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,6144,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,10240,0.01680355601840549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,4096,0.0403093331389957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,3584,0.2603573269314236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,6144,0.013830222189426422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,3584,0.05669866667853462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,3584,0.008438222110271454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,512,0.011709333293967776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,2048,0.0063493334584765965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,1536,0.003823111040724648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,128,0.006207110981146495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,512,0.00564177789621883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,2048,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,768,0.005655111124118169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,1536,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,128,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,3072,0.004249777644872665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,12288,0.05827022261089749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,512,0.0038515557017591265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,32,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,65536,0.008814222282833522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,16384,0.06570844517813788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,256,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,10240,0.13373955090840658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,51200,13.303595648871529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,5120,0.034632888105180525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,16384,0.06950577762391832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,1024,0.002826666666401757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,8192,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,768,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,16384,0.0961315565639072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,51200,0.04030844569206238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,3584,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,8192,0.018176888426144917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,32,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,65536,0.083697775999705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,5120,0.024022221565246582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,768,0.0034773333205117118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,6144,0.0650213360786438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,512,0.009594666461149851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,2048,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,10240,0.03894755575391982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,512,0.00553955551650789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,2048,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,512,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,768,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,256,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,51200,0.1331013308631049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,5120,0.17959200011359322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,2048,0.02649066680007511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,3584,0.14935289488898382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,256,0.054894222153557666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,3072,0.06361066632800631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,3072,0.013565333353148567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,2560,0.012482666307025485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,8192,0.1518835491604275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,3584,0.010139555566840703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,128,0.019720888800091214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,1024,0.013064888616402945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,3072,0.012491555677519904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1536,2048,0.029273778200149536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,10240,0.012418666647540199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,8192,0.09979110956192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,128,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,512,0.002765333279967308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,6144,0.06498755349053277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,16384,0.03643111056751675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,6144,128,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,65536,0.17257422871059844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,4096,0.09200711382759942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,6144,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,65536,0.3077644507090251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,256,0.002502222235004107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,768,0.005364444520738389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,10240,0.017952000101407368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,1024,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,8192,0.025954667064878676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,64,0.005519999812046687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,3584,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,12288,0.03121688961982727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,65536,0.08579999870724148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,10240,0.052036444346110024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,7168,0.03654399845335219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,128,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,3584,0.01574222246805827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,51200,51200,5.390486399332683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,7168,0.02091466718249851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,512,0.010047111246320937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,256,0.013771555489963956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,512,0.09536266989178127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,65536,0.05650222301483154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,7168,0.009831111464235518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,256,0.014402666025691561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,1536,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,3072,0.010919999745157031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,768,0.00562755556570159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,8192,0.01371911085314221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,2560,0.017200888858901132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,256,0.006671999891599019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,2048,0.017655111021465726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,768,0.0041946665280395085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,512,0.0068959999415609576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,12288,0.0350835555129581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,3072,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,2560,0.028896888097127277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,51200,0.051501333713531494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,7168,0.028558221128251817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,3584,0.0063928887248039246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,3072,0.019096889429622226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,4096,0.10740533139970566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,1024,0.01608711149957445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,12288,0.0221368885702557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,12288,0.18466133541531035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,4096,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,51200,0.05424000157250298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,10240,0.5294151306152344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,5120,0.013295110729005603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,1536,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,51200,0.050684445434146456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,6144,0.010336000058386061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,4096,0.023231999741660223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,512,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,8192,0.20997867319318983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,512,0.012844444149070315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,16384,0.3744693332248264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,2560,0.023652444283167522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,3584,0.024667556087176006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,65536,7.919522603352864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,3584,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,128,12288,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,65536,0.06996888584560819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,1536,0.02273511058754391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,7168,0.06708711385726929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,5120,0.01611822181277805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,16384,0.022286223040686712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,8192,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,51200,0.04275822308328417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,2560,0.07804799742168851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,128,0.007275555696752336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,512,0.02199555602338579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,16384,0.17548267046610513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,768,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,6144,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,65536,0.05239466826121012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,12288,0.2899271117316352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,12288,0.11131289270189072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,2560,0.030461334519916113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,256,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,4096,0.01311377767059538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,8192,0.014544889330863953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,8192,0.02030311193731096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,8192,0.011060444845093621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,7168,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,10240,3.9077118767632375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,65536,2.869560029771593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,128,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,12288,0.0229759994480345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,16384,0.05794577466117012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,3072,0.030686222844653662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,5120,0.011969777444998423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,1024,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,128,0.008096888661384583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,2048,0.004218666503826777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,12288,0.22383554776509604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,51200,0.651167975531684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3072,128,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,65536,0.22204532888200548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,5120,0.025349333882331848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,512,0.009431111315886179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,4096,0.00998577806684706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,5120,0.09240711397594875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,12288,2.177011489868164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,1536,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,8192,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,1024,0.015855110353893705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,2560,0.07321777608659533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,12288,0.01872711049185859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,16384,0.02146311104297638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,6144,0.033369776275422834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,1536,0.0322017769018809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,4096,0.5791609022352431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,8192,0.03148533238304986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,8192,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,2048,0.0064479998416370815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,2048,0.010771555205186209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,128,0.035727110173967146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,10240,0.03340355555216471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,512,0.007342221836249034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,256,0.05841955873701307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,51200,0.12886044714185926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,256,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,128,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,768,0.018244443668259513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,3072,0.022357333037588332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,32,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,6144,0.02995022137959798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,10240,0.04903022117084927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,12288,0.02594311038653056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,51200,0.4239555464850531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,7168,0.044799112611346774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,51200,0.04963555600908068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,1024,0.006060444646411472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,3584,0.07814400063620673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,2048,0.026399110754330952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,3072,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,1536,0.02916355596648322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,51200,0.061392002635531955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,65536,2.0662765502929688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,51200,0.053561776876449585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,4096,0.12114489078521729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,8192,0.02350577712059021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,512,0.02312533391846551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,6144,0.026366222235891554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,256,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,3584,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,51200,0.022962666220135156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,1024,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,2048,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,6144,0.013791999883121915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,4096,0.054175111982557506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,7168,0.07308977842330933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,2560,0.012720000412729053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,2560,0.007207999626795451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,3072,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,1536,0.03641511003176371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,5120,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,1536,0.019686222076416016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,2560,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,5120,0.011047999891969891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,512,0.017487999465730455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,2048,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,3584,0.015261333849694995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,1536,0.01311644415060679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,1536,0.06591111421585083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,2048,0.10279555453194512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,6144,0.021260443660948012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,8192,0.02199199961291419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,2048,0.014098667436175875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,2560,0.25557422637939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,16384,0.0899626678890652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,1024,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,768,0.04629688792758518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,4096,0.03275110986497667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,1024,0.011594666375054253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,128,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,64,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,3584,0.015048889650238885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,512,0.006303999986913469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,8192,0.05537155601713392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,51200,0.40508532524108887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,1024,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,8192,0.15829866462283665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,5120,0.06016800138685438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,12288,0.013401777380042605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,2048,0.02568799919552273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,1024,0.02717777755525377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,10240,0.017991110682487488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,6144,0.09342400232950847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,4096,0.01683200067943997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,2048,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,64,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,2048,0.012727999852763282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,51200,0.06070311201943291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,12288,0.09195199939939712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,10240,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,10240,0.12654666105906168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,256,0.02446488870514764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,4096,0.007751999629868402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,768,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,12288,0.01444533301724328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,16384,0.0479964448346032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,65536,0.22401867972479927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,51200,1.4694346321953669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,8192,0.24855465359157988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,3584,0.08029066854053073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,2048,0.10842488871680366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,10240,0.1327635579639011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,3584,0.06335822078916761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,256,0.013722666435771518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,256,0.0022159999029503926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,256,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,2048,0.007873777714040544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,3072,0.09671288728713989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,1536,0.03104089034928216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,1536,0.030253334177864924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,6144,2560,0.007348444726732042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,1024,0.026130666335423786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,51200,0.25276266203986275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,3584,0.12056710984971787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,5120,0.02875822120242649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,256,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,6144,0.021380444367726643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,64,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,16384,0.0052284445199701525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,5120,0.030243555704752605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,128,0.11698577139112686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,2048,0.02608266638384925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,768,0.008725333544943068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,6144,0.010327111515733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,32,0.01052266690466139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,512,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,128,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,1024,0.027224888404210407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,2048,0.05697866943147448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,128,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,12288,0.04161422120200263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,65536,0.06352888875537448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,128,0.08298399713304308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,6144,0.018366222580273945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,65536,0.4918213420444065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,10240,0.012427555190192329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,6144,0.013462222284740873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,5120,0.019088000059127808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,4096,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,6144,0.01852799952030182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,3072,0.014113777213626437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,2048,0.014246222045686511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,12288,0.23054043451944986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,51200,0.052547554175059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,8192,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,256,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,51200,0.05356088942951626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,256,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,5120,0.028413332170910303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,4096,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,512,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,1024,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,768,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,512,0.016707556115256417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,3584,0.007376888559924231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,256,0.03888622257444594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2048,51200,0.05687555339601305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,3072,0.14756533834669325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,4096,0.08033600118425158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,1536,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,65536,0.3472542232937283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,1024,0.015448000695970325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,3072,0.09360355801052517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,51200,0.047876444127824574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,4096,0.011906666888131036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,16384,1.7121022542317708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,1536,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,64,0.07306933403015137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,1024,0.00904444439543618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,10240,0.020741333564122517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,1024,0.01421511173248291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,6144,0.12998667028215197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,51200,0.06530577606625028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,2048,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,3584,0.03573066658443875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,7168,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,16384,0.5767937766181098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,2048,0.012411555482281579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,1536,0.0041946665280395085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,3584,0.4050079981486003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,3584,0.03496177660094367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,51200,0.027093332674768236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,512,0.0053431110249625305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,5120,0.027440889014138117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,16384,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,2560,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,1536,0.02557955516709222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,10240,3.112347496880425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,1536,0.011922666596041786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,7168,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,3584,0.033012443118625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,512,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,3584,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,3072,0.06724444362852308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,4096,0.010419555836253695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,2048,0.013059555656380124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,5120,0.08406577507654826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,1536,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,8192,0.07020799981223212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,16384,0.03943555553754171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,7168,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,512,0.06727377573649089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,1536,1.164770656161838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,64,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,128,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,256,0.025622222158643935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,1024,0.014535110857751636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,64,0.02628799941804674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,10240,0.03947644432385763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,256,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,128,0.018467555443445843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,65536,0.08455466561847264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,16384,0.20952088303036162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,7168,0.1256951093673706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,3584,0.17990133497450086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,512,0.008194666769769456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,768,0.009383111364311641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,128,0.012496888637542725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,12288,2.870320849948459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,65536,0.174291557735867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,1024,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,1024,0.02124711043304867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,256,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,5120,0.3185288906097412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,6144,0.012276444170210095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,2560,0.012786666552225748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,10240,0.22047732936011422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,5120,0.010043555663691627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,12288,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,10240,0.027031110392676458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,1536,0.007485333416197035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,1536,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,5120,0.018814222680197824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,51200,0.11526221699184841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,51200,0.374034669664171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,3584,0.07521510786480375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,128,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,3072,0.11365511682298447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,256,0.006601777755551868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,3584,0.04525688952869839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,128,0.01682844426896837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,5120,0.03750577900144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,2048,0.012548444171746572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,1024,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,3584,0.018953777021831937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,10240,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,4096,0.009540444446934594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,256,0.012442666623327466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,6144,0.16284888320498997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,2048,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,256,0.0023946666883097752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,256,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,8192,0.015967110792795818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,10240,0.025086222423447505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,2560,0.13270044326782227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,1024,0.05327377716700236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,1536,0.01478666729397244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,2560,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,12288,0.09295377466413711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,1024,0.018936889039145578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,3072,0.035092443227767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,2048,0.03303733468055725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,1024,0.019987555013762582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,5120,0.031481779283947416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,6144,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,64,0.0021911110315057966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,5120,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,1024,0.04775377776887682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,512,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,1024,0.006103111224042044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,768,0.005865777946180767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,8192,0.01534400052494473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,4096,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,1536,0.006204444501135085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,8192,0.015384000208642749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,2560,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,768,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,32,0.7328071064419217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,256,0.015174221661355762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,64,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,256,0.015777778294351365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,768,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,1024,0.035282668140199445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,5120,0.01769333415561252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,16384,0.02852177619934082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,768,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,3584,0.04354666670163473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,1536,0.02000711030430264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,8192,0.15375999609629312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,5120,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,3072,0.028816888729731243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,256,0.02180533276663886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,128,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,64,0.014905777242448596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,51200,0.067830224831899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,256,0.015447111593352424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,6144,0.062014222145080566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,512,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,2560,0.022738666998015508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,3584,0.03249511122703552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,65536,0.11367021666632758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,10240,0.03658488723966811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,128,0.020862221717834473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,3584,0.030039999220106337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,10240,0.02572533322705163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,2560,0.010436444646782346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,6144,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,10240,0.014200000299347771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,3072,0.007542221910423703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,8192,0.13970222738054064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,4096,0.22325065400865343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,64,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,512,0.0120328888297081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,5120,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,32,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,3072,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,768,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,3072,0.1946551137500339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,256,0.006762666834725275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,8192,5120,0.02308355602953169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,16384,0.0229751103454166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,2048,0.06315822071499295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,65536,0.21352622244093153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,3584,0.024454222785101995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,2560,0.4527715577019586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,8192,0.013027555412716336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,7168,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,7168,6144,2.124124526977539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,512,0.03175288769933913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,65536,0.8819333182440864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,10240,0.014930667148696052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,10240,0.02495555579662323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,7168,0.1316159963607788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,32,0.0041848888827694785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,1024,0.01979822251531813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,12288,0.1193351083331638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,5120,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,5120,0.022090666823916968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,12288,0.032327112224366926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,64,0.010668444136778513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,768,0.06347999970118205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,512,0.02739288906256358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,5120,0.14743822150760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,2048,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,6144,0.011984888878133563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,16384,0.3654133213890924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,5120,0.05164088805516561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,4096,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,5120,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,6144,0.019483556350072224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,512,0.04013244311014811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,1536,0.010066666536860997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,2048,0.004585777719815572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,128,0.01884533299340142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,16384,0.04297955499755012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,7168,0.014652444256676568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,5120,0.03904622130923801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,7168,10240,0.042136000262366406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,5120,0.019142222073343065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,1536,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,65536,0.12588266531626383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,128,0.0018764444523387486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,2048,0.004927111168702443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,8192,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,512,0.002591111179855135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,6144,0.13535822762383357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,2048,0.027837332752015855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,3584,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,1536,0.007848000360859765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,32,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,6144,0.020218667056825426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,5120,0.03088444471359253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,3584,0.007826666865083907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,12288,0.03137777911292182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,12288,0.03495733274353875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,2048,0.015819556183285184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,16384,0.1211164395014445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,3584,0.029008888536029395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,2560,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,5120,0.04915822214550442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,5120,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,10240,0.012568000290128918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,1024,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,3584,0.025722665919197932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,768,0.045863999260796436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,512,0.014848889576064216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,6144,0.02418666746881273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,3584,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,12288,0.12032711505889893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,512,0.005715555614895291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,65536,0.9064728418986002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,10240,0.01609777741962009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,7168,0.0626044472058614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,16384,0.22318577766418457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,512,0.01995022263791826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,1536,0.02558044426971012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,128,0.038944890101750694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,512,0.011808888779746162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,7168,0.25613334443834096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,128,0.016064888901180692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,10240,0.018936889039145578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,6144,0.04120355513360765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,6144,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,2560,0.0221404449807273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,51200,3.557915581597222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,3584,0.020170667105250888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,256,0.023842665884229872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,10240,0.03010933266745673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,4096,0.048956443866093956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,32,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,2560,0.011476444701353708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,5120,0.10641600026024713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,3584,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,16384,0.02314222190115187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,5120,0.024359110328886244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,51200,0.3364862336052789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,1024,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,3072,0.035818666219711304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,5120,0.01425244410832723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,5120,0.03877066572507223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,8192,0.04971555537647671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,512,0.02014488809638553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,6144,0.05688266621695625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,6144,0.7394124666849772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,2560,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,3584,0.010980444649855295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,65536,0.024850666522979736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,4096,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,1536,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,4096,0.12406044536166722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,768,0.01734755602147844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,4096,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,65536,0.7504560152689616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,1536,0.13377955224778917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,64,0.07372444205813937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,65536,0.11738667223188613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,6144,0.029198222690158423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,1024,0.023131555981106226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,5120,0.038076443804634943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,12288,2.081097708808051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,8192,0.1686462163925171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,6144,6144,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,768,0.015110222829712762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,2560,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,256,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,6144,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,8192,0.3279208872053358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,8192,1.184035513136122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,4096,0.39176532957288956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,256,0.0031822222388452957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,5120,0.020977778567208186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,8192,0.16863555378384062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,3584,0.008066666622956594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,5120,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,1024,0.06338666545020209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,4096,0.06628533204396565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,256,0.017160000072585214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,1024,0.04006933172543844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,128,0.0021564445147911706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,2560,0.022667555345429316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,4096,0.228836456934611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,256,0.012196444802814059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,64,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,51200,0.05731822384728325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,3072,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,768,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,5120,0.20769155025482178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,10240,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,65536,0.0748737785551283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,32,0.013047110703256396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,6144,0.040580444865756564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,4096,2.4469208187527127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,6144,0.050944000482559204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,2048,0.050352000527911715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,5120,0.0376853346824646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,6144,0.012223111258612739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,32,0.028821332587136164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,64,0.0018702221827374564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,10240,0.4309599929385715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,128,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,768,0.020554666717847187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,3584,0.20488888687557646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,2048,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,6144,0.033313777711656355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,10240,0.01977688901954227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,6144,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,64,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,6144,0.06702399916119046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,1536,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,768,0.05305599835183886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,128,0.015384000208642749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,51200,0.11193511221143936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,51200,12.322578430175781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,12288,0.043529778718948364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,7168,0.017528888252046373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,3072,0.02956977817747328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,2560,0.012180444267061023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,64,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,1536,0.015531554818153381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,16384,0.10582933160993789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,6144,0.14972710609436035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,1024,0.08822755681143866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,3584,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,512,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,7168,0.0695342222849528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,512,0.01240355521440506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,3584,0.15190045038859049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,512,0.010670222342014313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,8192,0.087135997083452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,256,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,32,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,3072,0.03175288769933913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,12288,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,7168,0.012489777472284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,8192,0.015301333533393012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,128,0.026894221703211468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,7168,0.16502043935987684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,16384,0.03275555703375075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,768,0.07529244158003065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,6144,0.13742578029632568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,4096,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,10240,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,1536,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,2560,0.017151110702090792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,512,0.01238933371173011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,16384,0.03366311060057746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,6144,0.11411200629340278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,65536,0.6241333219740126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,5120,0.0502800014283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,128,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,10240,0.18849333127339682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,1024,0.015463110473420886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,51200,0.016007999579111736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,768,0.005527110977305307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,3072,0.022258667482270136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,6144,0.3761315610673692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,2048,0.37133065859476727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,65536,0.07765244775348239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,256,0.2389146751827664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,1024,0.019259555472267997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,8192,0.044532444741990834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,5120,0.04968355430497063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,128,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,3584,0.018917333748605516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,5120,0.02642222245534261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,12288,0.023453333311610754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,2048,0.006746666712893381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,16384,0.02139733400609758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,51200,0.25069067213270396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,65536,0.058989332781897656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,768,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,6144,0.014001778430408902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,512,0.02660888930161794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,16384,0.042508443196614586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,2560,0.006892444358931647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,512,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,2048,0.019296889503796894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,10240,0.16020354959699842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,256,0.022226666410764057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,4096,0.02680622206793891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,7168,0.20680888493855795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,2048,0.006333333336644703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,3072,0.012140444583363004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,51200,3.468006134033203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,5120,0.01907999979125129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,3584,0.013329777452680798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,65536,0.012164444559150271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,8192,0.0837271081076728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,128,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,768,0.008640000389681922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,6144,0.17816532982720268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,65536,0.13203556007809109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,32,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,4096,0.0725839998986986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,1024,0.09029422203699748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,5120,0.07330933544370863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,3072,0.049264887968699135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,512,64,0.008016000191370646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,128,0.11654310756259495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,12288,0.05061155557632446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,3584,0.05985155370500353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,8192,0.06184444162580702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,3584,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,1024,0.014514666464593677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,4096,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,3584,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,512,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,65536,0.07685599724451701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,3072,0.020650666620996263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,512,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,6144,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,512,0.015259555644459195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,5120,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,3584,0.011027555498811932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,3072,0.007309333317809635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,12288,0.04338577720854017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,4096,0.008119111259778341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,51200,0.12989511754777697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,3584,0.008016000191370646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,4096,0.045040001471837364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,3072,0.016213332613309223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,2048,0.016221332881185744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,12288,0.049011554982927114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,4096,0.007207999626795451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,768,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,128,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,4096,5.226698557535808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,1024,0.005978666659858491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,128,0.020943111843532983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,10240,1.543558226691352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,12288,0.01627111103799608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,7168,0.046988444195853345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,10240,0.08807022041744655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,2560,0.018922666708628338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,7168,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,512,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,512,0.0025155554629034465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,10240,0.1427662240134345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,7168,0.026128000683254663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,1024,0.026504000027974445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,7168,0.024847999215126038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,3584,0.005640000104904175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,32,0.011972444752852121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,2560,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,10240,0.01273333364062839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,1024,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,5120,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,256,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,2560,0.15386666191948786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,7168,0.0343777769141727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,256,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,3072,0.029172443681293066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,12288,2.418808831108941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,10240,0.33408000734117294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,3072,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,2048,0.00887111077706019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,768,0.05435377690527174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,256,0.07759466436174181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,512,0.007349333001507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,32,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,5120,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,51200,0.047525333033667676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,10240,0.013284444808959961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,128,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,128,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,8192,0.030360887447992962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,16384,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,512,0.01850755512714386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,5120,0.04464088877042135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,256,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,128,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,8192,0.013431111143694984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,256,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,8192,0.07014844152662489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,256,0.004570666700601578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,4096,0.08264622423383924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,768,0.01789955629242791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,51200,1.990171644422743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,7168,0.03475288881195916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,2048,0.018792000081804063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,10240,0.6008782386779785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,256,0.023028444912698533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,256,0.015094222293959724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,2048,0.009735110733244155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,3072,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,1024,0.013814222481515674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,16384,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,2048,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,7168,0.016913778252071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,65536,0.700650691986084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,3072,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,4096,0.00795911086930169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,4096,0.009083555804358589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,128,0.0029226665695508323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,16384,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,2048,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,7168,0.021720889541837905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,10240,0.18565689192877877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,2560,0.03016266557905409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,5120,0.052970667680104576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,128,0.010430222584141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,32,0.02195022172398037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,10240,0.03688888748486837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,2560,0.033139556646347046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,3072,0.08111555708779229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,6144,0.10328800148434109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,32,0.1892568932639228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,1024,0.013975999421543546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,1024,0.025340444511837427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,4096,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,51200,0.0790319972568088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,2048,0.04993866549597847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,65536,0.1058186690012614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,8192,0.01831288966867659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,12288,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,10240,0.7751830948723687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,12288,0.040334221389558576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,5120,0.012649777862760754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,256,0.007920000288221572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,2560,0.06512977679570515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,2048,0.025427555044492085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,16384,0.19065777460734049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,6144,0.01794755624400245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,128,0.022778666681713525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,6144,0.020029332902696397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,1536,0.004872000051869286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,1536,0.035573333501815796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,768,0.03787644373046027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,3584,0.7413466771443685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,4096,0.2740631103515625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,6144,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,768,0.009180444810125563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,4096,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,8192,0.010749333434634738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,7168,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,32,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,2048,0.2373164494832357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,3072,0.028228445185555354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,768,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,1536,0.027663111686706543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,10240,0.04210755560133192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,512,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,10240,0.06156533294253879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,5120,0.2833271026611328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,768,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,3584,0.42406132486131454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,3584,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,2048,0.014281777871979607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,5120,0.1402684450149536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,12288,0.017489777670966256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,3584,0.07517155673768786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,5120,0.01149866647190518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,2048,0.013953778478834363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,2048,0.025694222913848028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,3584,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,3072,0.028588443994522095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,2560,0.02327999969323476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,16384,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,3072,0.00850133349498113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,12288,0.15557332833607992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,1536,0.028000887897279527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,32,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,256,0.01497866710027059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,12288,0.08983021974563599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,1536,0.015187554889255099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,12288,2.678062227037218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,768,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,2560,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,512,0.00406577769252989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,8192,0.09044444561004639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,10240,0.21760000122918022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,256,0.026132444540659588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,6144,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,1536,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,2048,0.033040887779659696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,51200,2.099103079901801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,8192,4096,0.011796444654464722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,3584,0.031587554348839655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,512,0.01513066722287072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,1024,0.061011552810668945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,3072,0.019229332605997723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,7168,0.15032977528042263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,65536,0.2346408896976047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,51200,0.18939110967848036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,768,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,12288,0.029358221424950495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,512,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,16384,0.05509066581726074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,1536,0.006321777900060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,3072,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,128,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,51200,1.9432729085286458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,1536,0.00867377801073922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,3072,0.06809422042634752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,4096,0.012084444363911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,2048,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,6144,0.04028266668319702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,4096,0.00832266691658232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,8192,3072,0.028733332951863606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,4096,0.00757422215408749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,3072,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,10240,0.06230044364929199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,10240,0.04251555601755778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,1024,0.011573333707120685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,8192,0.08888444635603164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,4096,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,6144,0.1116151147418552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,5120,0.01774222155412038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,2048,0.041314668125576444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,51200,0.07260888814926147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,768,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,51200,0.34596800804138184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,32,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,8192,0.05721155802408854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,2048,0.0424124432934655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,512,0.19719910621643066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,256,0.008653333617581261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,128,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,8192,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,512,0.029416001505321924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,12288,0.26172622044881183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,2048,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,768,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,1536,0.19188177585601807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,768,0.010803555448849997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,16384,0.02606222199069129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,7168,0.03290844294759963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,256,0.01699199941423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,3584,0.07399821943706937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,128,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,5120,0.023561777340041265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,64,0.0019084444890419643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,16384,0.016183111402723525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,16384,0.05071466830041674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,1024,0.021551999780866835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,7168,0.010009778042634329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,65536,0.2766248914930555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,16384,0.04582755433188545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,256,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,65536,0.2748026582929823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,3072,0.009880888793203566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,128,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,2048,0.004466666529575984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,4096,0.02274399995803833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,51200,7.0562498304579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,1536,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,10240,0.5509760114881728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,3072,0.9084978103637695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,12288,0.23089154561360678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,65536,0.08543999989827473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,3072,0.03070666723781162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,4096,0.039232889811197914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,1024,0.014633778068754407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,256,0.010669333239396414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,128,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,512,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,64,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,16384,0.1965493361155192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,65536,0.14559466309017605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,16384,0.021185777253574792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,65536,0.09552088710996841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,128,0.10965422126981948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,32,0.0052933332820733385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,6144,0.11062133312225342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,8192,0.09552177455690171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,1536,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,6144,0.009663999908500248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,2048,0.028468443287743464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,8192,0.0281368891398112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,51200,0.13986844486660427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,768,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,768,0.01793333391348521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,5120,0.04407199886110094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,2048,0.03719200028313531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,51200,0.1450293329026964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,10240,0.025246222813924152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,16384,0.052899556027518384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,7168,0.1564684444003635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,256,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,1024,0.003083555648724238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,5120,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,5120,0.03625866770744324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,3072,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,65536,0.1585635609096951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,1536,0.24089333746168348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,3072,0.030050665140151978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,128,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,1024,0.04098844528198242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,51200,0.9219840367635092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,32,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,1536,0.0038337777886125776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,12288,0.19969511032104492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,5120,0.04157955447832743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,7168,0.13068355454338923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,128,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,128,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,3584,0.02472266720400916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,1536,0.012317332956526013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,7168,0.05120533373620775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,12288,0.050442665815353394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,4096,0.007124444676770105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,256,0.005792888916201062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,2560,0.008618666893906063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,256,0.01074044406414032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,65536,0.7656284438239204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,2048,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,6144,0.11388977368672688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,65536,0.4506613413492839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,768,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,768,0.017677333619859483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,2560,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,4096,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,2048,0.009715555442704095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,1536,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,5120,0.010314666562610203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,2048,0.014915555715560913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,1024,0.15069244967566595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,64,0.012431110772821637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,10240,0.013521778086821238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,3584,0.036779556009504534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,256,0.007939555578761632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,5120,0.011146666275130378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,256,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,3584,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,256,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,4096,0.01705422169632382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,768,0.01612622208065457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,128,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,65536,0.08838399913575913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,3584,0.03031377659903632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,7168,0.13973155286577013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,32,0.004338666796684265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,4096,0.039000888665517174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,1536,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,128,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,6144,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,32,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,768,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,7168,0.045364442798826426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,10240,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,4096,0.16134133603837755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,768,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,1536,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,16384,0.018165333403481376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,256,0.023391111029518977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,2048,0.031890667147106595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,3584,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,1024,0.0473217765490214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,2048,0.015225778023401896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,32,0.020117332537968952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,51200,0.6606239742702907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,3584,0.542624897427029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,5120,0.3516382111443414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,2048,0.1263422171274821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,12288,0.06168533033794827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,2048,0.2966639995574951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,32,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,2048,0.01462577780087789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,1024,0.06820266776614718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,3584,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,4096,0.014872888724009195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,3072,0.007576888634098901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,4096,0.024217777782016333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,1536,0.03286400106218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,7168,0.03710222244262695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,512,0.586624887254503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,2048,0.00796888851457172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,32,0.0025315555847353404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,8192,0.06726577546861437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,16384,12288,0.5779244634840224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,512,0.009758222434255812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,65536,0.06493688954247369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,3584,0.06299289067586263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,1536,0.020608888732062448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,6144,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,2048,0.010936000280910067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,5120,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,2560,0.007743111087216272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,1536,0.018575110369258456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,3584,0.025574222207069397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,3584,0.012450666891203986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,3584,0.015024888846609326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,128,0.03131466772821214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,5120,0.03558133376969232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,2560,0.008978666530715095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,3072,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,6144,2560,0.03399466805987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,5120,0.04047911034690009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,16384,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,51200,0.09921244117948745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,12288,0.0676657756169637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,32,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,6144,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,64,0.04114133450720046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,3584,0.027419555518362258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,12288,0.0859839982456631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,10240,0.03145600027508206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,256,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,8192,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,3072,0.06593333350287543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,768,0.02295733326011234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,2048,0.036432888772752546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,4096,0.012710221939616732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,51200,0.259789334403144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,12288,0.0820888876914978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,512,0.009826666778988307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,4096,0.014729777971903482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,51200,0.13117599487304688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,128,0.016207110550668504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,3072,0.03428711162673102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,1536,0.06997866763008966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,512,0.007809778054555257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,1536,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,1536,0.003524444583389494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,768,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,7168,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,7168,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,51200,0.0995582209693061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,256,0.004464888738261329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,768,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,512,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,1024,0.03278577658865187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,5120,0.025977777110205755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,256,0.007788444558779399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,4096,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,7168,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,12288,0.14799643887413874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,768,0.010325333310498131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,7168,0.018952889574898612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1024,32,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,12288,0.04858044452137417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,8192,0.12135199705759685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,1536,0.0064533332155810455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,1536,0.017768000562985737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,7168,0.16866666740841338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,12288,0.06587022542953491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,1024,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,65536,0.5895582305060493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,10240,0.14851555559370253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,1024,0.02404533326625824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,7168,0.03133777777353922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,10240,0.03437422381507026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,768,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,2560,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,512,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,128,0.06930489010281034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,128,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,2560,0.07760977745056152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,65536,0.056779556804233126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,2048,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,2048,0.03139466709560818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,2560,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,65536,0.35292532708909774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,5120,0.1233946614795261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,2048,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,2048,0.01575555569595761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,6144,0.019946666227446664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,1536,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,8192,0.013196444345845116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,512,0.013404444687896304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,51200,0.0507004459698995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,256,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,3584,0.7981342209710015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,6144,0.03173955612712436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,16384,0.03834399912092421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,3072,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,128,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,6144,0.02743822170628442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,7168,0.10220711098776923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,3072,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,768,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,3072,0.10700711276796128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,8192,0.09237511290444268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,65536,6.329601711697048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,3584,0.7248782051934136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,768,0.026016889346970454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,128,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,512,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,768,0.012665777570671506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,3072,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,12288,0.03801511062516107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,4096,0.8955048455132378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,51200,0.06827288866043091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,2560,0.026890666948424444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,3072,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,2048,0.0488337775071462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,2560,0.015298666225539314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,16384,0.04914044340451559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,256,0.0679137772983975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,4096,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,10240,0.02916533417171902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,6144,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,16384,0.034994665119383074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,3072,0.004571555389298333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,256,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,2560,0.057612445619371205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,12288,0.01737333337465922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,2048,0.02198933396074507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,8192,0.2634604507022434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,128,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,3072,0.007840889195601145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,32,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,65536,0.1438631084230211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,65536,0.27907466888427734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,65536,0.44808530807495117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,128,0.018216888109842937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,64,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,1024,0.007096000015735626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,51200,0.3216808901892768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,1024,0.031790223386552595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,51200,0.3904000123341878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,512,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,2048,0.029602666695912678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,10240,0.013061333033773633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,64,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,16384,0.04675288995107015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,2560,0.014767999450365702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,768,0.006035555568006303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,10240,0.02753777801990509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,768,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,1536,0.5528737703959147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,8192,0.010701333483060202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,768,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,3584,0.12506043910980225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,256,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,1536,0.11660888459947373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,2560,0.008362666600280339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,12288,0.035651554663976036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,7168,0.023638221952650283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,2048,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,768,0.05999288956324259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,65536,0.3924986786312527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,6144,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,1536,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,1024,0.018201778332392376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,32,0.00423555572827657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,65536,0.06787377595901489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,6144,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,6144,0.08419377936257257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,8192,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,32,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,6144,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,65536,3.0359325408935547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,1024,0.03208533260557387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,5120,0.12347555160522461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,1536,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,2048,0.02881866693496704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,128,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,6144,0.1282364394929674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,512,0.012282667060693106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,5120,0.04695733388264974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,16384,1.1954737769232857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,1536,0.018575999471876357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,5120,0.30411020914713544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,12288,0.21136089166005453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,12288,0.012475555141766867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,51200,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,2560,0.04223111271858215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,12288,2.340040842692057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,512,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,128,0.03151377704408433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,64,0.0021173333128293357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,16384,0.027705777022573683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,10240,0.04182577795452542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,256,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,3072,0.014354666074117025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,12288,0.09591111209657456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,7168,0.009891555541091496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,3584,2560,0.007117333511511485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,12288,0.028875556257035997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,4096,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,3584,0.013772444592581855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,8192,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,64,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,10240,0.09592355622185601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,12288,0.030631999174753826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,7168,0.02462133268515269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,3072,0.01904444396495819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,51200,0.8978630701700846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,65536,0.2481137646569146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,512,0.01679644485314687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3072,4096,0.02156977852185567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,2560,0.015912888778580558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,65536,9.522566901312933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,8192,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,256,0.010564444793595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,5120,0.1284524467256334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,7168,0.014360888136757744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,5120,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,5120,0.03356533249219259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,4096,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,768,0.11305688487158881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,64,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,3584,0.028869334194395278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,51200,0.10999111334482829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,2560,0.007993777592976889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,256,0.010847999817795224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,5120,0.09955555862850612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,5120,0.05537777807977465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,65536,1.0693688922458224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,12288,0.02052622205681271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2560,6144,0.01694222291310628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,5120,0.02896889050801595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,65536,1.2657883961995442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,32,0.008455110920800103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,1536,0.06472622023688422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,128,0.027054222093688116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,3584,0.0074240002367231585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,3072,0.04924622178077698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,3584,0.03661866651640998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,4096,0.04525244567129347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,512,0.044381333722008594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,1024,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,65536,0.29643821716308594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,7168,0.01070755554570092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,128,0.0021119999388853707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,3072,0.0074373334646224976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,6144,0.45752535925971133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,3072,0.020883555213610332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,2560,0.026884444885783728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,65536,3.297242694430881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,3584,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,256,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,16384,0.28584443198310006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,51200,0.1135724385579427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,10240,0.029502222935358684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,65536,5.501447041829427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,16384,0.05189511179924011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,2048,0.08785688877105713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,65536,0.19299199846055773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,12288,3.1278497907850475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,16384,0.021895999709765118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,1536,0.048900445302327476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,6144,0.06972266568077935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,16384,128,0.01718577742576599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,6144,0.02440622283352746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,65536,256,0.03233155608177185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,12288,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,5120,0.029767112599478826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,256,0.016504888733228047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,65536,0.06425422430038452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,512,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,7168,0.03654488921165466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,12288,2560,0.018580444984965853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,3072,0.044815109835730664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,1536,0.0773199995358785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,5120,0.011794666449228922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,2048,0.006735111276308696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,768,0.01792266633775499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,1024,0.020097777247428894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,51200,0.040080888403786555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,10240,0.04663377669122484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,2048,0.020782222350438435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,768,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,8192,0.007535999847782983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,10240,0.4500240219963922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,7168,0.5976035330030653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,12288,0.2563564512464735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,65536,0.140721771452162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,768,0.004298666699065102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,1536,0.0273964438173506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,768,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,3072,0.27733778953552246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,6144,0.05288266804483202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,12288,0.031699554787741765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,128,0.09535377555423313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,3072,0.024271999796231587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,2560,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,8192,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,5120,0.024513777759340074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,7168,0.05262044403288099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,1536,0.049406223826938205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,768,0.013832000394662222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,4096,0.011012444065676795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,1536,0.00700800038046307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,51200,0.09509511126412286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,3584,0.015477332803938123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,6144,0.009334222310119206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,768,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,10240,0.06849422057469685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,12288,0.3385813236236572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,65536,0.4654124577840169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,12288,0.03262666530079312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,51200,0.4982124434577094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,65536,4.67192628648546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,512,0.027464888162083093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,2048,0.4776017930772569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,65536,0.7381671269734701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,1536,0.07115555471844144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,3584,0.01370488852262497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,4096,0.02643288837538825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,256,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,128,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,8192,0.25574133131239146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,10240,0.11504622300465901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,2048,0.01091644416252772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,256,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,16384,0.23639822006225586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,128,0.01779022150569492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,65536,0.061306668652428516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,256,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,4096,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,12288,0.031501332918802895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,7168,0.043331556850009494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,512,0.011044444309340583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,2048,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,65536,0.35014666451348203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,256,0.01757688820362091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,10240,0.020596444606781006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,768,0.057189332114325635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,5120,0.02514133354028066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,8192,0.26676710446675617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,3584,0.06825155682033963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,32,0.005389333185222413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,10240,0.06388444370693631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,12288,0.12065422534942627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,5120,0.03881333271662394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,128,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,4096,0.7014791170756022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,6144,0.029584887954923842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,128,0.011527111132939657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,4096,0.0166631109184689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,128,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,2048,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,51200,0.07743200328614977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,64,0.0024497778051429326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,5120,0.018431999617152743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,5120,0.06499999761581421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,5120,0.0517537792523702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,512,0.1287271049287584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,2048,0.029459555943806965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,128,0.022600000103314716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,5120,0.04783466789457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,7168,0.027298667364650305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,6144,0.03170311119821336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,7168,0.014866666661368476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,2560,0.24216977755228677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,1024,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,3072,0.17395022180345324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,1536,0.02401955591307746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,3584,0.13143732812669542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,65536,0.08988355265723334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,12288,0.036256889502207436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,1536,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,2560,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,64,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,65536,1.0938409169514973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,16384,0.15844089455074734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,10240,0.018464000688658822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,4096,0.024338665935728285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,2048,0.010019555687904358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,4096,0.01885333326127794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,768,0.025631111529138353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,65536,0.07381333245171441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,6144,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,10240,0.012287110918098025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,4096,0.10972622368070815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,512,0.016528889536857605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,2560,0.26821777555677623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,256,0.01055111073785358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,65536,0.5593973265753852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,1536,0.012774222426944308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,6144,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,16384,0.036914666493733726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,2048,0.014919999572965833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,2048,0.006825777805513806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,6144,0.02940533227390713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,51200,0.0606222218937344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,51200,0.05241511265436808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,65536,0.1407182216644287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,2048,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,64,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,768,0.0035093331502543557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,3072,0.05760177638795641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,5120,0.025774222281244066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,8192,0.10496266682942708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,7168,0.02258577777279748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,10240,0.20220088958740234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,8192,0.2795164320203993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,8192,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,768,0.04445777667893303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,3072,0.008360000120268928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,1536,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,3584,0.007789333661397298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,2560,0.02551377813021342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,64,0.031255112753974065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,128,0.022276444567574397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,10240,0.024101333485709295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,4096,0.022034666604465906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,256,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,16384,0.053256889184316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,1536,0.00499288903342353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,51200,0.04598133431540596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,512,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,768,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,51200,0.15189954969618055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,5120,0.0788488917880588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,768,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,128,0.006063111126422882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,2560,0.36601866616143125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,51200,0.27518044577704537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,10240,0.021703999903466966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,128,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,5120,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,256,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,1536,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,4096,0.008712000317043727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,16384,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,65536,0.39379111925760907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,1024,0.004392000122202767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,65536,0.16605956024593777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,2560,0.026576000783178542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,512,0.005424889011515512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,1536,0.026272889640596177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,51200,0.0669599970181783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,64,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,6144,0.04293955696953667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,128,0.024636444118287828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,6144,0.178765336672465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,16384,2.249293433295356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,1536,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,1024,0.011759999725553723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,4096,0.027806222438812256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,65536,2.022539562649197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,1024,0.04370577798949348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,128,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,65536,0.13689066304100886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,7168,0.020258666740523446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,12288,0.015187554889255099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,7168,0.01854044364558326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,512,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,12288,0.38817421595255536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,3584,0.008808888494968414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,128,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,1024,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,8192,0.013752889302041797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,256,0.044032000833087497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,4096,768,0.03013066781891717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,2048,0.015455999308162265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,16384,0.13751467068990073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,1536,0.0038622220357259116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,10240,0.029345777299669053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,32,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,32,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,4096,0.02437244521247016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,1024,0.027719111906157598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,1536,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,1024,0.024759110477235582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,128,0.01718933383623759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,256,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,1024,0.07331822315851848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,768,0.030385779009924993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,7168,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,1024,0.07580444547865126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,10240,0.017383999294704862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,5120,0.059333331055111356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,12288,0.027127110295825537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,10240,0.016128889388508268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,5120,0.26835910479227704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,2048,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,12288,0.035107556316587664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,8192,0.0705439978175693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,128,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,32,0.008175111479229396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,16384,0.018622222873899672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,12288,0.046394666035970054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,6144,0.03032355507214864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,1536,0.016903999778959487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,16384,0.20345689190758598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,32,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,51200,0.11030489206314087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,3072,0.010991111397743225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,10240,0.10364266898896958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,3584,0.011020444333553314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,7168,0.03413866625891791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,2560,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,7168,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,2560,0.013902222116788229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,3072,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1536,51200,0.0503111117415958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,256,0.010500444306267632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,3072,0.0748737785551283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,3072,0.015968888998031616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,4096,0.0296382208665212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,768,0.006732444382376141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,4096,0.018941332896550495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,128,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,128,0.05073688758744133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,6144,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,65536,1.3114657931857638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,512,0.004159111115667555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,5120,0.10536266697777642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,5120,0.009328888522254096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,768,0.006298666612969504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,256,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,1024,0.004376000000370873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,3584,0.019909333851602342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,2048,0.02370311154259576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,5120,0.008599999878141616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,2560,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,64,0.021526222427686054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,51200,0.05110311177041796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,10240,0.24400178591410318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,10240,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,256,0.013452444639470843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,8192,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,32,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,256,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,65536,0.24250578880310059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,65536,0.638214217291938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,64,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,5120,1.1736141840616863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,128,0.012423111332787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,3072,0.007484444313579136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,12288,0.056848002804650195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,5120,0.011334222224023608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,5120,0.013402666482660504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,7168,0.6895564397176107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,3584,0.030576888057920668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,3072,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,512,1536,0.03102488981352912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,512,0.03746844331423441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,8192,0.012142221960756512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,3072,0.024490666058328416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,768,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,1024,0.016157334049542744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,3072,0.18126400311787924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,12288,0.01796444422668881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,768,0.007762666377756331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,256,0.015020444989204407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,3072,0.062096887164645724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,10240,10240,0.056314667065938316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,4096,0.008570666942331526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,3072,0.023971555961502924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,1536,0.010341333018408881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,3584,0.01668177710639106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,7168,0.010112000008424124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,1024,0.0058782220714622075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,7168,0.10363466872109307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,65536,0.06110044320424398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,3584,0.013306666579511432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,512,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,3584,0.024280889166726008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,16384,0.03910400138960944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,512,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,6144,0.04889600144492256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,128,0.018775110443433125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,6144,0.017216889394654166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,16384,0.08938044309616089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,2048,0.12209867106543647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,12288,0.023994666006830003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,128,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,32,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,1536,0.008256889051861232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,5120,0.03985866573121812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,6144,0.024681778417693243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,65536,0.1428026623196072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,512,0.00543200017677413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,2560,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,6144,0.1466088824801975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,16384,0.5800586806403266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,16384,0.06968622075186835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,5120,0.041850666205088295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,16384,0.042917334371142916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,128,0.11060622003343369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,65536,0.009038222332795462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,256,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,5120,0.0285173323419359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,128,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,1536,0.02198222279548645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,12288,0.034661332766215004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,768,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,16384,0.15517866611480713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,12288,0.023306666149033442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,768,0.0140337778462304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,32,0.005056000004212062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,7168,0.017476444443066914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,12288,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,3072,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,7168,0.042713777886496655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,10240,0.011788444386588203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,6144,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,6144,0.011033777561452655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,3584,0.13055466281043157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,2048,0.009523555636405945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,2560,0.02017333275742001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,5120,0.09144622087478638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,2560,0.0063982220987478895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,256,0.02930755416552226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,256,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,256,0.011021333436171213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,10240,0.040948443942599826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,1024,0.027573333846198186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,1536,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,768,0.006227555374304454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,6144,0.011357333097192975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,5120,0.10638311174180771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,7168,0.11943288644154866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,1536,0.02407377792729272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,65536,0.04987733231650459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,16384,3.0643626319037547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,51200,0.13689866330888537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,8192,0.07843466599782307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,2048,0.004603555632962121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,512,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,4096,1.0054435729980469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,8192,0.02453600035773383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,7168,0.18722489145067003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,6144,0.8382959895663791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,51200,1.8505689832899306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,512,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,5120,0.032363557153277926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,3072,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,256,0.005403555515739653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,4096,0.12238400512271458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,51200,0.29744354883829754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,65536,1.0455030865139432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,256,0.018333334061834548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,10240,0.19313511583540174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,256,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,512,0.016950221525298226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,12288,0.10229244497087266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,6144,0.021664000219768945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,2560,0.007541333635648091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,128,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,3072,0.0612693362765842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,65536,0.2722871038648817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,2048,0.06526488727993436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,64,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,128,0.006567111031876669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,3584,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,12288,0.16792444388071695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,256,0.017982222967677645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,10240,9.820849948459202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,51200,0.07018133004506429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,256,0.023668444818920557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,2048,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,1536,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,8192,0.4796746571858724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,3584,0.01716711123784383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,3584,0.04075555668936835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,7168,0.2718506654103597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,128,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,4096,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,1024,0.2213431066936917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,64,0.010847110715177325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,3584,0.010228444304731157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,3584,0.02325511144267188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,51200,0.9831057654486762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,16384,0.052821334865358144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,32,0.010016889207892949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,3584,0.021944888763957556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,1024,0.007709333466158972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,51200,0.6070951355828179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,4096,0.07862044705284967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,768,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,1024,0.046257777346505054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,3584,0.018206222189797294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,12288,0.034281777011023626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,12288,0.014477334088749357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,7168,0.02232622272438473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,128,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,512,0.010085333552625444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,128,0.02947910957866245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,512,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,10240,0.012367111113336353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,6144,0.6253280109829372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,3072,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,3072,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,6144,0.0417795545525021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,6144,0.07137155532836914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,256,0.02971822354528639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,8192,0.010764444039927589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,12288,0.049970666567484535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,128,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3584,7168,0.012316444681750404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,10240,0.025715554753939312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,2048,0.005231110999981563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,65536,4.4904908074273004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,8192,0.09377600087059869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,12288,0.07773422532611422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,8192,0.07418488793902926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,2048,0.009815110928482479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,7168,0.1193973355823093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,3072,0.02258311046494378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,51200,177.74545627170139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,16384,0.020285333196322124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,3072,0.07656355698903401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,2560,0.006763555523422029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2560,6144,0.011239111423492432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,4096,1.1072568893432617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,8192,0.016138666205936007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,2048,0.0618399977684021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,256,0.014456889695591398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,2048,0.11218221982320149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,10240,0.17667733298407662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,16384,0.027228444814682007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,512,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,1536,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,1536,0.017851556340853374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,512,0.01688888834582435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,10240,0.042039109600914844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,32,0.012664889295895895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,1024,0.008619555168681675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,512,0.026271111435360376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,51200,0.04160977734459771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,12288,0.05113600028885735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,256,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,1536,0.02419022222359975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,1024,0.0069982219073507525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,7168,0.009219555391205682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,256,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,8192,0.21553866068522134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,3584,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,51200,0.09880622227986653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,256,0.014757333530320061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,7168,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,128,0.022309333086013794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,3072,0.0053582220441765254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,2048,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,8192,0.4008275667826335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,64,0.0018657777044508194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,5120,0.01742577718363868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,1024,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,1024,0.01183999992079205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,256,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,3072,0.007608888877762689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,12288,0.02846755584081014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,3072,0.017247110605239868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,1536,0.8800640106201172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,65536,0.1436835527420044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,1536,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,32,0.0018835556175973679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,8192,0.028318223026063707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,65536,0.0671164459652371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,1024,0.03496711121665107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,6144,0.027222222752041284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,10240,0.12378488646613227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,128,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,2560,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,4096,0.016519111063745286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,4096,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,4096,0.03333244389957852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,1536,0.007912000020345053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,51200,0.0665644473499722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,32,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,6144,0.024448888169394598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,1024,0.003519999898142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,6144,0.018566222654448617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,256,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,512,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,1024,0.003555555724435382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,3072,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,10240,0.02555911077393426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,12288,0.07962133487065633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,512,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,2560,0.15483200550079346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,10240,0.13315644529130724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,51200,3.0368514590793185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,512,0.08194488949245878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,8192,0.03090488910675049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,12288,0.0528053343296051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,5120,0.3546008798811171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,1024,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,65536,0.3205457793341743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,4096,0.020228443874253165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,51200,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,512,0.010495999621020423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,51200,0.2512799898783366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,128,2560,0.03439999951256646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,768,1536,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,6144,0.01053066634469562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,16384,0.13104266590542263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,65536,4.429341210259332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,4096,1.0091191397772896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,51200,0.07869778076807658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,128,0.015811555915408664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,3072,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,51200,0.12384355068206787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,768,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,10240,0.02755822241306305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,64,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,10240,0.02903555499182807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,8192,128,0.00884622252649731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,4096,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,3072,0.07720800240834554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,16384,0.028049776951471966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,6144,0.07017866770426433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,256,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,256,0.014856888188256158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,10240,0.014354666074117025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,12288,0.057077334986792676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,12288,0.1363582213719686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,6144,0.013164444102181328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,51200,0.12547732724083796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,1024,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,512,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,768,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,64,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,256,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,3584,0.1345697773827447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,1536,0.04749155706829495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,32,0.02370044423474206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,8192,0.13856532838609484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,3584,0.050063110060162015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,10240,0.0920142200258043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,10240,0.034789333740870156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,3584,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,2048,0.024951110283533733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,32,0.009099555512269339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,768,0.008627555436558193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,3584,0.029881778690550063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,16384,0.034789333740870156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,3584,0.011100444528791638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,256,0.014918223023414612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,1024,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,7168,0.10532710949579875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,16384,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,4096,0.101583997408549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,2048,0.0447599987188975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3584,2048,0.021962665849261816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,512,0.006014222072230444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,3584,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,512,0.005981333139869902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,1536,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,51200,0.5173990991380479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,768,0.012401777837011548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,8192,0.03107200066248576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,16384,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,4096,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,256,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,51200,0.05696622530619303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,1024,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,3072,0.02648266653219859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,32,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,2560,0.10755111111534967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,5120,0.2788951132032606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,1536,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,3072,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,2048,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,8192,0.014373333917723762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,16384,0.23562754525078666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,1536,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,256,0.007728888756699032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,512,0.005966222120655908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,1536,0.0074026667409473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,5120,0.08851199679904515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,51200,0.02327377763059404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,3072,0.008763555851247575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,768,0.03555644551912943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,16384,0.5454151365492079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,10240,2.1721519894070096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,6144,0.3325057824452718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,512,0.006990222467316522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,12288,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,512,51200,0.045744889312320285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,10240,0.0359173317750295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,4096,0.14579555723402235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,128,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,16384,0.015800888339678448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,65536,0.15614755948384604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,7168,0.3037857744428846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,768,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,128,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,5120,0.02714666724205017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,65536,2.066031985812717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,12288,0.08479466703202988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,1536,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,768,0.014331556028789945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,1536,0.09519200192557441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,32,0.012829333543777466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,51200,0.5751973258124458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,7168,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1024,6144,0.024033778243594702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,1024,0.5847751299540201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,6144,0.17555556032392713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,7168,1.3342435624864366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,51200,0.7494115299648709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,256,0.042281776666641235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,8192,0.24213867717319063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,6144,0.012837332983811697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,3584,0.02090577781200409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,3584,0.009191110730171204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,12288,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,128,0.01516711049609714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,10240,2.9764196607801647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,768,0.022294221652878657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,16384,0.022963555322753057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,51200,3.025849872165256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,4096,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,1536,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,51200,0.05846755372153389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,512,5120,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,3072,0.06401155392328899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,65536,0.20108444160885283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,16384,0.10501155588361953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,256,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,51200,0.9392639795939127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,16384,0.016064888901180692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,256,0.023227555884255305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,2048,0.08907288975185818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,2048,0.13607199986775717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,2048,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,2560,0.04360266526540121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,1024,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,7168,0.03259377678235372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,1024,0.014172444740931192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,2048,0.009661333428488838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,12288,0.26761865615844727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,51200,0.08355110883712769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,65536,0.016013332539134555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,4096,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,10240,0.04771555463473002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,768,0.027672888504134283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,4096,0.00853066643079122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,2560,0.010523555179437002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,768,0.020945777495702107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,3584,0.015882665912310284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,65536,0.6896542443169488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,3584,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,51200,0.04893777767817179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,51200,0.25642045338948566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,256,0.02184444500340356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,16384,0.04166933231883579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,65536,0.06402755445904203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,6144,0.013890667094124688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,128,0.09637244542439778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,1024,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,128,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,16384,0.06939378049638537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,1024,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,2560,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,1536,0.00958933350112703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,2048,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,256,0.00612088872326745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,7168,0.03003022074699402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,16384,0.09410488605499268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,512,0.012311110893885294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,1536,0.004535111288229625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,768,0.009690666364298927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,10240,0.04678488771120707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,10240,0.012201777762836881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,7168,0.12296799818674724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,1536,0.010467555787828235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,8192,0.0781715578503079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,1024,0.0314977765083313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,3584,0.13811288939581975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,2048,0.24796801143222383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,1024,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,768,0.01292622254954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,16384,0.13816800382402208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,3072,0.007177777588367462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,2048,0.007256000406212277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,6144,0.057144886917538114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,128,0.01922044489118788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,8192,0.03297777639495002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,51200,0.22381422254774305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,4096,0.028132445282406274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,768,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,8192,0.03317955467436049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,3584,0.02034311162100898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,65536,0.13983466890123156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,3584,0.6322791311475966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,12288,0.23412977324591744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,512,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,768,0.011016888750924004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,1024,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,51200,8192,0.07848533656862047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,16384,0.04576711191071404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,1024,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,3584,0.010392888552612728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,3072,0.037117332220077515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,512,0.012737777498033313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,7168,0.6734959814283582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,8192,0.030136889881557886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,51200,0.12515555487738714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,5120,0.01886311173439026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,256,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,16384,0.03853866789076064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,8192,0.028667555914984807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,3072,0.019116444720162284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,512,0.010052444206343757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,4096,0.02604888876279195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,51200,0.6139884524875218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,256,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,4096,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,512,0.03683822353680929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,1024,0.004979555391603046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,7168,0.027181333965725366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,512,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,768,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,3072,0.11727021800147162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1024,5120,0.03221333358022902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,128,0.008276444342401292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,1536,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,7168,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,768,0.012740444805887012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,12288,0.01923022170861562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,10240,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,7168,0.03631110986073812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,16384,0.015251555376582675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,16384,0.08128977484173246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,128,0.008311111066076491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,51200,0.08836088577906291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,3584,0.02106133268939124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,8192,0.16144800186157227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,2048,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,4096,0.014840889308187695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,6144,65536,0.12108977635701497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,1024,0.018179555733998615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,1536,0.0041857775714662345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,4096,0.03230399886767069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,3072,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,2048,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,7168,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,768,0.017640888690948486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,2560,0.013031110995345645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,5120,0.009710222482681274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,10240,0.20968266328175864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2048,6144,0.009499555660618676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,256,0.01537333263291253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,1024,0.0046640001237392426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,512,0.05148622062471178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,64,0.01312711089849472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,4096,0.09607199827829997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,256,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,2560,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,7168,0.1655742194917467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,6144,0.011166222393512726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,16384,0.05236266718970405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,64,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,768,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,1536,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,8192,0.05160444312625461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,10240,0.057483553886413574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,2048,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,7168,0.5643119812011719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,2560,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,1024,0.013069333301650153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,512,0.05750222338570488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,7168,0.09434844387902154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,8192,0.05229155553711785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,1536,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,65536,0.06493066416846381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,32,0.037823110818862915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,2048,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,512,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,51200,0.5056062274509007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,65536,0.09266844722959731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,32,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,7168,0.011391999820868174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,6144,0.02456533412138621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,10240,0.017821333474583097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,6144,0.028190222051408555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,4096,0.028601778878106013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,768,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,3584,0.029409776131312054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,6144,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,32,0.004928888960017098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,10240,0.026473777161704168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,32,0.0028568889117903183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,256,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,6144,0.02365155518054962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,16384,0.01852888862291972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,51200,3.251649856567383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,4096,0.20889066325293648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,5120,0.017186666528383892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,512,0.0141004439857271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,5120,0.10262222422493829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,6144,0.03685599896642897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,65536,0.17733689149220785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,12288,0.02854222059249878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,65536,2.3337217966715493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,256,0.03359999921586778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,1024,0.039489779207441546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,3584,0.018519111805491976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,7168,0.05342844459745619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,2048,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,12288,1.8370496961805556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,3072,0.008760000268618265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,2048,0.006903999795516332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,128,0.00924177798959944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,8192,0.1775439977645874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,16384,0.020337777005301583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,7168,0.013412444127930535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,12288,4096,0.015014222926563688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,65536,0.08280355400509305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,2560,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,1536,0.5851715405782064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,51200,0.6057502428690592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,6144,0.7094906700981988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,2048,0.008808888494968414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,10240,0.02053155501683553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,32,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,768,0.018254222141371835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,128,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,2048,0.01146044416560067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,6144,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,12288,0.057710223727756076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,2048,0.02681244413057963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,8192,0.14065689510769314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,4096,0.010203555226325989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,512,0.012822222378518848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,16384,0.05414666732152303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,3584,0.012739555703269111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,16384,0.05249422126346164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,65536,0.6086648835076226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,4096,0.17305244339836967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,16384,0.030583110120561387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,16384,0.26312979062398273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,3584,0.007600888609886169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,1024,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,64,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,3072,0.0533591111501058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,3584,0.008046222229798635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,7168,0.0292559994591607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,10240,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,128,0.00655377780397733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,1024,0.0038720000949170855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,2048,0.06598755386140612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,256,0.002769777758253945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,3072,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,8192,0.7224888801574707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,7168,0.01993600030740102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,768,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,2560,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,2560,0.3367111153072781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,10240,0.3754568894704183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,32,0.027789334456125896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,2048,0.017826666434605915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,7168,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,7168,0.19377599822150338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,3584,0.038768887519836426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,768,0.015471110741297403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,51200,0.04225511021084256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,1536,0.08574755324257745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,3072,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,256,0.006133333262470033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,16384,0.027764444549878437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,65536,0.37912800576951766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,128,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,51200,0.33537689844767254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,16384,2.0498621198866105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,2560,0.038698666625552706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,2560,0.04278044568167793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,256,0.0031840000301599503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,256,0.012465777496496836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,3584,0.5583270920647515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1536,1536,0.007199111084143321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,10240,7168,0.03840800126393636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,512,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,12288,0.07255555523766412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,3072,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,16384,0.26187200016445583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,8192,0.030234667989942763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,64,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,12288,0.08426666922039455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,3072,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,1536,0.006259555617968242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,1024,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,256,0.002324444345302052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,32,0.021681777305073206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,512,0.006666666517655055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,3584,0.007689778175618913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,256,0.016178665889634024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,3584,0.018174222773975797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,3584,0.007044444481531779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,512,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,256,0.007282666862010956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,10240,0.06432533264160156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,32,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,8192,0.013464888764752282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,512,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,7168,0.010327111515733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,512,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,65536,0.10267111327913071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,3584,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,2048,0.020396444532606337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,2048,0.005384888913896348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,12288,0.02979733215437995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,51200,24.878602769639755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,1024,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,1536,0.0206631107462777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,2560,0.013056889176368713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,128,128,0.0019884444773197174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,51200,0.1868924432330661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,768,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,128,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,512,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,1536,0.020773332979944017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,16384,0.19411110877990723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,256,0.019890666007995605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,12288,0.03775733378198411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,768,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,65536,0.016302223006884258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,2560,0.0273964438173506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,2048,0.024175110790464614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,3072,0.010276444256305695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,65536,4.144744873046875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,7168,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,12288,0.05340444379382663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,16384,0.054216000768873424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,65536,3.014675564236111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,10240,0.2215262254079183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,256,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,6144,0.09873333242204453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,1024,0.05094666613472832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,1536,0.03080355458789402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,8192,0.0519644452465905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,51200,0.8871484332614474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,51200,0.10090933243433635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,6144,0.01943733294804891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,4096,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,512,0.07884444130791558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,12288,0.023988443944189284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,3584,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,12288,0.03570311268170675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,4096,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,12288,0.05163910984992981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,10240,0.10238489177491929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,1536,0.0047022220161226065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,4096,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,32,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,512,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,1536,0.0064382221963670515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,64,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,3584,0.029834667841593426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,4096,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,2048,0.010142222046852112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,4096,0.05946222278806898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,10240,0.02535199953450097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,8192,0.04645422101020813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,10240,0.012542222109105853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,4096,0.023375110493765935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,16384,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,2560,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,6144,0.010214221974213919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,6144,0.024455111887719896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,1024,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,6144,0.0373306671778361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,10240,0.011761777930789523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,8192,0.022111111217074927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,256,0.030360887447992962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,256,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,1024,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,10240,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,512,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,5120,0.0199644449684355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,1024,0.08322666751013862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,3584,0.03053511182467143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,7168,0.41884178585476345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,1536,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,512,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,8192,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,2048,0.0820559991730584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,2048,0.03916355636384752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,6144,0.009395555489593083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,1024,0.007104888558387756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,7168,0.028127110666698877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,2048,0.015121777852376303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,2560,0.08238310946358575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,8192,0.05273422267701891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,1024,0.018976888722843595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,1536,0.004586666822433472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,256,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,6144,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,3584,0.06040710873074002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,5120,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,12288,2.059326171875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,32,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,512,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,1536,0.034895999564064875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,1536,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,16384,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,4096,0.04589777853753832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,128,0.01477777792347802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,64,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3584,2560,0.03218577636612786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,768,0.006186666587988536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,7168,0.012251555919647217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,8192,3.143992953830295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,51200,0.15993066628774008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,4096,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,12288,0.1638151142332289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,16384,0.034661332766215004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,1024,0.017823111679818895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,512,0.01296622223324246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,3584,0.03427377674314711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,8192,0.029483556747436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,4096,0.016799110505315993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,12288,0.0361697773138682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,256,0.054358220762676664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,256,0.00217333332531982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,51200,0.054699556695090405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,256,0.011574221981896294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,768,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,65536,0.4578542179531521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,512,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,3072,0.022108443909221228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,768,0.008622222476535374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,768,0.01647111111217075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,4096,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,256,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,512,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,256,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,1536,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,8192,0.11980622344546848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,3072,0.03492088781462775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,256,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,1536,0.023112888137499493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,64,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,2048,0.007133333219422235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,3072,0.01668444441424476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,1024,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,512,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,2048,0.019262222780121695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,8192,0.024718221690919664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,65536,0.14048266410827637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,8192,0.011030221978823343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,6144,0.05754488706588745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,32,0.1076800028483073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,16384,0.4399537775251601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,128,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,3072,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,1024,0.030039111773173015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,64,0.011737777955002256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,1024,0.013764444324705334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,65536,0.13392355706956652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,512,0.012544000314341651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,51200,0.04818577898873223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,8192,0.025103110406133864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,3072,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,256,0.02784622377819485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,51200,0.16889511214362252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,3072,0.00796088907453749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,128,0.0028764444092909494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,1024,0.0038568890757030914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,3072,0.03188622328970168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,12288,0.3928862147861057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,5120,0.016415999995337594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,512,0.01346577786737018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,16384,0.026124444272783067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,16384,0.018127111925019156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,32,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,5120,0.010520888699425591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,12288,0.13086044788360596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,2048,0.023159111539522808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,51200,0.2946631113688151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,512,0.032481776343451604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,512,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,256,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,32,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,8192,0.1058079997698466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,2048,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,768,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,12288,0.004922666483455234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,256,0.026849778162108526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,10240,0.01256444470749961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,7168,0.014373333917723762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,10240,0.046294222275416054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,128,512,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,4096,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,10240,0.06943644417656793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,2560,0.011674666570292579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,65536,0.6882506476508247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,12288,0.050740443997912936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,512,0.005516444643338521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,512,0.0979057749112447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,128,0.020440000626775954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,768,0.020836444364653692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,51200,0.40726221932305234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,256,0.025057777762413025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,16384,0.629876454671224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,512,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,512,0.008002666963471307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,32,0.005403555515739653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,256,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,12288,0.0912826657295227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,3072,0.18326311641269258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,16384,0.027124444643656414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,12288,0.9262515174018012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,16384,0.11076888773176406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,7168,0.0693253344959683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,5120,0.5893093215094672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,64,0.0224906669722663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,5120,0.027290667096773785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,7168,0.07576444413926867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,5120,0.027388445205158655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,2560,0.01686844395266639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,256,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,4096,0.09364533424377441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,1024,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,768,0.013960889644092984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,3072,0.019676445258988273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,1024,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,8192,0.03759288787841797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,8192,0.1690035528606839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,10240,0.05544266766972012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,256,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,3072,0.0074417781498697065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,7168,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,128,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,12288,0.16213777330186632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,6144,0.02193244463867611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,7168,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,2048,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,3072,0.031545778115590416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,3584,0.006236444330877728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,4096,0.3278026580810547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,1536,0.20173777474297416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,3072,0.010464888479974536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,3584,0.02128977742460039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,3072,0.49947378370496964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,256,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,1536,0.04036088784535726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,6144,0.015551111764378019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,7168,0.23800977071126303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,5120,0.0184542222155465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,1536,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,51200,0.0625928905275133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,2048,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,6144,0.027794665760464136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,512,0.0217848883734809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,16384,0.10144444306691487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,1024,0.015458666616015963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,7168,0.15661956204308405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,6144,0.009085333181752099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,6144,0.02130311065249973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,128,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,5120,0.013159111142158508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,8192,0.13943289385901556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,2560,0.1059546669324239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,1024,0.008656000097592672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,128,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,64,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,3584,0.07982310983869764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,4096,0.008215111162927415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,6144,0.019523556033770244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,12288,0.0162773331006368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,7168,0.1224764453040229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,6144,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,7168,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,1024,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,256,0.030823998981051978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,6144,0.05330488748020596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,12288,0.20703111754523384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,16384,0.029300444655948218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,2560,6.3730256822374125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,768,0.028263999356163874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,4096,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,65536,0.09370844231711493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,7168,0.04366488920317756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,6144,7168,0.012495111260149213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,1536,0.0188737776544359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,3584,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,2560,0.007245333658324347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,512,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,32,0.011416888899273343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,10240,0.059736000166998975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,6144,0.03280355532964071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,6144,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,2048,0.025872000389628943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,8192,0.042230221960279674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,16384,0.040592888991038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,4096,0.014182221558358936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,128,0.008023111356629265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,8192,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,32,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,16384,0.30082933108011883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,4096,0.12059999836815728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,128,0.01791111131509145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,2560,0.00572533326016532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,51200,0.05940710835986667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,12288,0.14168711503346762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,3072,0.02534755567709605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,512,0.0034933334423436057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,10240,0.024107555548350017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,8192,0.011552888486120435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,1536,0.023844444089465674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,2560,0.01961777773168352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,16384,0.9847688674926758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,8192,0.14962044027116564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,2048,0.6982986662122938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,3072,0.048282666338814624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,10240,0.0613280004925198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,32,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,10240,0.10182844267951117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,5120,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,12288,0.017758222089873422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,8192,0.018536888890796237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,64,0.006321777900060018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,3072,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,1536,0.020985777179400127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,51200,0.13001778390672472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,3584,0.017082666357358296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,3584,0.011852444046073489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,65536,0.07855377594629924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,768,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,16384,0.22677956687079537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,4096,0.45656977759467227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,7168,0.48254934946695965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,8192,0.03796799977620443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,512,0.012851555314328937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,256,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,256,0.008344000412358178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,2560,0.028775112496482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,512,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,5120,0.014064888159434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,12288,0.021200888686709937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,3072,0.02804533309406704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,3072,0.0166275550921758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,3584,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,5120,0.008459555606047312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,1536,0.006513777706358168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1536,256,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,51200,0.04942755566702949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,65536,0.07226044601864286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,65536,0.6849084430270724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,4096,0.009343110852771336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,51200,0.026029333472251892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,1024,0.022665777140193518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,2048,0.013724444641007317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,2048,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,12288,0.07258577479256524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,7168,0.19228266345130074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,12288,0.10583288802040948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,2048,0.39403377638922793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,512,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,7168,0.02425244450569153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,10240,0.13215466340382895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,16384,0.1726222170723809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,7168,0.10097333457734849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1536,512,0.010399999717871347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,8192,0.10818311240937974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,64,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,7168,0.1404515504837036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,65536,0.05676888757281833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,512,0.009337777892748514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,32,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,16384,0.018910222583346896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,16384,0.018589332699775696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,6144,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,2048,0.01703466640578376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,768,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,512,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,12288,0.10416444142659505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,768,0.02751200066672431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,3072,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,51200,0.6315617561340332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,3072,0.016951999730534024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,2048,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,8192,0.07857955826653375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,768,0.013610666824711693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,6144,0.015574221809705099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,128,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,256,0.0052826665341854095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,2048,0.004206222378545337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,512,0.008943999807039896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,12288,0.07009689013163249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,256,0.01550222271018558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,1024,0.00904977818330129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,12288,0.0243484444088406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,16384,0.07565422190560235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,768,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,12288,0.050613333781560264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,3584,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,65536,0.20037421915266249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,2560,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,65536,0.20889422628614637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,8192,0.024451555477248296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,1536,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,1536,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,64,0.09477333227793376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,8192,0.01110400011142095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,6144,0.004583999928500918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,3072,0.037393778562545776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,768,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,2560,0.028774221738179524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,7168,0.350309345457289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,512,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,64,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,5120,6.2353931003146705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,5120,0.16159378157721624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,8192,0.2536657651265462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,256,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,6144,0.019485332899623446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,3584,0.017575111654069688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,65536,0.11623911062876384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,7168,0.03900444507598877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,768,0.01891288823551602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,51200,0.06866133213043213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,5120,0.01202222208182017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,3584,0.08391111426883274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,2048,0.015400888191329109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,8192,0.020986666282018025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,1536,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,256,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,3584,0.008205333517657386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,2048,0.014862222803963555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,3072,0.015341333217091031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,128,0.01819644371668498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,768,0.004288889053795072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,51200,0.13060088952382407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,6144,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,3584,0.008934222161769867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1024,8192,0.02126311096880171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,3072,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,1024,0.015103111664454142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,1024,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,7168,0.02621511121590932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,16384,32,0.02639644510216183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,4096,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,3072,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,768,0.009687110781669617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,3072,0.03806311223242018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,10240,0.08980000019073486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,1024,0.006069333189063602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,128,16384,0.06556977828343709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,8192,0.12781333923339844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,128,0.020503110355801053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,512,0.008117333054542542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,4096,0.008239111138714684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,4096,0.09023911423153347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,512,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,6144,0.42204266124301487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,512,0.024838222397698298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,64,0.01276711126168569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,128,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,512,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,12288,0.12282133102416992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,4096,0.1542311112085978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,512,0.026854222019513447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,512,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,4096,0.02514044443766276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,6144,0.03492444422509935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,8192,0.03618933426009284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,3584,0.3557448916965061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,65536,0.13172266218397352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,5120,0.015370666980743408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,512,0.006760889043410619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,6144,0.015056000815497505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,768,0.0191920002301534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,1024,0.008991111483838823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,2048,0.028071112102932397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,3584,0.13174577554066977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,512,0.005919111271699269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,10240,0.016227554943826463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,5120,0.6897004445393881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,6144,0.010103111465771994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,1024,0.07086666425069173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,1536,0.04044444362322489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,4096,0.07020444340176053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,3584,0.21100089285108778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,5120,0.00886933339966668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,4096,0.014873777826627096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,10240,0.03640000025431315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,8192,0.34180532561408145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,8192,0.018048889107174344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,51200,0.05717600054211087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,51200,0.1635635561413235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,64,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,768,0.011536888778209686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,8192,0.02514222264289856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,5120,0.01700444519519806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,65536,0.14333866702185735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,1024,0.06380889150831434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,6144,0.020098666350046795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,5120,0.037851555479897395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,64,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,4096,0.03837955660290188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,51200,0.03552088803715176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,512,0.0630319979455736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,65536,0.17651377783881295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,6144,0.01868711080816057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,256,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,3584,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,16384,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,512,0.021487111846605938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,128,0.004959110998445087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,768,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,2048,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,1536,0.028988444142871436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3072,3584,0.03159555461671617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,16384,0.015815110670195687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3072,6144,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,768,0.04347733325428433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,2048,0.03436888919936286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,2560,0.04242044356134203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,3584,0.12405155764685737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,512,0.02933244572745429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,3072,0.012479110724396177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,12288,0.040799998574786715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,51200,0.2590160104963514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,768,0.2261777718861898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,3072,0.08701866865158081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,10240,0.011176000038782755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,10240,0.02884711159600152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,64,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,512,0.02695200012789832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,10240,0.053026666243871055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,16384,0.009487999810112847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,8192,0.10706844594743516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,3584,0.03868177864286635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,12288,0.32576534483167857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,128,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,5120,0.01089777797460556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,3584,0.04860533277193705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,1024,0.20487466123369005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,8192,0.012827555338541666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,4096,0.05968799855973986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,8192,0.14647289117177328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1536,1536,0.017621333400408428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,5120,0.01014755583471722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,512,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,16384,0.031586666901906334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,5120,1536,0.01699022286468082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,12288,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,768,0.002936888900068071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,256,0.0029351111087534162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,12288,0.09800622198316787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,65536,0.05137688914934794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,3072,0.01236533373594284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,1024,0.024720888998773363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,8192,0.09205333391825359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,4096,0.020627554919984605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,768,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,3584,0.027722666660944622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,512,0.005658666706747479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,8192,0.03706222110324436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,8192,0.01038755559258991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,1024,0.027076444692081873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,8192,0.03427733315361871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,768,0.02276533345381419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,65536,0.05299999978807238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,65536,0.07095644209120008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,12288,0.12465688917371963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,65536,0.34847021102905273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,32,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,2560,0.016812443733215332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,512,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,1536,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,10240,0.7347857687208387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,7168,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,2560,0.019256888164414298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,32,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,3072,0.11567555533515082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,65536,0.08697955475913154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,768,0.02017333275742001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,3072,0.03211466802491082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,128,0.036891556448406644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,32,0.007659555309348636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,128,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,6144,0.020029332902696397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,2048,0.009873777627944946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,2048,0.03405688868628608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,3584,0.024697777297761705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,16384,0.016131555040677387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,256,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,3584,0.01202400028705597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,1024,0.04251377781232198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,6144,0.020477333002620272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,8192,1.4368764029608831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,12288,0.4742435349358453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,3584,0.027222222752041284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,8192,1.0894418292575412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,7168,0.35911199781629777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,1536,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,51200,0.18452621830834282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,10240,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,2048,1.5532809363471136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,51200,0.07012710968653361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,8192,0.0993280013402303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,4096,0.007992888490358988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,6144,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,3584,0.022494221727053326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3584,256,0.02156977852185567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,65536,0.05909866756863064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,256,0.016181333197487723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,51200,0.1498159964879354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,16384,0.6171021991305882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,512,0.05075644453366598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,256,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,768,0.0226248883538776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,1024,0.03569688730769687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,7168,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,2560,0.021000888612535264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,16384,0.7119448979695638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,65536,1.7364159690009222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,65536,0.7013271119859484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,5120,0.028000000450346205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,65536,3.4805609385172525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,8192,0.011056889262464313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,5120,0.10911289188596938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,5120,0.013150222599506378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,10240,0.021124444074100916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,512,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,7168,0.015896000795894198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,12288,0.23662400245666504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,1024,0.2481013403998481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,2048,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,4096,0.052608887354532875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,64,0.0181040002240075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,2048,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,128,0.012552888856993781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,128,0.00612088872326745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,1536,0.023552000522613525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,1024,0.006649777707126405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,1536,0.02906222144762675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,16384,0.07520177629258898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,7168,0.013603555659453074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,7168,0.21034222178988984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,512,0.012701333396964602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,64,0.01161688897344801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,256,0.02587466604179806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,3072,0.010063110954231685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,2048,0.0064702220261096954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,3584,0.017100445098347135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,5120,0.01583377851380242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,2560,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,2560,0.03315466642379761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,3584,0.3977217674255371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,256,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,3072,0.013789333403110504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,8192,0.03151199883884854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,3584,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,6144,0.23765243424309623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,3584,0.021866665946112737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,2048,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,256,0.005464000006516774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,256,0.005780444376998478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,1024,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,256,0.013912889692518445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,3072,0.01315822203954061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,3072,0.01884888940387302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,5120,0.027442667219373915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,256,0.015254222684436373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,128,0.0030382221771611106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,3072,0.05760711431503296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,2560,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,32,0.01865511139233907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,32,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,2048,0.020242666204770405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,51200,0.010454222559928894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,3072,0.007517333659860823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,3072,0.10402222474416097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,1024,0.007686222592989604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,51200,0.06199200285805596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,7168,0.011171555353535546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,51200,0.4114853276146783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,8192,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,5120,0.010469333165221743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,7168,0.010613333019945355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,128,0.03017510970433553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,4096,0.20670399400922987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,128,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,8192,0.010852444503042432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,3584,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,128,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,8192,0.0570639967918396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,7168,0.03951822386847602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,128,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,12288,0.035784890254338585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,51200,0.4661004278394911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,768,0.005970666805903117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,1024,0.014409777190950183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,512,0.0041653331783082755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,3072,0.009132444030708736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,128,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,51200,0.9716995027330187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,7168,0.07791022459665935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,128,0.015494222442309061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,512,0.003970666478077571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,3584,0.009768889182143742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,32,0.0019502221710152095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,65536,0.06528266933229235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,768,0.016495111915800307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,64,0.0019066666977273093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,768,0.02367288867632548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,7168,5120,0.09633688794242011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,10240,0.032596445745891996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,16384,0.9379546907213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,8192,0.040844443771574236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,768,0.02462755474779341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,12288,0.0192284451590644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,3072,0.02195377813445197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,7168,2048,0.0068897778789202375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,3072,0.013421333498424955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,768,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,32,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,16384,0.3607475492689345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,5120,0.02162666618824005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,512,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,128,0.10255555311838786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,512,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,7168,0.5018177562289768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,12288,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,65536,0.28495287895202637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,128,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,12288,0.45397334628634983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,1536,0.017511111166742113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,128,0.0021102221475707158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,32,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,128,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,2048,0.008287111090289222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,5120,0.45633602142333984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,10240,0.09414488739437527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,3072,0.012465777496496836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,1536,0.021274665991465252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,10240,0.03503555390569899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,7168,0.03551377852757772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,512,0.01595288846227858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,6144,0.0127368892232577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,4096,0.009840000006887648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,4096,0.01998044384850396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,51200,0.20028089152442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,1024,0.016516443755891588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,51200,0.05138755506939358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,6144,0.030440890126758154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,1536,0.007719111111429002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,8192,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,1024,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,51200,0.052896890375349254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,51200,0.0400622222158644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,6144,0.01423911088042789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,5120,0.021922666165563796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,512,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,128,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,12288,0.575110223558214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,10240,0.07588355408774482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,768,0.022611555125978258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,6144,0.009641778137948778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,2048,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,1536,0.018529777725537617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,256,0.025123554799291823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,32,0.004568888909286923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,7168,0.01928711103068458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,128,0.006073777874310811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,7168,0.017137777474191453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,2560,0.03352888756328159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,1024,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,256,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,12288,3.7907617357042103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,1024,0.004980444494220945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,16384,0.036320000886917114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,256,0.022962666220135156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,8192,65536,0.08381778001785278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,6144,0.04519733455446032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,6144,0.00964355551534229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,10240,0.047713776429494224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,3072,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,1536,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,512,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,1024,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,4096,0.10715999868181016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,3584,0.07574666870964898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,51200,0.022980444961124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,128,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,7168,0.17085422409905326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,7168,0.3001555601755778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,2048,0.007378666765160031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,1536,0.007330666813585494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,3072,0.03573777940538194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,768,0.013255111045307584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,1024,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,768,0.012081777883900536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,768,0.009602666729026372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3072,16384,0.023056889573733013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,256,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,1024,0.031680001152886286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,16384,0.1371955606672499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,10240,0.02902311086654663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,6144,0.004855111241340637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,1024,0.016712889075279236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,65536,0.12418133682674831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,512,12288,0.04051911168628269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,4096,0.01886311173439026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,65536,0.3052604463365343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,7168,0.05111733410093519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,256,0.011626666618718041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,1024,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,1536,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,512,0.018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,2560,0.05832622448603312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,10240,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,3584,0.007184888753626082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,8192,0.14269599649641249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,64,0.005712000032265981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,128,0.015428443749745687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,6144,0.028205331828859117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,51200,3.342287911309136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,4096,0.3866364426083035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,1024,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,3072,0.05764444669087728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,768,0.003597333199448056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,51200,0.05608177847332425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,512,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,64,0.006667555620272954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,1536,0.013286222186353473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,5120,0.014888889259762235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,8192,0.03531555665863885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,10240,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,512,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,1024,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,12288,0.08369866344663833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,16384,0.023071999351183575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,1024,0.017514665921529133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,32,0.019488889310095046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,512,0.01443733274936676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,12288,0.022879999544885423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,3072,0.053640888796912305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,65536,0.14471288522084555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,12288,0.013473778135246701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,5120,1.1090141932169597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,16384,0.056052446365356445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,12288,0.013554666605260638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,3072,0.05720622009701199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,8192,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,1024,0.005720888988839255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,128,0.02164622147878011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,512,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,64,0.006334222025341458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,8192,0.02403644389576382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,1024,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,5120,0.008994667066468133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,4096,0.3556186623043484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,512,0.025677333275477093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,6144,0.013992889059914483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,4096,0.03840444485346476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,5120,0.009520889156394536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,16384,6144,0.5688835779825846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,768,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,7168,0.01475288967291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,16384,0.14146133263905844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,64,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,16384,0.025183111429214478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,1536,0.018227555685573153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,1536,0.02530044482813941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,1024,0.0038319999972979226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,51200,0.06212177541520861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,2560,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2560,64,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,16384,0.15874932871924505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,10240,0.010040889183680216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1024,256,0.008908444808589088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,256,0.07919289006127252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,2048,0.027790221903059218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,8192,0.03241599930657281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,32,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,768,0.018941332896550495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,768,0.011735111474990845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,2560,0.0196426659822464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,512,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,7168,0.07395822472042508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,12288,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,768,0.01238222254647149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,16384,0.03667644328541226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,128,0.003519999898142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,512,0.003589333345492681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,1024,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,128,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,7168,0.037900444534089826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,5120,0.5543155670166016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,10240,0.04462577899297079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,51200,0.0672266681989034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,3584,0.0042035554846127825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,10240,0.015336889359686108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,10240,0.021198223034540813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,51200,0.15822933779822454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,65536,0.0710666643248664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,51200,0.4282302326626248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,7168,0.029566221767001685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,8192,0.5086737738715278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,2560,0.015121777852376303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,10240,0.08427377541859944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,12288,256,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,128,0.006305777778228124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,8192,0.0566435522503323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,8192,0.1858044465382894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,128,0.0032631111227803757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,6144,0.15185866090986463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,1536,0.01789422167672051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,5120,0.01387555566098955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,7168,0.15375555886162653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,65536,0.21358933713701037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,5120,0.012867555850081973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,3072,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,1024,0.009005332986513773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,1024,0.03607022100024753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,5120,0.01703288820054796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,16384,0.024953777591387432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,8192,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,3072,0.01516355574131012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,16384,0.24038932058546278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,1536,0.41071467929416233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,4096,0.07415111197365655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,5120,0.024703111913469102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,2560,0.010028444230556488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,51200,0.5252657996283637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,3072,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,3072,0.007737778127193451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,5120,0.03837688763936361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,2048,0.023056000471115112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,6144,0.0162453336848153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,12288,0.21990133656395805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,2560,0.18413244353400338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,2048,0.00927111092540953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,4096,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,16384,0.01739644507567088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,4096,6144,0.03091111116939121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,64,0.011086222198274402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,6144,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,51200,0.03784977727466159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,2048,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,32,0.013889777991506787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,5120,0.014631999863518609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,16384,0.041103111373053655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,1024,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,256,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,65536,0.16654577520158556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,1024,0.006598222172922558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,2560,0.03125244379043579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,6144,0.06312444474962023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,10240,1.1966213650173612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,65536,0.6930124494764539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,2560,0.020399999287393358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,3584,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,128,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,4096,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,5120,0.02390755547417535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,12288,0.027440889014138117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,512,0.10772266652848984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,2560,0.007409777906205919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,51200,0.5147848659091526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,16384,3584,0.1675377819273207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,10240,0.015517334143320719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,512,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,3584,0.11996622880299886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,5120,0.018597332967652213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,7168,0.04850044515397814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,256,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,8192,4096,0.032953778902689614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,4096,0.07248089048597547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,32,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,32,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,768,0.016547555724779766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,768,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,6144,0.005660444498062134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,65536,0.027348443865776062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,768,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,51200,0.045271111859215625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,1024,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,512,0.010437332921557956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,3584,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,2048,0.015189333094490899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,1024,0.009389333426952362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,512,1.1507013108995225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,65536,0.8583191235860189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,16384,0.24577511681450737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,7168,0.031210667557186548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,4096,0.008033778104517195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,32,0.007662222617202335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,1536,0.007382222347789341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,64,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,128,0.01684799955950843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,51200,0.33799200587802464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,51200,2.0440292358398438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,3584,0.04224977890650431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,51200,0.04674222071965536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,1536,0.019872888922691345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,12288,0.7163048850165473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,3072,0.04103022151523166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,2560,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,5120,0.012479999827014076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,3072,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,512,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,5120,0.03141244583659702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,6144,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,768,1024,0.010769777827792697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,6144,0.10513510968950059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,256,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,65536,3.4916551378038196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,5120,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,5120,0.14948799875047472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,3072,0.01791733337773217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,768,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,768,0.010457778142558204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,1536,0.08892889155281915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,12288,0.055507553948296436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,3072,0.09839200311236912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,4096,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,256,0.0036675557494163513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,5120,0.033152888218561806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,8192,0.02747111188040839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,65536,0.14577778180440268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,10240,0.18760177824232313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,3072,0.205533332294888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,2560,0.028295109669367473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,2560,0.020960888928837244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,5120,0.03950044512748718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,1024,0.00305600009030766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,512,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,256,0.01818933255142636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,128,0.03193333413865831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,12288,0.07784266604317559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,4096,0.02466222147146861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,51200,0.15109954939948186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,1536,0.0038577777643998465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,3072,0.010672888822025724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,6144,0.14204622639550105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,12288,0.04085955686039395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,3584,0.08946044577492608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,8192,0.03331111205948724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,16384,0.1909653345743815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,256,0.03196533189879523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,51200,0.12979288895924887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,32,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,64,0.011190222369299995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,10240,0.05275999837451511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,65536,0.06282844146092732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,16384,0.04570666617817349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,5120,0.21388533380296496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,65536,0.1003191139962938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,16384,0.016537777251667447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,1024,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,5120,0.04367288947105408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,1024,0.010418666733635796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,2048,0.058021333482530385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,512,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,4096,0.008620444271299574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,10240,0.06408088737063937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,512,0.007166222565703922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,12288,0.03296711047490438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,1536,0.021134222547213238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,51200,0.47238842646280926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,4096,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,1024,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,768,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,768,0.01854933301607768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,7168,0.24622578091091582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,64,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,512,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,3584,0.008087111016114553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,51200,0.07113689184188843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,8192,0.09296622541215684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,3072,0.011172444456153445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,768,0.025070221887694463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,16384,0.020544888244734872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,6144,0.14082488748762342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,1024,0.005711110929648082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,256,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,51200,0.9807110892401801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,2048,0.012063110868136087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,3584,0.008910222185982598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,3072,0.007744000189834171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,10240,0.043819556633631386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,65536,0.06126311090257433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,2560,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,7168,0.030496888690524634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,7168,0.027460444304678176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,2560,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,7168,0.02533600065443251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,16384,0.33394577768113876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,8192,0.05840622054206001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,3072,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,12288,0.28000354766845703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,768,0.020376000139448378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,6144,0.05626844697528415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,64,3072,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,51200,0.1357920037375556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,10240,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,1024,0.24026399188571504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,512,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,3584,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,128,0.005223999834722943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,512,0.06263822317123413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,10240,0.14317688677046034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,10240,0.046804444657431714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,2048,0.0069884442620807225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,128,0.013295110729005603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,10240,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,4096,0.04010844561788771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,1536,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,12288,0.08948533402548896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,1024,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,12288,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,7168,0.018813333577579923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,1024,0.003938666648334927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,32,0.0018533333722088074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1536,2048,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,7168,0.011354666617181567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,512,0.02477600011560652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,51200,0.2466355429755317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,1536,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,4096,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,768,0.01218311074707243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,16384,0.04599377844068739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,32,0.005427555491526921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,128,0.012102222277058495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,1536,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,7168,2048,0.008722667064931657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,3072,0.5931608941819932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,2560,0.022712000542216833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,3072,0.1634933286243015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,4096,0.003875555677546395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,10240,0.08323111136754353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,1536,0.005705777969625261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,2048,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,1024,0.4363244374593099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,2048,0.04334400097529093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,3072,0.012031111452314587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,128,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,12288,0.037188443872663714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,5120,0.011095110740926532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,5120,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,1024,0.0058328890138202244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,3072,0.016180444094869826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,65536,0.8719671037462023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,12288,0.06785600052939521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,768,0.009753777749008602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,4096,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,51200,0.197014225853814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,3072,0.01586666703224182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,4096,0.030048890246285334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,51200,0.031982223192850746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,5120,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3072,1024,0.021148444877730474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,128,0.007749333149856991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,65536,0.16362755828433567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,3072,0.09731111261579727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,10240,0.015564444992277356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,64,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,16384,0.18067999680836996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,3072,0.06795466608471341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,51200,8.331908331976996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,512,0.01460355520248413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,65536,0.09285422166188557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,2560,0.03973155551486545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,3584,0.14803289042578802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,10240,0.008615111311276754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,2560,0.0203493336836497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,2560,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,768,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,128,0.032604446013768516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,4096,0.016114667057991028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,65536,0.3966666592491998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,512,0.013040000365840064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,1024,0.027577777703603108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,5120,0.05732533004548815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,8192,0.012830221818553077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,8192,0.018198221921920776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,1536,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,3584,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,8192,10.43117438422309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,51200,0.057506667243109814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,2048,0.08010755644904242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,3584,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,256,0.020614221692085266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,10240,0.021586666504542034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,768,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,256,0.020595555504163105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,51200,0.0771848890516493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,5120,0.021581333544519212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,65536,0.5236835479736328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,3072,0.06181333462397257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,7168,0.10420266787211101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,3584,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,7168,0.02711822258101569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,2048,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,1536,0.027309333284695942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,7168,0.018792889184421964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,2560,0.12134933471679688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,51200,1.0994008382161458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,8192,0.013725333743625216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,768,0.0031431110368834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,768,0.026033777329656813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,2560,0.19041955471038818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,1024,0.029863999949561224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,1024,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,2048,0.034155554241604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,512,0.007552000383536021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,1024,0.018466666340827942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,512,0.005280888742870755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,1024,0.0038968887594011095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,1536,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,1024,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,768,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,4096,0.05571733580695259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,8192,0.025527111358112756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,1536,0.018221333622932434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,128,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1024,65536,0.0627982219060262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,3072,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,12288,0.03884355558289422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,3584,0.007792000141408708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,1536,0.011986666255527072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,3072,0.010667555862002902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,12288,0.03272799981964959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,6144,0.1220666699939304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,128,0.005377777748637729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,1024,0.00452622233165635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,8192,0.24834932221306694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,7168,0.023409777217441138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,51200,0.9613546795315213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,2560,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,3072,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,256,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,6144,0.03514400124549866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,12288,0.06620711088180542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,512,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,64,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,2560,0.025478222303920325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,3584,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,768,0.02533333334657881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,1536,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,256,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,8192,0.04880355464087593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,512,0.042246222496032715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,7168,0.06283999813927545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,51200,3.9809218512641062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,1024,0.006536888993448681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,8192,0.011428444749779172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,8192,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,3584,0.1416071123547024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,12288,0.013844444519943662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,768,0.12703911463419595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,4096,0.009207111265924241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,1024,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,10240,0.017475555340449016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,16384,0.3605786694420709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,3072,0.014363555444611443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,128,0.01016444464524587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,4096,0.012601777911186218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,3072,0.014840889308187695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,8192,0.022721777359644573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,3584,0.15341333548227945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,6144,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,1536,0.003911111089918348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,65536,0.37067288822597927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,512,0.01258044441541036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,2560,0.008295999632941352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,64,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,2048,0.004217777815130022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,3584,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,2560,0.028156446086035833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,7168,0.0351217786471049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,51200,0.060235553317599826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,51200,0.42842311329311794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,7168,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,7168,0.36094223128424746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,8192,0.39997243881225586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,51200,0.8488622241550021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,12288,0.042103999190860324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,10240,0.02769777841038174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,10240,0.011117333339320289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,768,0.009726222190592024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,3072,0.008403555386596257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,512,0.05543555484877693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,65536,0.2823733223809136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,2048,0.012061333490742577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,7168,0.01923288901646932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,128,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,4096,0.01904177831278907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,1536,0.0038951109680864546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,8192,0.2094551059934828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,768,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,6144,0.17821510632832846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,3072,0.010020444790522257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,1024,0.014447111222479077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,4096,1.2437715530395508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,3072,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,6144,0.025722665919197932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,6144,0.01457333399189843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,3584,0.14978133307562932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,65536,1.2244444953070748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,32,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,128,0.018943111101786297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,5120,0.13744978109995523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,3072,0.01993600030740102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,5120,0.02994933393266466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,51200,0.11213066842820908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,512,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,2560,0.009121778110663096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,12288,10240,0.31484532356262207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,51200,0.4296257760789659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,1536,0.08691022131178115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,3584,0.021891555852360193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,2560,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,7168,0.013291555146376291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,65536,0.09055288632710774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,768,0.027090667022599116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,51200,0.0513591104083591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,6144,0.05868888563579983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,256,0.005241777747869492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,5120,0.008709333009190029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,8192,0.07429333527882893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,16384,0.04872622092564901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,768,0.14800711472829184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,6144,0.4965324401855469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,8192,0.03445155421892802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,6144,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,4096,0.24737866719563803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,7168,0.033725334538353816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,10240,0.05565333366394043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,8192,0.17960800064934623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,6144,0.018586667047606576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,16384,0.02134399943881565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,64,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,16384,0.05532000131077237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,5120,0.012155555188655853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,512,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,64,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,10240,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,3584,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,6144,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,8192,0.02128177715672387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,512,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,32,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,4096,0.05929778019587199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,10240,0.5052888658311632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,6144,0.018664888209766813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,128,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,16384,0.02176977859603034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,32,0.0021964444054497615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,12288,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,7168,0.0595360000928243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,32,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,64,0.002226666650838322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,512,0.0027946666296985415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,8192,3.287507587009006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,512,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,768,0.00452977791428566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,1536,0.01683999929163191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,5120,0.2575742138756646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,768,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,1024,0.017463111215167575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,512,256,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,4096,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,256,0.009363555245929295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,12288,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,7168,0.0229768885506524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,10240,0.03604088889227973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,16384,0.03632888860172696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,512,0.03059555424584283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,7168,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,6144,0.1018035544289483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,256,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,6144,0.010071111222108206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,128,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,51200,0.05196355448828804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,8192,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,8192,0.010675555302037133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,7168,0.02342133389578925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,768,0.042025778028700084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,3072,0.03175466590457492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,2048,0.018414222531848483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,1536,0.06591111421585083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,256,0.006704000135262807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,1536,0.0618399977684021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,3072,0.013095999757448832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,16384,0.1489715576171875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,1024,0.053416000472174756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,5120,0.03833333320087857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,3072,0.09648444255193074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,3072,0.013924444715181986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,256,0.005384888913896348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,2048,0.004795555439260271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,256,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,512,0.01459111107720269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,3072,0.018079110317760043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,7168,0.021928000781271193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,51200,0.10123555527793036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,3584,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,6144,0.10982666412989299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,2560,0.10553689135445489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,2048,0.00675022229552269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,10240,0.026384888423813715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,3584,0.024333332975705464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,5120,0.2912702295515272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,10240,0.1565262211693658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,768,0.021287999219364587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,51200,0.06103822257783678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,4096,0.03970844546953837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,7168,0.07688266701168485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,1536,0.041687998506757945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,3072,0.031855109665128924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,65536,3.359640969170464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,3072,0.015997333659066092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,1536,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,5120,0.02522755497031742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,4096,0.030552887254291113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,16384,0.023002665903833177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,5120,0.017139555679427255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,16384,0.09080000056160821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,128,0.02286400066481696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,10240,0.013912000589900546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,51200,0.23784266577826607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,128,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,512,0.007546666595670912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,128,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,1536,0.010967999696731567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,12288,0.4326959980858697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,32,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,5120,0.27325688468085396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,768,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,32,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,6144,0.01036622209681405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,12288,0.2786533302730984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,3584,0.008645333349704742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,1536,0.04348977737956577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,768,0.015104000767072042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,1024,0.02602044410175747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,2048,0.04842933350139194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,1024,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,256,0.01716977854569753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,256,0.016913778252071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,7168,0.036751998795403376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,4096,0.1713742282655504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,10240,0.027004443936877783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,7168,0.14016621642642552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,256,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,3072,0.009082666701740688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,12288,0.057725330193837486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,10240,0.022827555735905964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,2048,0.047622223695119224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,12288,0.0418906675444709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,128,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,16384,0.05794577466117012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,65536,0.04671377936999003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,3072,0.07117688655853271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,2560,0.024158222807778254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,16384,0.07217599948247273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,512,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,4096,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,512,0.016908443636364408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,65536,0.48353158103095156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,51200,0.04086044430732727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,5120,0.009330666727489894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,12288,0.12667200300428602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,3584,0.04945422212282816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,32,0.006798222247097228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,512,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,6144,0.009297777381208207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,12288,0.05077955457899305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,65536,0.06990933418273926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,256,0.017879999346203275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,5120,0.08600799904929267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,2048,0.11509599950578477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,3584,0.019487111104859244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,12288,0.024736889534526404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,512,0.006211555666393704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,2048,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,8192,0.7199849022759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,768,0.023355555203225877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,3072,0.05900622076458401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,3072,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,3072,0.014411555396185981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,6144,0.05927466683917575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,2048,0.02162577708562215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,5120,8192,0.1607955561743842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,2048,0.010919999745157031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,5120,0.04135555691189236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,128,0.13535466459062365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,3072,0.0681555536058214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,7168,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,2048,0.0709493358929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,8192,0.030689779255125258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,3584,0.005220444252093633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,16384,0.05726044707828098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,1024,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,1536,0.018904000520706177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,1536,0.022065778573354084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,5120,0.05878133243984646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,3072,0.045695109499825366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,5120,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,2048,0.010447999669445885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,128,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,4096,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,65536,0.26319999165005153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,512,0.02439733346303304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,2048,0.013312000367376538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,32,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,51200,0.07283200158013238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,4096,0.02815999918513828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,1024,0.018210666047202218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,8192,0.12943466504414877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,6144,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,51200,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,65536,0.14829599857330322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,5120,0.2815111213260227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,2560,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,12288,0.05497866537835863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,6144,0.12451467249128555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,256,0.015462223026487561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,12288,0.10665777656767104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,8192,0.09765511088901097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,1024,0.027753776974148218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,512,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,512,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,256,0.016897777716318767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,128,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,512,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,1024,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,1536,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,512,0.107515553633372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,65536,0.0830871131685045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,256,0.013193777865833707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,64,0.02854666776127285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,6144,0.015094222293959724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,6144,0.03734133309788174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,8192,0.09806578026877509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,2560,0.010119111173682744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,51200,0.45613601472642684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,12288,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,64,256,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,128,0.01035733355416192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,32,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,128,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,16384,0.3943919870588515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,512,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,768,4096,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,256,0.003080000066094928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,65536,0.2717848883734809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,768,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,512,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,2048,0.03411111235618591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,3584,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,1024,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,51200,0.6021795802646214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,3584,0.0042444442709287005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,32,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,2048,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,6144,0.02916533417171902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,7168,0.016601777738995023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,6144,0.24220887819925943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,512,0.015054222610261707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,8192,0.021983999345037673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,128,0.03573066658443875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,12288,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,32,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,1536,0.03451022174623277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,1024,0.07925688558154635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,768,0.023768888579474554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,65536,0.052391111850738525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,4096,0.01818933255142636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,2048,0.028985778490702312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,1024,0.031155556440353394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,1024,0.007705777883529663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,512,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,8192,0.014592889282438489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,16384,0.3682417869567871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,10240,0.27350576718648273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,3584,0.5941155751546224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,1536,0.01236533373594284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,65536,0.6707866456773547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3584,65536,0.08053421974182129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,256,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,3584,0.03917155663172404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,7168,0.16679022047254774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,1536,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,2048,0.06363288561503093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,16384,0.03753066725201077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,3072,0.13531377580430773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1536,2560,0.007239110767841339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,12288,0.15827288892534044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,5120,0.49179379145304364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,2560,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,128,0.0046942221621672315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,2560,0.012511110968059964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,128,0.02177422245343526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,5120,0.012459555433856117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,6144,0.02442488902144962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,65536,0.14855199389987522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1024,768,0.04733955529001024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,8192,0.010758221977286868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,12288,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,12288,0.04799911048677233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,7168,0.03315911028120253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,16384,0.026346666945351496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,12288,1.1467644373575847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,128,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,16384,0.04365333252482944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,10240,0.015816888875431485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,12288,0.04741600155830383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,256,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,10240,0.08460177977879842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,51200,0.5272275606791178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,128,0.011878222227096558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,4096,0.013456000222100152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,2048,0.02126666737927331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,8192,0.023000889354281958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,768,0.01977599991692437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,2048,0.021135111649831135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,6144,0.009744000103738572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,12288,0.03360889024204678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,128,0.005884444548024072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,6144,0.01109155515829722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,1536,0.00810222244924969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,5120,0.03328355484538608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,1024,0.027296889159414504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,128,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,16384,0.021368000242445204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,16384,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,256,0.06167288621266683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,2560,0.298239098654853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,256,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,6144,0.0820373362965054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,256,0.01751377847459581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,5120,0.033287998702791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,1024,0.015485333071814643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,7168,0.011687111523416309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,2560,0.10553066598044501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,3584,0.030309332741631403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,16384,0.3807520071665446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,64,4096,0.003544888976547453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,256,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,256,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,7168,0.26830045382181805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,51200,0.042993777328067355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,64,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,768,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,4096,0.015258666541841296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,2048,0.009964444571071202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,16384,0.023753777146339417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,2048,0.008257778154479133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,3584,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,32,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,1024,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,768,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,4096,0.015951111912727356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,1536,0.038506666819254555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,768,0.025409777959187824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,512,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,12288,0.028954668177498713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,3584,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,256,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,7168,0.0551182230313619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,8192,0.012834666503800286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,768,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,16384,0.5055911276075575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,2048,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,512,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,10240,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,16384,0.017364444004164804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,1536,0.0053031109273433685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,128,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,512,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,12288,0.0527724458111657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,16384,0.5013528929816352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,32,0.00591466658645206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,256,0.00601244428091579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,128,0.006320000108745363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,4096,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,2048,0.01036444471942054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,7168,0.02409066590997908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,51200,0.042265776130888194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3072,8192,0.057554668850368924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,2560,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,51200,0.0379342238108317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,3584,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,16384,0.07690222395790948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,3584,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,10240,0.012141333685980903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,1536,0.19023377365536162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,5120,0.02464177707831065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,1024,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,5120,0.03792444533771939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,768,0.020946666598320007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,12288,0.31414932674831814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,16384,0.11857955985599095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,6144,0.09084266424179077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,5120,0.14733421802520752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,3072,0.03906044363975525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,1024,0.007544000115659501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,2048,0.020266667008399963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,768,0.006858666737874349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,1536,0.04612888892491659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,7168,0.021014221840434607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,1024,0.040031999349594116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,2560,0.007689778175618913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,12288,768,0.018608889646000333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,128,0.018054222067197163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,2560,0.006960888703664144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,10240,0.17109333144293892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,128,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,3072,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,12288,0.059839109579722084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,1536,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,12288,0.068996442688836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,12288,0.10596711105770534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,4096,0.06122666597366333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,2560,0.016327111257447135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,6144,0.027791112661361694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,1024,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,32,0.0025031111306614345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,10240,0.01034755590889189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,3072,1.2457670635647244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,51200,2.0765698750813804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,51200,0.8760071330600314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,6144,0.1402746703889635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,7168,0.0107342220014996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,65536,0.3780755466885037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,32,0.0057546666098965546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,12288,0.10210222005844116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,6144,0.056596444712744824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,2560,0.024784889486100938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,64,0.012533333566453723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,1536,0.019322666856977675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,16384,0.04093866546948751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,5120,0.11796532736884223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,7168,0.010461333725187512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,2048,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,7168,0.04203111264440748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,256,0.03269688950644599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,12288,0.2616533438364665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,2560,0.028455999162462022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,3584,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,4096,0.00868000007337994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,128,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,10240,0.011768889096048145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,2560,0.011173333558771344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,1536,0.010943111446168689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,65536,0.800904909769694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,2048,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,5120,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,1536,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,65536,0.1478631099065145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2048,768,0.009767110976907942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,512,0.014763555592960782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,3072,0.039705776506000094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,128,0.10639644331402248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,1536,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,8192,0.034210665358437434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,1536,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,8192,0.02858577834235297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,5120,0.018580444984965853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,51200,0.032448000378078885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,16384,0.09912800126605564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,5120,0.00907555553648207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,2560,0.01443555619981554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,128,0.019864888654814828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,32,0.005969777703285217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,7168,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,1536,0.027620444695154827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,128,0.010327999790509542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,2048,0.01204355557759603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,256,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,65536,0.4520791371663411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,6144,768,0.16857155164082846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,7168,0.031952887773513794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,2560,0.01239733315176434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,4096,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,4096,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,64,0.0058142224119769196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,128,0.017444445027245414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,3584,0.007894222107198503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,2560,0.028792887926101685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,10240,0.02696711156103346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,7168,0.45404709710015195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,128,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,512,128,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,3072,0.013605333036846586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,12288,0.19467022683885363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,1024,0.00944622192117903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,3072,0.024306666519906785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,1024,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,51200,0.15831022792392307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,1024,0.012496000362767113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,3072,0.09054044220182632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,512,0.0566613343026903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,128,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,1536,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,16384,0.07543377743826972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,16384,0.02311822275320689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,768,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,65536,0.012114666402339935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,10240,0.012307555311255984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,10240,0.0325857765144772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,2048,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,256,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,7168,0.01789422167672051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,16384,0.16304800245496961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,1024,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,10240,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,65536,0.7226559850904676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1024,1024,0.027106665902667578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2560,3584,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,512,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,6144,0.01796533332930671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,1536,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,1536,0.04430933462248909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1024,5120,0.034613334470325045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,768,0.02940888868437873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,12288,32,0.03383822242418925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,1024,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,2048,0.0038542221817705366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,512,65536,0.05686044692993164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,12288,0.024958221448792353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,2560,0.302872896194458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,51200,0.8630275726318359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,2560,0.011915555430783166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,51200,0.08548088868459065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,16384,0.03714488943417867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,256,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,2048,0.032856888241238065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,8192,0.07086133294635348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,4096,0.06691822078492907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,32,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,256,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,65536,0.021036444438828364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,6144,1.4868479834662542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,8192,0.19122933016883004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,128,0.022271111607551575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,5120,0.019652444455358718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,768,0.004159111115667555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,12288,0.038196444511413574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,51200,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,7168,0.01998044384850396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1536,12288,0.03716622127426995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,128,0.004933333231343163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,512,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,1024,0.0032888888898822996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,768,0.01608800060219235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,128,0.009481777747472128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,8192,0.04660177893108792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,10240,0.043507556120554604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,10240,0.022629333866967097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,6144,1.0490763982137044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,1024,0.008904000123341879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,3072,0.013811555173661975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,1024,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,51200,0.13446133666568333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,3584,0.008098666866620382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,32,0.1210462252298991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,64,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,1536,0.02437777817249298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,4096,0.036390221781200834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,65536,0.07311377922693889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,5120,0.08128888739479913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,32,0.007324444750944774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,10240,0.014944000376595391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,1536,0.005869333528810077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,5120,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,7168,0.02093066606256697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,12288,0.014349333114094205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,768,0.017855111095640395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,8192,0.011735111474990845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,2560,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,5120,0.4397146701812744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,32,0.018365333477656048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,8192,0.03199022346072727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,12288,0.011359111302428775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,16384,0.043268445465299815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,8192,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,256,0.03084711233774821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,1024,0.006723555425802867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,65536,0.140775998433431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,1024,0.09786844253540039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,16384,0.032448000378078885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,10240,0.2524782286749946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,3072,0.01962844365172916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2048,128,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,512,0.0024906665914588505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,128,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,7168,0.018560000591807894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,512,0.016567111015319824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,64,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,32,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,51200,0.957713762919108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,16384,1.2248569064670138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,64,0.01720977822939555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,16384,0.2667084534962972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,256,0.01516266663869222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,2048,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,6144,0.023563555545277063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,512,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,4096,0.07916000154283312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,128,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,3584,0.004997333304749595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,16384,0.03999200132158067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,3584,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,32,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,1024,0.023911999331580266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,7168,0.021640888518757288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,3072,0.03160444564289517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,8192,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,2560,0.040630221366882324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,3584,0.024361777636739943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,6144,0.019349333312776353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,2560,0.010227555202113258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,1024,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,65536,0.367794672648112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,4096,0.015990222493807476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,128,0.0021422221842739317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,1024,0.011549333731333414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,51200,1.157131512959798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,1536,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,128,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,3072,0.02088977727625105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,6144,0.021328000558747187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,512,0.007134222322040134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,65536,0.7987910906473795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,6144,0.1139591137568156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,6144,0.013230221966902414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,5120,0.025051555699772302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,10240,0.023690667417314317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,3072,0.0058151111006736755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,2560,0.024838222397698298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,5120,0.00904977818330129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,16384,0.02032444377740224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,32,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,10240,0.07553066809972127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,4096,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,8192,0.07281955745485094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,512,0.011972444752852121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,64,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,3584,0.007882666256692674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,1024,0.029823111163245306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,8192,0.011955555942323474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,1536,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,128,0.015861334072219003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,1024,0.03665688965055678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,3584,0.020607999629444547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,5120,0.08988000286950006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,2048,0.008906666603353288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,65536,0.09392622444364761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,4096,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,5120,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,768,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,6144,0.04551466637187534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,5120,0.17164711157480875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2560,1024,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,3072,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,12288,0.2625039948357476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,128,0.022557333111763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,3072,0.00755555596616533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,12288,0.028280887338850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,768,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,4096,0.02975644336806403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,7168,0.023131555981106226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,3072,0.2644497818417019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,7168,0.009518221848540837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,3584,0.0074177781740824384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,10240,0.26620178752475315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,7168,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,768,0.02608888844648997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,10240,0.02127200033929613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,256,0.1211671166949802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,6144,0.03557422094874912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,4096,0.01606666710641649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,256,0.0019786666250891155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,65536,0.5256480111016167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,3072,0.011864888999197217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,768,0.002809777855873108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,6144,0.4464400079515245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,512,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,8192,0.04474488894144694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,768,16384,0.03330488999684652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,4096,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,5120,0.15545245011647543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,1024,0.011181332998805575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,256,0.020671111014154222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,4096,0.02051199972629547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,512,0.023020444644822016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,4096,0.0655297769440545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,6144,0.04103999998834398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,4096,0.041321777635150485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,2560,0.025630222426520452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,3584,0.013784888717863293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,5120,0.015830222103330824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,3584,0.0076248885856734375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,768,0.022522666388087805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,768,0.002573333266708586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,1536,0.011101333631409539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,512,0.015882665912310284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,1024,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,128,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,1024,0.003543111185232798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,10240,0.1286097764968872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,51200,8192,0.44390132692125106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,2560,0.09159288803736369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,512,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,51200,0.050796445873048574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,5120,0.021964444054497614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,6144,0.021270222134060327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,128,0.009320889082219865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,2048,0.016887111796273124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,51200,0.3281493451860216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,768,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,51200,0.1828835540347629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,2560,0.01146044416560067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,65536,0.05970933039983114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,64,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,3584,0.15547200043996176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,6144,0.018199111024538677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1024,512,0.01388711151149538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,5120,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,3584,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,4096,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,16384,0.09236266877916123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,4096,0.031537777847713895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,16384,0.39394132296244305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,5120,0.018556444181336295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,768,0.0029173334025674393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,65536,0.5571902063157823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,256,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,10240,0.027035555905765955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,51200,2.5700719621446395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,1024,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,256,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,7168,0.010852444503042432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,1024,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,3072,0.06697955396440294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,1536,0.008389333056079017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,128,0.005528888768619961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,5120,0.2053377760781182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,512,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,2560,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,2560,0.009431111315886179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,8192,0.01368533323208491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,51200,0.25378044446309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,16384,0.04109155469470554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,1024,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,3584,0.0074311114019817775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,8192,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,65536,0.40060000949435765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,65536,0.09090666638480292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,1536,0.00869955536392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,64,0.0018968888454967076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,65536,0.062055998378329806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,512,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,12288,0.41775022612677676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,256,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,1536,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,256,5120,0.013902222116788229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,32,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,1024,0.02049688829316033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,65536,0.4477235476175944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,3584,0.013239999612172445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,4096,0.021692444880803425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,768,0.0063884444534778595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,3072,0.015178667174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,3584,0.014144000079896716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,4096,0.019104889697498746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,8192,0.16438133186764187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,3584,0.02904799911710951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,2048,0.02335111134582096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,12288,0.12751911746131048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,256,0.020971554848882887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,256,0.016466667254765827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,51200,0.9228550593058268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,4096,2048,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,1024,0.02126666737927331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,10240,0.09575377570258246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,16384,0.06200710932413737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,10240,0.016032889485359192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,16384,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,51200,0.14140710565778944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,7168,0.016304888659053378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,4096,0.01236977759334776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,8192,51200,16.388358222113716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,51200,10.222774929470486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,3072,0.3253600067562527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,1024,0.011415999796655444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,256,0.02813777658674452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,6144,0.28337155448065865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,256,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,7168,64,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,7168,0.019903111788961623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,2560,0.1009591089354621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,8192,0.013582222163677216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,768,0.008905777500735389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,7168,0.00998399986161126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,51200,0.10736355516645645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,65536,3.82185787624783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,64,0.0041991112132867174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,1536,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,4096,0.01595288846227858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,6144,0.2036222219467163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,51200,0.47704622480604386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,2048,0.05553422371546427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,2560,0.015263110399246216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,1024,0.00424177779091729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,2048,0.03386222322781881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,3072,0.04303555687268575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,12288,0.01980444457795885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,1024,0.019241778386963736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,256,0.03501244386037191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,512,0.050211555427975126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,65536,0.37448443306816953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,12288,0.1614417764875624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,51200,0.7722711033291287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,2048,0.04093599981731839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,7168,0.018632888793945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,256,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,2048,0.00543200017677413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,512,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,1536,0.09626489215426975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,3072,0.010564444793595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,6144,0.1294977797402276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,2560,0.013100444442696042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,1024,0.0031902222997612427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,256,5120,0.004880889008442561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,128,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,128,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,768,0.02784088916248745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,7168,0.02313066687848833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,3072,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,12288,15.49656507703993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,256,3584,0.04343822267320421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,512,64,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,6144,0.03388266762097677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,4096,0.026368000441127356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,768,0.10498133632871841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,2560,0.00720000018676122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,3584,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,32,0.004515555583768421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,51200,1.3922008938259547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,16384,0.019883554842736986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,3072,0.04231377773814731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,3072,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,12288,0.023256889647907678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,7168,0.03482666611671448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,2560,0.02695111102528042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,1536,0.023744000328911677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,65536,0.013222222526868185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,1536,0.006703111032644908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,768,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,64,0.010075555907355415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,8192,0.25270401106940377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,65536,0.6819804509480795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,512,0.015429332852363586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,7168,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,51200,0.17028444343143037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,16384,0.028199109766218398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,2048,0.023315555519527856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,2560,0.0325582226117452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,6144,0.02639111048645443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,2048,0.025306666890780132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,64,0.012725333372751871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,5120,0.023476445012622412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,12288,0.016372443901167977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,2048,0.023312888211674158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,51200,0.07664533456166585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,5120,32,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,2048,0.006803555621041193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,2048,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,1024,0.008626666333940294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,51200,0.05293600095642937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,51200,0.30002578099568683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,65536,0.12469955285390218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,12288,0.11486755477057563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,4096,0.06255733304553561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,4096,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,512,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,16384,0.03549155592918396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,10240,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,8192,0.018897778458065458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,768,0.06178222099939982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,10240,0.028785778416527644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,3584,0.041217777464124895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,5120,0.023669333921538457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,10240,0.03087377879354689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,10240,0.028600888119803533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,64,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,2560,0.23147734006245932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,7168,0.012072000238630505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,1536,0.020995555652512446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,51200,1.1957715352376301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,3072,0.017700443665186565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,2560,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,3584,0.04421066575580173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,768,0.05196711089875963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,5120,0.02123822271823883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,2048,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,2048,0.026924444569481745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,12288,0.05630222294065687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,12288,0.013431111143694984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,51200,0.02091911103990343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,4096,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,4096,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,2048,0.02074044446150462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,64,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,6144,2048,0.0229795558585061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,51200,0.09478044509887695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,7168,0.020581333173645865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,2560,0.021214221914609272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,10240,0.06012533108393351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,2048,0.027373333772023518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,10240,0.012123555772834353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,3584,0.01717511150572035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,51200,0.019156444403860304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,3584,0.015100444356600443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,6144,0.11085244682100083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,2560,0.014757333530320061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,1536,0.01661422186427646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,7168,0.24101599057515463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,128,0.008574222524960836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,7168,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,128,0.004926222066084544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,10240,0.050203555160098605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,1024,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,5120,0.02260977857642704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,128,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,1024,0.018551111221313477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,7168,0.038069334295060896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,51200,0.4292951160007053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,256,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,2560,0.007704888780911763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,3072,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,12288,0.015535111228624979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,2560,0.010679999987284342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,2048,0.004565333326657613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,3584,0.010642666783597736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,16384,0.04145155681504144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,768,0.01722577710946401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,51200,0.17053243849012586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,65536,0.12519733111063638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,16384,0.07356178098254733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,3584,0.17039199670155844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,3584,0.06292444467544556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,2048,0.062459554937150746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,12288,2.6206692589653864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,512,0.012266666524940066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,65536,0.07084444496366712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,8192,0.02997777859369914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,1024,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,3072,0.010834666589895884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,32,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,4096,0.023922666907310486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,16384,0.08425155613157485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,256,0.11530844370524089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,12288,0.10246044397354126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,2048,0.02591200007332696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,65536,0.7510106828477648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,4096,6.760068257649739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,5120,0.09530222415924072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,2048,0.08234844605127971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,10240,0.013435555828942193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,2560,0.02682755556371477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,64,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,32,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,256,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,65536,0.06974489159054227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,2560,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,2048,0.06980799966388278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,7168,0.014519110321998596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,5120,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,768,0.17114578353034127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,7168,0.03453777896033393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,2048,0.08402044243282741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,2048,0.0045422220395671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,768,0.02437955637772878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,6144,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,8192,0.040383110443751015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,8192,0.027024888330035742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,2560,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,6144,0.16092533535427517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,512,0.20791110727522108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,1536,0.025007999605602686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,1024,0.015885333220163982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,8192,0.06552000178231134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,12288,0.028379556205537584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,2560,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,8192,0.01273155543539259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,1024,0.06088888645172119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,7168,0.009870222045315636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,3584,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,7168,0.10560444328520034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,51200,0.26744090186225045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,6144,0.046296000480651855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,2560,0.012524444195959302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,12288,0.01331555595000585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,128,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,3584,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,10240,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,256,0.02259644369284312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,32,0.01035288886891471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,2048,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,32,0.001726222224533558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,128,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,512,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,512,0.056040002240075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,7168,0.01130844404300054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,16384,512,0.0069448889957533936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,128,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,4096,0.018986667195955913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,12288,0.06320266591178046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,51200,1.2937146292792427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,3584,0.02858755654758877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,3072,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,3584,0.25114133622911244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,65536,0.050344000260035195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,10240,51200,0.7270666758219401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,51200,0.05218133330345154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,768,0.01719111038578881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,7168,0.13355199495951334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,12288,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1536,8192,0.011718221836619906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,16384,0.035840001371171736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,4096,0.04964444372389051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,1024,0.013776000175211163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,128,0.008018666671382057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,32,0.008386666576067606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,5120,0.07965511083602905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,256,0.0035866668654812705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,65536,0.43516890207926434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,16384,0.13112444347805446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,2560,0.5342355304294163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,4096,0.0090506664580769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,6144,0.012949333422713809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,12288,0.01608088943693373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,6144,0.061814222070905894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,51200,0.07677688863542345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,128,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,8192,0.009617778162161509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,32,0.003527111063400904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,7168,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,128,0.005416888743638992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,64,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,10240,0.6929386456807455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,1024,0.029842668109469946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,4096,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,256,0.009650666680600908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,16384,4.89223395453559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,1536,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,12288,0.038951999611324735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,64,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,10240,0.02920266654756334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,256,0.013442666994200813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,4096,0.02199733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,1024,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,1024,0.00755555596616533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,64,0.00527555536892679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,5120,0.020868443780475195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,3584,0.08086844285329182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,4096,0.036644445525275335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,2048,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,10240,0.05308889018164741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,2048,0.011234666738245221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,4096,0.018530666828155518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,1024,0.1965493361155192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,3584,0.02464977734618717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,7168,0.02472355630662706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,5120,0.0862426691585117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,7168,11.388818528917101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,12288,0.04459377792146471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,51200,0.42919288741217715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,2048,0.03137866655985514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,768,0.003551111039188173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,8192,0.3586933347913954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,12288,0.23323822021484375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,2560,0.008661333057615492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,65536,0.7535386615329318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,7168,0.036770668294694685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,51200,2.32968627081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,7168,0.013818666338920593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,64,0.008611555728647444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,3584,0.0861555536588033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,1024,0.0491866668065389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,512,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,1536,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,10240,0.0478151109483507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,10240,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,8192,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,1024,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,128,0.0070266665683852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,65536,0.08656622303856744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,51200,2.5775538550482855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,6144,0.017997332745128207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,2560,0.07308622201283772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,65536,1.521040916442871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,32,0.013410666750537025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,6144,0.02905866503715515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,2560,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,16384,0.2570426728990343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,3584,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,3072,0.12758044401804605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,64,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,4096,0.01718222267097897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,51200,0.2315546671549479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,8192,0.05389688743485344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,768,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,1024,0.14152267244127062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,1024,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,768,0.0028151110228565005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,51200,0.050322665108574756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,16384,0.11115733782450359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,1536,0.03086755673090617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,65536,0.7155315611097547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,65536,0.07616177532407972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,128,0.025811556312772963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,256,0.015630222029156156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,51200,0.22685776816474068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,1536,0.028039111031426325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,51200,0.05094666613472832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,1024,0.007204444044166141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,5120,0.011231111155615913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,16384,0.019575110740131803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,12288,0.026774222652117412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,10240,0.0999164448844062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,51200,0.273214234246148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,51200,3.126106686062283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,2048,0.010211555494202508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,2560,0.011700444751315646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,7168,0.06308622492684259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,128,0.06423199839062162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,1024,0.03968800107638041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,32,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,3072,0.01740800009833442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,1024,0.010782221953074137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,8192,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,8192,0.04564355479346382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,5120,0.010391111175219217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,1536,0.011071999867757162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,51200,20.356657240125866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,8192,0.024783111280865137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,51200,0.444193787044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,3584,0.024199111594094172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,8192,0.027810666296217177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,8192,0.027859555350409612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,16384,0.020238222347365487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,1024,0.004418666578001446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,768,0.006791111081838608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,5120,0.04150311152140299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,2560,0.07310399744245741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,16384,0.19509866502549914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,8192,0.013506666653686099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,1536,0.03820533222622342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,2560,0.04283110962973701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,128,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,512,0.02295733326011234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,3584,0.023909333679411147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,128,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,2560,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,12288,0.10877866877449884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,16384,0.10782844490475124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,4096,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,12288,0.28314844767252606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,10240,0.18692889478471544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,512,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,6144,0.43628621101379395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,512,0.005642666584915585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,12288,0.031192888816197712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,7168,0.025981333520677354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,256,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,8192,32,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,7168,0.026009778181711834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,3072,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,256,0.019983111156357657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,3584,0.2916693422529432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,10240,0.050438221957948476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,8192,0.03295733200179206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,2048,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,1024,0.006314666734801398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,512,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,64,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,128,0.015844444433848064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,6144,0.12025600009494358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,4096,0.04085777865515815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,3072,0.0834311114417182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,6144,0.9628213246663412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,256,0.00797155582242542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,1024,0.01568000018596649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,8192,0.46790398491753477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,3072,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,6144,0.0170151111152437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,2048,0.1233031087451511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,2048,0.0225582222143809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,65536,3072,0.042499555481804736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,32,0.019919999771647982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,16384,0.021703110800849065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,3584,0.029838220940695867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,12288,0.42645777596367734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,1024,0.025940444734361436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,1536,0.012997333374288348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,3584,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,65536,6.107607099745009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,128,0.014509333504570855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2560,8192,0.02330133318901062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,4096,0.059396445751190186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,4096,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,768,0.25411732991536456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,12288,0.10080799791547988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,256,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,512,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,32,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2048,128,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,1536,0.030159999926884968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,2560,0.04069066709942288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,65536,0.6869920094807943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,6144,0.009385777844323052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,7168,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,7168,0.5870746506585015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,16384,0.2071982224782308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,2560,0.025049777494536504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,256,0.026075555218590632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,6144,0.13710578282674155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,3584,0.953738636440701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,7168,0.011415111521879831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,7168,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,7168,0.049976888630125255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,1024,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,1024,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,51200,13.437017652723524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,512,0.009645333720578088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,16384,0.033795555432637535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,65536,3.9642346700032554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,10240,0.06537689103020562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,8192,256,0.0035706667436493766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,1536,0.015656888484954834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,2560,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,65536,0.47094223234388566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,51200,0.2746284537845188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,4096,0.03374933203061422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,512,0.014021333720948962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,6144,0.038811554511388145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,7168,0.022554667459593877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,51200,0.38553065723843044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,768,0.03235022226969401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,7168,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,128,0.00924622184700436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,5120,0.11563377910190159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,4096,0.029824889368481104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,7168,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,2560,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,512,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,1024,0.034929778840806745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,16384,0.01644888851377699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,32,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,32,0.03946844405598111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,32,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,51200,0.129786663585239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,1536,0.03779733180999756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,3584,0.04137244489457872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,1024,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,51200,0.10132622056537205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,5120,0.029976000388463337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,1024,0.003532444437344869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,65536,0.06682933039135404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,16384,0.9495742585923937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,4096,0.024674667252434626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,6144,128,0.035170667701297335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,65536,0.05736622545454237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,6144,0.011050666371981302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,4096,0.005072000126043956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,3584,0.038521776596705117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,4096,0.007962666451931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,8192,0.20812800195482042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,8192,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,256,0.01683466633160909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,6144,0.04096711013052199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,768,0.020035554965337116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,7168,0.04353955719206068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,2048,0.07294488615459867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,1024,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,7168,1.1237493091159396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,7168,0.05088355475001865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,512,0.02036088870631324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,16384,0.045681777927610606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,3584,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,2048,0.008038221961922115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,16384,4096,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,1024,0.0700862209002177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,1024,0.01956355571746826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,2048,0.019542222221692402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,7168,0.036312000619040594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,8192,0.11460444662306045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,6144,0.016207999653286405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,128,0.06208622455596924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,5120,0.018571555614471436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,128,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,8192,0.06370311313205294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,3072,0.013755555782053204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,6144,0.0099973330895106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,128,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,16384,0.01982666717635261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,65536,0.46954668892754453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,6144,0.9255680508083768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,8192,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,8192,0.5963955455356175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,128,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,6144,0.01942488882276747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,1024,0.013922666509946188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,256,0.01480888823668162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,7168,0.03975733452373081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,256,0.024681778417693243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,512,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,10240,0.042152000798119434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,2560,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,5120,0.010796444283591377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,10240,0.24039199617173937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,128,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,3584,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,10240,0.028584000137117174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,32,0.05761155817243788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,768,0.02842666705449422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,32,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,1536,0.023584889041052923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,10240,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,768,0.028761777612898085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,768,0.010622222390439777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,51200,0.05134133497873942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,7168,0.13280888398488364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,51200,0.04610400067435371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,3584,0.01991555591424306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,6144,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,256,0.07954133219189115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,1536,0.08846844567192926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,65536,0.047669334544075855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,7168,3.5767190721299915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,32,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,7168,0.0533537765343984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,128,0.023183110687467787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,2048,0.03006311257680257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,16384,0.05859555800755819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,65536,0.45337597529093426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,32,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,6144,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,6144,0.022893332772784766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,256,0.01926933394538032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,256,0.028929776615566675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,65536,128,0.010279110736317104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,256,0.006481777876615524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,7168,0.09285955296622382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,8192,0.025459556115998164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,16384,0.017482666505707633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,2048,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,5120,768,0.01699822147687276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,12288,0.011649777491887411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,16384,0.025215110845035974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,7168,0.10458933644824558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,1536,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,12288,0.03885066509246826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,4096,0.06782666842142741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,12288,0.03055022160212199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,3072,0.03082488973935445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,16384,0.0439208878411187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,64,0.02642222245534261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,2560,0.020682666036817763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,3072,0.14720977677239314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,768,0.003571555432346132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,256,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,6144,0.04363822274737888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,64,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,65536,1.549107551574707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,512,0.011649777491887411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,8192,0.03986755675739712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,4096,0.04810133245256212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,7168,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,1536,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,3072,0.03439022103945414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,3072,0.02769688930776384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,1536,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,4096,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,512,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,256,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,1536,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,4096,0.03065599997838338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,32,0.005611555443869696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,65536,0.38084620899624294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,6144,0.05680266353819105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,128,2560,0.009400889277458191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,32,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,1024,0.0107004443804423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,1024,0.027219555444187585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,8192,0.013637333280510373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,12288,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,16384,0.019460444649060566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,64,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,768,0.03334755698839823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,2048,0.08575555351045396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,51200,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,16384,0.06112088759740194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,2048,0.04323022233115303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,64,0.006216889040337668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,6144,0.036088890499538846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,51200,0.07671199904547797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,512,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,65536,0.12737511263953313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,65536,0.08751644690831502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,8192,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,51200,1.9172390831841364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,7168,0.1329022248586019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,2048,0.02539288832081689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,768,0.03461422191725837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,256,0.015108444624476962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,12288,0.014027555783589682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,65536,0.07168622149361505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,3584,0.02665600015057458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,3072,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,3072,0.019225777851210702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,128,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,1024,0.006208000083764394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,51200,0.013570666313171387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,8192,0.014089778065681458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,256,0.012494222157531314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,7168,0.01144266708029641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,8192,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,2560,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,51200,0.04661866691377428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,51200,0.9106186760796441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,8192,0.036650667587916054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,16384,0.24590754508972168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,128,128,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,5120,0.015146666102939181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,2048,0.012321777641773224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,65536,0.05836977561314901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,32,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,5120,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,3584,0.006828444285525217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,1536,0.019134221805466544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,3072,0.021317332983016968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,16384,0.07973511351479425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,2560,0.059917330741882324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,3072,0.19416622320810953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,2048,0.00665511108107037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,1536,0.00665866666369968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,3072,0.060497780640920006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,3584,0.023447111248970032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,10240,0.08137066496743096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,12288,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,3584,0.2030142280790541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,1536,0.023023999399609033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,8192,0.027798222170935735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,10240,0.023378666904237535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,768,0.003489777859714296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,128,0.01957244508796268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,512,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,65536,0.6444480154249403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,5120,0.027028444740507338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,3072,0.013864888913101621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,8192,0.02500711050298479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,512,0.0026337777574857077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,512,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,12288,0.35472088389926487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,12288,0.017416889468828838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,7168,0.03702755437956916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,512,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,32,0.005640888793600931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,8192,0.014670222997665405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,4096,0.6338444285922581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,768,0.0032177778581778207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,768,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3584,3072,0.05173066589567396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,4096,0.024307555622524683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,5120,0.06083555354012383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,768,0.01646399994691213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,5120,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,128,0.004229333251714706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,512,0.015525332755512662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,256,0.0028906667398081887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,2048,0.05356088942951626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,8192,0.13489866256713867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,3584,0.027330666780471802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,65536,11.076861063639322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,512,0.016864889197879367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,128,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,3072,0.08066133658091228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,2048,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,32,4096,0.004555555681387584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,768,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,51200,0.9987884097629123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,51200,1.1000400119357638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,8192,0.11893867121802436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,3072,0.01651022169325087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,128,0.0031822222388452957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,5120,0.039910223748948835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,6144,0.11433688799540202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,65536,0.6634204652574327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,10240,0.13750133249494764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,256,0.0053031109273433685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,12288,0.060200889905293785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,65536,0.3010488881005181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,4096,0.11639732784695095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,2560,0.04791288905673557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,65536,0.12659377521938747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,3584,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,8192,0.24026044209798178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,128,0.014452444182501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,2048,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,768,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,16384,0.04808800088034736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,5120,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,6144,0.02612888813018799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,12288,0.018547554810841877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,3584,0.04008000095685323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,4096,0.1471031109491984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,3584,0.24795820977952746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,1024,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,12288,0.08700978093677098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,4096,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,4096,0.09269777933756511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,12288,0.3552106751335992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,16384,0.010751999914646149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,768,0.0418906675444709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,51200,0.060552000999450684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,512,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,3072,0.055926223595937095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,7168,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,5120,0.08575733502705891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,7168,0.07505333423614502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,65536,0.32183289527893066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,768,0.01685333251953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,6144,0.04286311070124308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,10240,0.029901332325405542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,2048,0.0072266666425599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1024,256,0.016185777054892648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,1536,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,8192,0.3015742301940918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,768,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,4096,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,5120,0.017652443713612027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,51200,0.21990133656395805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,10240,0.015056000815497505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,4096,0.04657155606481764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,65536,1.276679144965278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,6144,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,10240,0.07608266671498616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,2560,0.033378667301601835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,10240,0.03980088896221585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,4096,0.08112711376614042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,2048,0.06575110885832045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,768,0.04624355501598782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,256,0.014801777071423002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,1024,0.012106666962305704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,768,32,0.0018782222436534034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,10240,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,7168,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,7168,0.06921066840489705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,3584,0.007672000262472365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,10240,1.089079115125868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,3584,0.08505688773261176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,768,0.03590933481852213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,7168,7168,0.029093331760830347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,4096,0.04134577843878005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,16384,0.0226239992512597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,256,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,2048,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,512,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,5120,0.01685422162214915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,3584,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,3584,0.01831288966867659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,128,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,512,0.056361774603525795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,768,0.0028168888141711554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,7168,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,2048,0.07776177591747709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,128,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,10240,0.012610666453838348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,65536,0.22627467579311797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,6144,0.5231128798590766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,2048,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,128,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,2560,0.253532436158922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,768,0.004848000076082018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,7168,0.03046577837732103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,65536,0.010271111296282874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,128,768,0.0031991110493739447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,128,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,3584,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,512,0.012095111111799875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,128,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,32,0.013612444202105204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,51200,0.05961777766545614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,64,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,8192,3584,0.027505778604083594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,2560,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,32,0.009335999687512716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,3584,0.017162667380438913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,51200,0.14355111122131348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,768,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,10240,0.015222221612930298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,6144,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,5120,0.18048889107174346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,10240,0.2578710979885525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,128,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,768,0.013067555924256643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,10240,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,128,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,5120,0.02504355543189579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,3072,0.013798221945762634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,65536,0.5346835454305013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,65536,0.08710310856501262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,8192,0.07565689086914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,1536,0.026766222384240892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,3072,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,768,0.022654222117529974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,12288,0.056523554854922824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,1536,0.01888444523016612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,256,0.012498666842778524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,10240,0.16390489207373724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,8192,0.06846844487720065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,768,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2048,4096,0.02573866645495097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,12288,0.01607822212908003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,12288,0.02961155441072252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,2560,0.5074675348069932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,16384,0.07378044393327501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,8192,2.3907688988579645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,65536,0.46776713265313047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,1536,0.03295200069745382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,256,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,65536,0.7356719970703125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,512,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,2560,0.10521244340472752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,4096,0.021889777647124395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,16384,0.029398222764333088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,3584,0.009787555370065901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,768,0.036229332288106285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,2560,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,8192,0.30130844646030003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,128,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,7168,3072,0.10342222452163696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,4096,12288,0.03136266602410211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,2048,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,12288,0.08803377548853557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,4096,0.028213332096735638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,6144,0.011061333119869232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,32,0.0018586667461527719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,2048,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,3584,0.15277243985070124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,512,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,32,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,3584,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,6144,0.021148444877730474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,51200,0.21871199872758654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,7168,0.035290665096706815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,2560,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,7168,0.018163555198245578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,5120,0.010547555155224271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,1024,0.015471999843915304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,1536,0.012005333271291522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,3584,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,65536,0.06961599985758464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,12288,0.04723822077115377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,4096,0.14527289072672525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,3584,0.009179555707507664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,3584,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,2560,0.012614222036467658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,7168,0.009399111072222391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,16384,0.09085688988367717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,512,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,256,0.024680000212457445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,8192,0.047931555244657725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,1024,0.014547554983033074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,2560,0.02515999972820282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,512,0.0355751117070516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,6144,0.0418488879998525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,2048,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,3072,0.017782222893502977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,6144,0.5449768702189127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,6144,0.0957075556119283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,768,0.005684444473849402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,2560,0.021408889028761122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,10240,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,64,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,768,0.0028000000036425064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,8192,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,128,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,6144,0.024680000212457445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,7168,0.025627555118666753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,3072,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,3072,0.015747555428081088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,51200,0.3873599900139703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,6144,0.012598222328556908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,7168,0.1038702196545071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,7168,0.0233608881632487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,3584,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,2560,0.004912000149488449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,5120,0.21918843852149117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,5120,0.018907555275493197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,5120,1536,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,3584,0.041689776711993747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,1536,0.011674666570292579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,4096,0.014127110441525778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,7168,0.1547795534133911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,51200,0.13935911655426025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,65536,0.03548889027701484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,1536,0.022502221994929846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,2560,0.004934222333961063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,7168,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,128,0.03247733248604669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,8192,0.027461333407296076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,6144,7168,0.04185600082079569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,10240,0.3608115514119466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,2048,0.012800888882742988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,10240,0.16846044858296713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,1536,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,5120,0.033189333147472806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,3072,0.012794666820102267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,65536,0.098380446434021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,128,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,256,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,128,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,12288,0.014532445205582513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,3584,0.043475555049048535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,256,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,51200,0.2745137744479709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,65536,0.19886222150590685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,512,0.015984889533784654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,256,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,2560,0.004860444615284602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,10240,0.5458524492051866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,128,0.004566222015354368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,2560,0.026753778258959453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,1024,0.017638223038779367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,6144,0.004233777936961916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,6144,0.013487999637921652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,12288,0.0887591110335456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,256,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,64,0.006650666809744305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,1024,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,4096,5120,0.02235377828280131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,10240,0.7867306603325738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,1024,0.00830488900343577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,2048,0.038373334540261164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,256,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,12288,0.06209511227077908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,8192,2560,0.049695998430252075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,2560,0.09286310937669541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,3072,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,768,0.13230844338734946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1536,3584,0.013260444833172692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,6144,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,2048,0.06481689214706421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,16384,0.0120319997270902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,4096,0.16674221886528862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,1536,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,512,0.014155555102560254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,3584,0.02089866664674547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,65536,0.12268088923560248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,51200,8.257688734266493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,5120,0.13005155987209743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,512,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,10240,4096,0.014399111270904541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,128,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,4096,0.23159911897447374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,256,0.01109155515829722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,1024,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,51200,0.041839109526740186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,768,0.019536889261669584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,4096,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,5120,0.24208354949951172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,4096,8192,0.08936533663007949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,1024,0.008560000194443597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,5120,0.0384897788365682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,256,0.009906666974226633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,5120,0.06594044632381864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2048,12288,0.016530666086408827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,4096,0.022646221849653456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,4096,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,2560,0.008669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,2048,0.027518222729365032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,256,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,12288,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,128,0.01590844492117564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,8192,0.1442613336775038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,8192,0.014106666048367819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,1024,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,1024,0.0036728887094391715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,10240,1.4167004691229925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,10240,0.06939378049638537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,1024,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,1024,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,16384,0.051983998881445996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,3072,0.026102221674389307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,65536,0.34900445408291286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,65536,0.061074667506747775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,768,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,256,16384,0.14292800426483154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,2048,0.04638400011592441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,2048,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,5120,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,8192,0.037851555479897395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,8192,0.01072799993885888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,16384,0.3279866642422146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,2560,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,16384,0.08217422167460124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,256,128,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,51200,0.052060445149739586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,128,0.012346666720178394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,3072,0.03074044320318434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,8192,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,2560,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,2048,0.0854311121834649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,7168,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,32,0.003548444559176763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,10240,0.00519822206762102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,32,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,5120,0.025608888930744592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,5120,0.04377866784731547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,12288,0.06107644240061442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,4096,0.00775911079512702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,4096,0.09972355763117473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,2048,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,64,0.029242667886945937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,1024,0.029919998513327703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,12288,0.020647999313142564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,6144,0.011250666446155973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,768,0.0105253333846728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,3584,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,256,0.0032008888406885993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,256,0.06044266621271769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,4096,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,32,0.08664888805813259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,3584,0.1163075500064426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,65536,0.41945955488416886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,51200,0.03155466583040025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,12288,0.013413333230548434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,3072,0.00889866633547677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,51200,0.15748355123731825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,768,0.012889777620633444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,128,0.013101333545313941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,16384,0.08593866560194229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,128,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,8192,65536,0.1467911137474908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,16384,0.019164444671736825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,1024,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,32,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,10240,6144,0.019560888409614563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,65536,0.06888533300823636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,128,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,3072,0.035904887649748064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,6144,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,768,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,4096,0.026167111264334783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,65536,2.2363866170247397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,128,0.011261333194043903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,3072,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,768,0.014550222290886773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,7168,0.11266756057739258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,12288,0.04997155401441786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,8192,0.012666666673289405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,1024,0.1270000007417467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,2048,0.023072888453801472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,65536,1.134126239352756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,12288,7168,0.03404088815053304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,65536,0.15787377622392443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,12288,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,10240,0.19992710484398735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,10240,0.9167155159844292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,65536,3.0963573455810547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,512,0.02295555505487654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,65536,0.20264355341593424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,768,0.029048889875411987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,8192,0.026370666093296472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,1024,0.3381804360283746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,65536,0.2884684403737386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,1536,0.34192532963222927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,1024,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,1024,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,256,0.014797333214018079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,65536,0.5652986632453071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,512,0.018550222118695576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,7168,0.01464355488618215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,3072,0.01799199978510539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,4096,0.025565332836574976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,3584,0.017343110508388944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,7168,0.03402933478355408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,128,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,3072,0.06626311275694105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,512,2048,0.011410666836632622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,5120,0.03363200028737386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,3584,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,1024,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,4096,0.21683022711012098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,51200,0.04006310966279771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,2560,0.018259555101394653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,12288,0.37259822421603733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,10240,0.029831111431121826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,12288,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,12288,256,0.0038204445607132385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,32,0.0017946666727463405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,12288,0.023204444183243647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,3072,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,256,0.015449777245521545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,3072,0.0063866666621632045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,2048,0.009370666411187913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,12288,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,32,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,51200,128,0.1209919982486301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,6144,0.04074933462672763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,16384,0.248468452029758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,1536,0.023288889063729182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,7168,0.035544888840781316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,16384,0.033259554041756526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,7168,0.06250311268700494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,6144,51200,0.5713173548380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,1024,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,5120,0.008852444589138031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,65536,7168,0.1323875586191813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,1536,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,10240,0.06702222426732381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,2560,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,128,0.0063031112982167145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,3072,0.06926311386956109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,768,0.02015199926164415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,512,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,7168,0.008280889027648503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,2560,0.03977155685424805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,6144,0.025104888611369666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,8192,0.04707644383112589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,1024,0.0031848889258172777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,512,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,6144,0.26967022154066295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,65536,0.34724799791971844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,32,0.007110222346252865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,256,0.03572533196873135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,3584,0.024141333169407312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,3072,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,51200,1.6042711469862196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,256,0.04497155547142029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,16384,0.05145777927504646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,32,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,3584,0.020911999874644812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,4096,0.3163928985595703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,768,0.0075599998235702515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,51200,0.0712968905766805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,512,0.10098755359649658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,3072,0.037295109695858426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,3072,0.08574044704437256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,256,0.011341333389282227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,10240,0.08844533231523301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,1024,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,7168,0.13765954971313477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,3584,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,8192,0.35263554255167645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,65536,0.152328888575236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,12288,0.04019022319051955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,1024,0.00315022220214208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,51200,0.13039555814531115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,51200,0.06040177742640177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,3072,0.023003555006451074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,128,0.013573333621025085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,65536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,16384,0.05787110990948147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,128,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,256,0.016699555847379897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,2560,0.15061510933770075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,128,0.01482311056719886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,256,0.01738666660255856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,16384,768,0.054987556404537626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,32,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,32,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,1024,0.022503111097547743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,4096,0.028875556257035997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,1024,0.00600000015563435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,1024,0.02794755498568217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,3072,0.025643555654419795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,5120,0.01928622192806668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,256,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,16384,0.07454578081766765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,128,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,128,0.014151111245155334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,128,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,2048,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,256,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,512,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,2560,0.01200711147652732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,3584,0.017300445172521803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,16384,0.057061334451039634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,7168,0.05128711130883959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,32,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1024,1024,0.011204444699817233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,256,0.006034666465388403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,128,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,65536,1.2378933164808485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,2560,0.03371644351217482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,3072,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,256,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,8192,0.020928889513015747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,2048,0.012788444757461548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,768,0.0214044451713562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,8192,0.013538666897349887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,7168,0.010678222609890832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,65536,2.7772755093044705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,8192,0.1544142166773478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,4096,0.037421332465277776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,2048,0.019917334119478863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,8192,0.06031555599636502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,256,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,768,2560,0.023079999619060095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,65536,0.19954933060540092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,16384,0.02665511104795668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,768,0.015804444750150044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,512,0.02014666630162133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,4096,0.06500977940029569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,65536,0.13443733586205378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,7168,3.0381840599907766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,3584,0.008629333641793992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,7168,0.022202667262819078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,8192,0.025126222107145522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,65536,0.06292800108591716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,12288,0.5392879909939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,10240,0.17720977465311685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,64,0.0018684443914228016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,512,0.0192239996459749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,6144,0.025191111697090998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,5120,0.014031999640994601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,768,0.03686133358213637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,2560,0.03984355595376756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,256,0.0036311112344264984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,4096,0.24034754435221353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,10240,0.06897866725921631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,16384,0.4014435609181722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,8192,0.016018667154841952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,5120,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,7168,0.02991022335158454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,1024,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,65536,0.5654613176981608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,1536,0.008269333177142674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,6144,0.022671999202834234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,8192,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,2560,0.01148622234662374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,4096,0.011482666763994427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,1024,0.05125688844256931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,256,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,2560,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,8192,0.25428978602091473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,64,0.00499199993080563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,3584,0.007975111405054728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,64,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,4096,0.029490666257010564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,10240,0.050511999262703784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,16384,7168,1.5001128514607747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,16384,0.48751375410291886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,1536,0.003581333491537306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,10240,0.2330533398522271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,12288,0.05252977874543932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,5120,0.07434044281641643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,6144,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,5120,0.014134221606784396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,3072,0.005614222337802251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,65536,3.6373617384168835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,5120,0.028507557180192735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,51200,0.08037244611316256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,64,0.02163822286658817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,2560,0.02128177715672387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,10240,0.024147555232048035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,16384,0.03311555584271749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,512,0.0328053335348765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,768,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,5120,0.028206222587161597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,1536,0.13235467010074192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,12288,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,1024,0.01348444405529234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,5120,0.009649777577983009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,12288,0.20700621604919434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,16384,0.019863999552196927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,64,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,2560,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,12288,0.12585600217183432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,51200,0.07104621993170844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,32,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,4096,0.055162668228149414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,6144,0.03349688980314467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,3584,0.15828888946109348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,8192,0.35065510537889266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,4096,0.0347031123108334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,16384,0.06523822413550483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,3072,0.032116442918777466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,3584,0.03403644429312812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,12288,0.01705777810679542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,2560,0.20827199353112114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,512,0.01684533390733931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,65536,3.1649191114637585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,16384,0.03546844588385688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,128,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,3584,0.017659554878870647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,2560,0.05170222123463949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3584,8192,0.01575555569595761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,1024,0.014512888259357877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,5120,0.023538667294714186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,16384,12288,0.3452204333411322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,16384,0.030135999123255413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,64,0.001830222292078866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,51200,0.5019609133402506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,3072,0.021535111798180476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,1536,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,2048,0.016161777906947665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,3072,0.006497777584526274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,5120,0.022633777724372015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,1536,0.01332711097266939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,10240,0.1975555552376641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,3584,0.027433777848879497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,65536,0.4083404541015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,64,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,65536,0.04961511161592272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,256,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,256,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,12288,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,4096,0.034978667894999184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,12288,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,1536,0.022820444570647344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,2048,0.0052266667286554975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,7168,0.09215822484758164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,1024,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,6144,0.04126222266091241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,12288,0.018087110585636564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,5120,0.15032177501254612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,1536,0.044545776314205594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,1024,0.05323733223809136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,1024,0.04042311178313361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,5120,0.11110577980677287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,4096,0.06764444377687243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,128,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,512,7168,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,51200,0.19917333126068115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,2560,0.023777777949968975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,1536,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,512,0.07058844301435682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,16384,0.18973244561089408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,3584,0.03356355428695679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,6144,0.022844443718592327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,5120,0.19704443878597686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,256,0.0021457777669032416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,4096,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,5120,0.017761778500345018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,768,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,1024,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2048,32,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,8192,0.23613156212700737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,7168,0.02864177690611945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,512,0.003216888962520493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,12288,0.3495813210805257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,5120,6144,0.013388444152143268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,1024,0.018535999788178336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,512,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,3584,0.009006222089131674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,1024,0.020564445190959506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,3072,0.04530044396718343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,128,0.0021786666992637846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,16384,5120,0.02197155521975623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,4096,0.15394222736358643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,51200,1.4740506278143988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,6144,0.027032888597912256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,16384,0.017307556337780423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,3072,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,5120,0.05005066593488058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,2048,0.04087288843260871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,3584,0.017421333326233756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,10240,0.3149511019388835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,5120,0.03430755601988898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,3584,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,3072,0.09295555618074204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,12288,0.0843120018641154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,512,0.0370399985048506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,6144,0.02329155637158288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,3072,0.009332444104883406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,3584,0.05572799841562907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,5120,0.18044533994462755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,12288,0.03865689039230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,256,0.002586666701568498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,8192,0.0136835558546914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,32,0.003510222252872255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,2560,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,2048,0.08529066377215916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,128,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,5120,0.013663999736309052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,8192,0.011971555650234222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,768,16384,0.035191999541388616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,65536,0.08329421944088407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,2048,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,3584,0.03348355491956075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,65536,0.13202399677700466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,512,0.003192000091075897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,10240,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,7168,0.06274666388829549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,16384,0.025145777397685583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,8192,0.04439999990993076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,5120,0.028423110644022625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,1536,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,768,0.02238133384121789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,512,128,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,1024,0.07463644610510932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,8192,0.011752888560295105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,12288,0.01623911162217458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,4096,0.024031110935741003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2048,32,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1536,768,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,768,10240,0.08085333638721041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,12288,0.15423911147647434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,6144,0.17268178198072645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,128,0.013166222307417126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,5120,65536,0.07403466436598036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,1024,0.02570844524436527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,128,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,16384,0.07131110959582858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,10240,0.044616887966791786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,3072,0.012791111237472959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,6144,0.04821866750717163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,4096,0.04070311122470432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,128,0.030459556314680312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,2048,0.09454044368531968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,5120,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,1536,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,256,0.00907466643386417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,6144,0.018979556030697294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,10240,1.6503243976169164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,8192,0.15368266900380453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,7168,0.03759200043148465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,256,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,768,0.007043555378913879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,16384,0.45107290479871964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,6144,0.06992089086108737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,64,0.1796097755432129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,12288,0.8231431113349067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,6144,0.035134222772386335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,6144,0.10679999987284343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,128,0.016535111599498324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,256,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,7168,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,1024,0.006344888773229387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,6144,0.015378667248619927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,4096,0.05211288730303446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,128,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,6144,6144,0.019191111127535503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,6144,0.012982221941153208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,2048,0.018971555762820773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,1536,0.009692444569534725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,128,0.014183110660976835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,7168,0.025042666329277888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,65536,0.19532355997297499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,32,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,7168,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,12288,0.03580177823702494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,2048,0.011172444456153445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,512,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,16384,0.02854222059249878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,7168,0.339483552508884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,1024,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,12288,0.014765333798196582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,16384,0.041606220934126116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,64,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,256,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,12288,0.0280693338976966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,256,0.006527110934257507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,4096,0.0832915571000841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,1536,0.03973333372010125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,256,0.010992888775136737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2560,1536,0.03758577836884393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,3584,0.01867022282547421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,512,0.005713777823580637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,5120,0.01854933301607768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,6144,0.030341333813137476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,51200,0.10888622204462688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,8192,0.04396355483267042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,3072,0.017331555485725403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,3072,0.014947555131382413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,8192,0.08130666944715711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,3584,0.020459555917316012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,8192,0.016196444630622864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,32,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,8192,0.02053955528471205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,1536,0.006414222220579783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,768,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,4096,0.026946667167875502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,51200,0.07777866390016344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,6144,0.04680622286266751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,768,0.009659555223253038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,16384,0.02978488802909851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,1536,0.10630844699011908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,2560,0.09648266765806411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,5120,0.018858666221300762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,7168,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,3072,0.00980088859796524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,10240,0.07151555352740817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,12288,0.3448817729949951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,7168,0.08865244521035089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,3072,0.01888799998495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,3584,0.8122933175828723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,3072,0.006204444501135085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,51200,0.06300622224807739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,16384,2.0804311964246964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,10240,0.07601333326763578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,512,1536,0.005212444398138259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,65536,0.08960444397396511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,3584,0.012715555727481842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,32,0.018346667289733887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,3072,0.11052355501386855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,65536,8.570548163519964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,128,0.0138586668504609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,64,0.002524444419476721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,16384,0.029790222644805908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,2560,0.02165066699186961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,3584,0.052894221411810986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,256,0.008639111287064022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,2560,0.0052471111218134565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,5120,0.008728000024954477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,3584,0.07910933097203572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,51200,0.1549288829167684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,256,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,5120,0.10804977681901719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,32,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,1536,0.008950221869680617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,12288,0.2886008951399061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,1024,0.02257333364751604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,1024,0.014855111638704935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,768,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,4096,0.011369778050316704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,6144,0.047360890441470675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,51200,51200,4.2910105387369795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,3584,0.640419536166721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,2560,0.010104888843165504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,65536,0.1768702268600464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,256,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,768,0.018294221825069852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,65536,0.6571635670132107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,3072,0.011125333607196808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,3072,0.03349688980314467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,5120,0.01722133325205909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,3072,0.015528000063366361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,7168,0.19957866933610705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,512,0.011773332953453064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,32,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,2560,0.01959466603067186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,6144,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,256,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,6144,0.1424942281511095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,6144,0.02125511070092519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,128,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,3584,0.024335111180941265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,8192,0.032568001084857516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,32,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,7168,0.02421155571937561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,7168,0.05500266618198819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,16384,0.0975724458694458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,51200,0.26711644066704643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,7168,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,12288,0.014306666122542487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,5120,0.10383911265267266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,3072,0.0333280000421736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,8192,0.18858133422003853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,8192,0.2583040131462945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,2560,0.11094044314490424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,7168,0.10803644524680243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,3584,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,64,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,2048,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,10240,4096,0.03861155443721347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,128,0.016198222835858662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,8192,0.5101751221550835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,1536,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,10240,0.1314506663216485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,4096,0.015604444675975375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,3584,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,3072,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,4096,0.024157333705160353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,4096,0.03197688857714335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,256,0.015848888291252982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,1536,0.5923804706997341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,51200,0.43351467450459796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,5120,0.013367111484209696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,7168,0.6391644477844238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,51200,0.10245866907967462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,16384,0.023063111636373732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,65536,0.08822933170530532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,2048,0.007328000333574083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,65536,0.18896799617343477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,51200,0.12573777304755315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,128,0.015862221519152325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,4096,0.058203558127085366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,4096,0.3000657823350695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,8192,0.013959111438857185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,1024,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,3584,0.02329866588115692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,3072,0.07909511195288764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,16384,0.34047910902235246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,512,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,6144,0.036836445331573486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,1536,0.021754667162895203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,5120,0.009418666362762451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,51200,0.05570844146940443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,3584,0.029297779003779095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,2048,0.006673777682913675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,2560,0.014879999889267815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,128,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,7168,0.020743111769358318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,2560,0.017657778329319425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,128,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,512,0.04270222120814853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,3584,0.020967110991477966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,128,0.007332444190979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,1536,0.03119644522666931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,5120,0.022495110829671223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,51200,0.0688942207230462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,1536,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,8192,0.022459555003378127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,12288,0.01795733306143019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,16384,0.09607466724183823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,3584,0.015807110402319167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,10240,0.13118933306799993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,12288,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,2560,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,4096,0.008568888737095727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,12288,0.164429333474901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,32,0.0017786667578750187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,16384,0.0123848890264829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,2048,0.04320889049106174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,6144,0.008833777573373582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2048,16384,0.021369778447681006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,1024,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,256,0.05848089191648695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,65536,0.07885777950286865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,51200,0.09422577752007379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,7168,0.02752799954679277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,512,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,5120,0.008279999925030602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,3584,0.04423822296990288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1024,64,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,512,0.03032711148262024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,768,0.021624000536070928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,8192,0.1532177792655097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,10240,0.3709004455142551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,768,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,10240,0.022107554806603327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,7168,0.021361778179804485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,64,0.006106666806671355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,2560,0.010839111275143094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,65536,0.45853598912556964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,10240,0.04204533497492472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,8192,0.05334844523006015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,10240,0.03945066531499227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,768,0.03724977705213759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,16384,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,12288,0.01924088928434584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,16384,2.8868221706814237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,65536,0.1511955526140001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,6144,0.019384000036451552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,1024,0.020252444677882724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,3072,0.04141244292259216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,65536,0.4278506702846951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,12288,0.007990222010347579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,12288,0.018957333432303537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,6144,0.02090399960676829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,512,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,256,0.020576889316240948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,128,0.014118222726715935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,12288,0.5493146578470866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,51200,0.07384800248675875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,5120,0.10639911227756077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,5120,0.14680177635616726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,768,0.02629866699377696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,768,0.0041031113101376426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,12288,0.017535110314687092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,768,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,1024,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,65536,0.1177191072040134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,5120,0.037821332613627114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,8192,0.016003555721706815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,16384,2.9211146036783853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,6144,2560,0.008383110993438298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,7168,0.021877333521842957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,65536,1.5734817716810439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,65536,0.3621928956773546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,7168,0.016184889607959323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,7168,0.09908355606926812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,12288,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,6144,0.14948266082339817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,4096,0.019952888290087383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,5120,0.022340445054901972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,3072,0.9606471591525607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,6144,0.01220266686545478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,51200,6144,0.10545421971215142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,5120,0.036100443866517805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,65536,1.2687866422865126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,16384,0.19058667288886177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,2048,0.19880266984303793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,12288,0.4222453435262044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,12288,0.08423822455936009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,6144,0.03223022156291538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,5120,0.8836133215162488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,32,0.002116444417172008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,10240,0.04120711154407925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,2560,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,2048,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,768,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,3072,0.0678408874405755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,1024,0.02293155590693156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,1024,0.0035822221802340615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,65536,2.3371367984347873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,3584,0.008023999631404877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,768,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,16384,0.11063288980060154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,6144,0.879601796468099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,128,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,2048,0.04204710986879137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,8192,0.04850488901138306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,8192,0.011770666473441653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,2048,0.018533334136009216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,1536,0.22063999705844453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,3584,0.01110755569405026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,256,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,65536,0.08694844113455878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,5120,0.07213866710662842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,1024,0.00630044440428416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,65536,0.7473342153761121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,512,0.016519111063745286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,32,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,3072,0.05634933047824436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,12288,0.023747555083698694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,2048,0.01108088923825158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,512,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,768,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,10240,0.0912017756038242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,4096,0.014302222265137566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,65536,0.07550399833255343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,1024,0.02951822347111172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,6144,0.013473778135246701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,64,0.010317333042621613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,64,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,12288,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,768,256,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,12288,0.015275556180212231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,10240,0.022898667388492163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,2560,0.017854221993022494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,1536,0.013341333303186627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,512,0.013587555951542325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,128,0.04148177636994256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,768,0.01995733380317688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,6144,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,7168,0.08367821905348037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,1536,0.009353777600659264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,12288,0.10033422046237522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,2048,0.027219555444187585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,4096,0.008298666940795051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,5120,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,1536,0.06472088893254598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,65536,1.8179200490315754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,64,0.005199111170238919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,2560,0.03771377934349908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,512,0.034147553973727755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,256,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,2048,0.009676444033781687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,6144,0.0207360006040997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,768,0.20395378271738687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,12288,0.09675111373265584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,6144,0.01738933391041226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,256,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,7168,0.019222221440739103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,5120,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,12288,0.013765333427323235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,3584,0.006672888994216919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,8192,0.059302224053276904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,4096,0.03824977742301093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,4096,3072,0.023711111810472276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,16384,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,4096,0.016528000434239704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,128,0.053166220585505165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,5120,0.009939555492666032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,7168,0.058693336115943066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,512,0.12799378236134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,512,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,5120,0.030361778206295435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,1024,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,51200,0.06706666946411133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,128,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,768,0.01699822147687276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,2048,0.02295822236273024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,5120,0.04893689023123848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,12288,0.1466239955690172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,1536,0.017532444662517972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,2048,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,3072,0.0669999983575609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,128,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,512,0.019307555423842538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,10240,0.38461510340372723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,51200,0.11273954974280463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,256,0.005647111270162795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,10240,0.2039946715037028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,8192,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,768,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,2048,0.04051377707057529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,64,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,5120,0.2531377739376492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,768,0.003991111285156674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,8192,0.15954755412207708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,65536,0.5459635522630479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,12288,0.1559048891067505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,256,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,51200,0.0645244452688429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,10240,1.691269344753689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,6144,0.021926222576035395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,128,0.020040000478426617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,6144,0.02572711143228743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,5120,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,7168,0.1078924470477634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,32,0.006862222320503659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,5120,0.02164799968401591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,8192,0.05236533284187317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,32,0.13632622030046251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,16384,0.028725332683987085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,65536,0.06635910934872098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,2048,0.08920622534222072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,6144,16384,0.04281066523657905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,128,0.023246222072177466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,16384,30.985514322916668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,12288,0.5773190922207302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,8192,0.04048533240954081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,51200,0.06651555829577975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,128,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,65536,0.08519288566377427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,65536,0.06393866406546699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,512,0.016694222887357075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,1536,0.010199999643696679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,512,0.009639110830095079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,768,0.19447733296288383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,2048,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,3584,0.09382666481865777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,5120,0.01352355546421475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,7168,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,5120,0.009285333255926767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,4096,0.15400889184739855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,3072,0.006692444284756978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,1536,0.005576000031497743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,2048,0.02719199988577101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,1024,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,6144,0.5091866917080349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,1536,0.004239999999602635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,32,0.022892443670166865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,51200,0.45897865295410156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,3584,0.012655999925401477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,5120,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,3072,0.10810310973061456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,8192,0.053063111172782056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,51200,0.031870222753948636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,51200,0.24824799431694877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,5120,0.040425777435302734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,128,0.04010399844911363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,2560,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,7168,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,10240,0.01237422227859497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,12288,0.03244977858331468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,7168,0.0924631092283461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,4096,0.03277955452601115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,5120,0.16271200444963244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,6144,0.022636445032225713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,3584,0.15897599856058756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,5120,0.043503112263149686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,6144,0.01921333372592926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,2048,0.14305333296457926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,10240,0.48157601886325413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,512,0.011258666714032492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,6144,0.019121777680185106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,7168,0.029866665601730347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,128,0.1207546657986111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1536,512,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,32,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,65536,0.28756711218092174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,10240,0.03752977649370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,51200,0.17917688687642416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,10240,0.015133332875039844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,7168,0.013085333009560903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,3072,0.027268444498380024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,10240,0.015915556086434256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,1024,0.015876443849669564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,6144,0.0744124452273051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,3072,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,3072,0.16372266080644396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,64,0.018557333283954196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,256,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,768,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,7168,0.12227733929951985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,7168,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,4096,0.7969359821743436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,6144,0.2537582185533312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,5120,0.03404355711407132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,8192,0.4665733443366156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,768,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,1024,0.03425600131352743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,12288,0.035651554663976036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,3584,0.03405333227581448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,4096,0.033957332372665405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,7168,0.044016887744267784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,51200,0.6565724478827583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,5120,0.28747910923428005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,1536,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,8192,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,2560,0.01759111053413815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,32,0.019609777463806998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,2560,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,2048,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,3072,0.00552977787123786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,10240,0.0140684445699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,6144,0.013386666774749756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,8192,1.2690337498982747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,2560,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,256,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,12288,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,51200,0.09642844729953343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,128,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,2048,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,16384,0.024288000331984624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,6144,0.011762667033407422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,3584,0.22902933756510416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,2048,0.010538666612572141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,12288,0.06897866725921631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,10240,0.020941333638297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,51200,0.04653066727850172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,6144,65536,0.07998933394749959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,5120,0.5214791297912598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,2048,0.004857777721352047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,1536,0.021615111165576514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,7168,0.023032888770103455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,512,0.002854222224818336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,768,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,3072,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,512,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,5120,0.008791999684439765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,32,16384,0.026203556193245783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,1024,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,768,0.020066666934225295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,8192,0.06104177898830838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,16384,0.05020622081226773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,16384,0.23202133178710938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,12288,0.040970666540993586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,3072,0.006588444527652528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,12288,1.0443341996934679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,51200,0.937877337137858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,1536,0.0753306680255466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,6144,0.029162665208180744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,3072,0.07964444160461426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,10240,0.05715644359588623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,3072,0.11265867286258274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,1024,0.090393775039249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,4096,0.04964888758129544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,32,0.007213333414660559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,2048,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,16384,0.016195555528004963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,5120,0.028343998723559912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,3072,0.06328711244795057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,128,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,12288,0.01984444426165687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,10240,0.051282667451434664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,6144,0.02273155583275689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,32,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,4096,0.03087200058831109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,12288,0.07641422086291842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,1024,0.012856889102194043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,10240,0.2569439940982395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,12288,0.6070968839857314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,3072,0.20679467254214814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,7168,0.019144000278578866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,768,0.020595555504163105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,2560,0.019479999939600628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,7168,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,128,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,6144,0.016206221448050607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,16384,0.05858844518661499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,6144,0.04271555609173245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,6144,0.022259554929203455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,12288,0.02119111186928219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,8192,0.16035200489891901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,1024,0.019205333458052743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,1536,0.030922667847739324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,5120,0.045801778634389244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,3584,0.020644444558355544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,65536,1.0125928454928927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,4096,0.0274000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,7168,0.01495733360449473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,64,0.011369778050316704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,768,0.010957332948843638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,7168,0.05333244469430712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,7168,0.035805334647496544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1024,512,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,6144,0.017656000124083627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,12288,0.30659466319613987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,6144,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,5120,0.07514311207665338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,12288,0.017050666941536795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,4096,0.009528000321653154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,3584,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,5120,0.03434222274356418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,3584,0.026319111386934917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,256,0.13835644721984863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,16384,3.4459849463568797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,64,0.0019404445257451797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,128,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,1536,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,2048,0.012984889249006907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,6144,0.02609599961174859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,128,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,8192,0.024604444702466328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,12288,0.2841200033823649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,512,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,16384,0.20500977834065756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,16384,2560,0.0669155584441291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,2560,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,65536,1.4060648812188044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,16384,0.19090133243136934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,512,0.011482666763994427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,10240,16.787339952256946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,51200,0.01850311126973894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,128,0.02010577751530541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,256,0.011666666302416058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,5120,0.018239999810854595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,768,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,65536,0.15422488583458796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,768,5120,0.07045777638753255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,1024,0.040912889771991305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,64,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1024,2560,0.014329777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,768,0.014349333114094205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,1536,0.04398133357365926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,3584,0.04509422183036804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,16384,0.33286576800876194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,2560,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1536,256,0.021874666213989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,2560,0.0362115568584866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,6144,0.01274311128589842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,7168,0.02790133489502801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,8192,0.012479999827014076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,6144,0.018974221414989896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,2560,0.027866668171352808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,32,0.007294221884674496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,16384,0.038775109582477145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,5120,0.007874666816658443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,1536,0.11974578433566624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,768,0.008303999900817871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,1536,0.042245331737730235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,4096,0.02556177808178796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,8192,0.029231998655531142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,6144,0.08698044220606486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,2048,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,1024,0.03171022070778741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,1536,0.06626400020387438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,768,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,1536,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,32,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,2048,0.023668444818920557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,2560,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,768,0.007981333467695449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,65536,6144,0.2404551241132948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,5120,0.030962665875752766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,65536,0.2242977884080675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,128,0.009655111365848118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,7168,0.23562754525078666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,5120,0.13601156075795492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,2048,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,256,0.006675555474228329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,1024,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,7168,0.010375111467308467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1024,4096,0.016568889220555622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,10240,0.2931777901119656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,51200,1.8146887885199652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,4096,0.031911998987197876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,512,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,12288,0.047648890150917896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,6144,0.07424889008204143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,256,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,512,7168,0.04954222175810072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,2048,0.3119937843746609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,512,0.0028888889484935333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,12288,0.05968711111280653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,128,0.11138578255971272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,2560,0.02397866712676154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,7168,0.03103644318050808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,256,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,7168,0.1034239994155036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,768,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,768,0.014145778285132514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,7168,0.025791999366548326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,1024,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,8192,0.40478844112820095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,3072,0.018894222047593858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,2560,0.033019555939568415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,51200,0.24553423457675508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,2048,0.013227555486891003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,4096,10240,0.02773511078622606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,3584,0.011710222396585675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,6144,0.004902222090297275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,12288,0.692171573638916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,512,0.011400889191362592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,512,0.12648710939619276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,1536,0.012729778057999082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,65536,0.12949244181315103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,65536,2.3231279585096574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,4096,0.09026577737596299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,12288,0.02051999999417199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,128,0.002817777709828483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,512,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,10240,0.08683377504348755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,6144,0.11076088746388753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,3584,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,3072,0.007679999702506595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,1536,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,8192,0.014553777045673795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,3072,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,64,0.007846222155623967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,65536,0.9377608829074435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,4096,0.019279999865425956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,3584,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,10240,0.027738667196697656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,2048,0.03057510985268487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,2560,0.03558311197492812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,768,12288,0.0536702209048801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,8192,0.009291555318567488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,3072,0.006789333290523953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,256,0.00907733374171787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,8192,0.05205688873926798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,3072,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,10240,0.011702222128709158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,256,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,10240,0.02835822105407715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,6144,0.018214222457673814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,2560,0.011061333119869232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,512,8192,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,512,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,768,0.006014222072230444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,4096,0.008469333251317343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,128,0.0018604445374674266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,65536,0.7309475474887424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,3584,0.007789333661397298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,512,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,6144,0.009705777797434065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,7168,0.1057822240723504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,51200,0.42420800526936847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,512,0.04704622096485562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,3072,0.008657777474986183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,256,0.013853333062595792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,8192,0.011704888608720569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,2048,0.15255199538336858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,12288,8192,0.30401065614488393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,1536,0.02163644466135237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,2560,0.006641777853171031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,512,7168,0.07272533575693767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,4096,0.03871466716130575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,8192,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,8192,0.9577769173516167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,6144,0.0127697777416971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,7168,0.6300986607869467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,768,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,16384,0.00518488883972168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,768,0.008609777523411645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,768,0.01812266641192966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,6144,0.01700355609258016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,2048,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,7168,0.039311998420291476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,768,0.011782222323947482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,768,10240,0.025489777326583862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,1024,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,32,0.12976177533467612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,256,0.016857778032620747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,10240,0.08713333474265204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,4096,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,16384,0.01613333324591319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,128,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,128,0.0021191111041439907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,7168,0.021998221675554912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,16384,0.33969778484768337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,16384,0.04618488748868307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,4096,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,3584,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,512,0.011594666375054253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,10240,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,65536,1.872695075141059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,10240,1.755664931403266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,7168,0.14150133397844103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,512,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,2560,0.011147555377748279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,7168,0.02695288923051622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,128,0.008624888956546783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,3072,0.019338667392730713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,64,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,12288,0.025792888469166223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,128,0.015440000428093804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,10240,8192,0.3143840101030138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,65536,0.0886497762468126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,8192,0.1882399982876248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,512,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,128,0.024736889534526404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,4096,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,512,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,128,0.014479999740918478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,2048,0.02052177819940779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,6144,0.02903111113442315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,4096,0.169649773173862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,16384,0.06461333566241793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,1024,0.12822222709655762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,512,10240,0.029623111089070637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,256,0.004867555366622077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2048,2560,0.018223111828168232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,12288,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,64,0.0038364442686239877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,1024,0.008013333711359236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,32,2560,0.012726222475369772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,3072,0.08052000072267321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,16384,0.1389813290701972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,51200,0.06531644529766507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,8192,0.3487404452429877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,1536,0.029488000604841445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,2560,0.026030222574869793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,1024,0.11195733812120225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,6144,0.06331200069851346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,7168,0.010065777434243096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,3584,0.027878221538331773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,512,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,5120,0.00903911143541336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,3584,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,7168,0.12290399604373509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,512,0.006151999864313338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,768,0.006096000058783426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,1024,0.004031110968854693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,12288,0.19873421721988252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,768,0.01850311126973894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,3584,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,2560,0.00574933323595259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,3584,0.07572444279988606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,1024,0.036835554573271014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,3584,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,1024,0.004252444538805219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,3584,1.180088890923394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,2048,0.010690666735172272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,32,0.004270222038030624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,128,0.014170666535695394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,1024,0.031389332479900785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1024,2560,0.007070221834712558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,1024,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,6144,0.011029333704047732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,128,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,8192,0.010821333361996545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,8192,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,4096,0.007994666695594788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,2048,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,12288,0.023405333360036213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,7168,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,51200,0.26098667250739205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,12288,0.04086577892303467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,6144,0.0310791101720598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,5120,0.038908445172839694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,2560,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,3072,0.023364444573720295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,51200,0.4204844368828668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,768,0.049288001325395375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,2560,0.017084444562594097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,7168,0.037343111303117536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,16384,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,7168,0.01015022231472863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,7168,0.02801777919133504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,5120,0.008976889153321585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,256,0.06424444251590304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,512,0.038019554482565984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,256,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,4096,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,64,0.005038222091065513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,10240,0.22581511073642305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,3584,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,16384,0.1918168862660726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,3584,0.008380444513426887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,1536,0.0057608890864584185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,1536,0.08469600147671169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,16384,0.04527022110091316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,6144,0.06087733639611138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,3072,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,2048,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,51200,1.0736399756537545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,768,0.02474044428931342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,2560,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,10240,0.04215822286076016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,3072,0.041625777880350746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,2048,0.01975022256374359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,3072,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,8192,0.015429332852363586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,768,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,8192,0.08839555581410725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,51200,88.6895751953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,65536,0.15199022822909886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,4096,0.035191112094455294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,16384,0.04136355717976888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,2048,0.015264888604482016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,51200,7168,0.06445244285795423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,128,0.018234666850831773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,512,0.0052044445441828836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,512,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,16384,0.05490488807360331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,1536,0.009543110926946005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,128,0.016859556237856548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,6144,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,2048,0.024975111087163288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,16384,0.034557332595189415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,12288,0.21261066860622832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,768,0.007317333585686154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,16384,0.08669333325492011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,10240,0.04860622353023953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,1024,0.018443556295500863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,51200,0.050649778710471265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,5120,0.010775110787815519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,5120,0.009219555391205682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,768,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,128,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,51200,1.4594684176974828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,10240,0.03238133258289761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,2560,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,2560,0.011927111281288994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,3072,0.02366666661368476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,10240,0.15228177441491023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,4096,0.344343105951945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,12288,0.053173333406448364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,16384,0.13933689064449734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,256,0.1395911110772027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,3584,0.013998222019937305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,4096,0.01868355605337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,6144,0.03867822223239475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,4096,0.022447110878096685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,32,0.0035182221068276297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,6144,0.029345777299669053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,2560,0.006698666761318843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,5120,0.14706399705674914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,1024,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,1024,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,12288,0.09400266408920288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,12288,0.02404266595840454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,2048,0.01221777747074763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,256,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,51200,2.7031307220458984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,1536,0.017039111918873258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,1536,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,768,0.10138489140404595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,65536,0.13900710476769343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,3072,0.024308444725142583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,1536,0.034095999267366194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,2048,0.01789066692193349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,2048,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,32,0.0075733330514695905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,32,0.0035422220826148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,6144,0.017895999881956313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,3584,0.02278577784697215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,768,128,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,256,0.0024480000138282776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,1024,0.05535555548138089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,51200,3.2574301825629344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,5120,7168,0.045372443066702954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,256,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,3072,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,128,0.014866666661368476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,3072,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,5120,0.020607110526826646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,7168,0.15415378411610922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,1536,0.006487111250559489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,128,0.006665777828958299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3072,768,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,768,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,65536,51200,0.4399866792890761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,1536,0.013628444737858243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,65536,0.24563021130032012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,65536,1.2961484061347115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,12288,0.026885333988401625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3584,768,0.0029306666304667792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,3072,0.010704889065689512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,8192,0.04985333482424418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,65536,0.09462133381101821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,10240,0.08259022235870361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,8192,0.031742221779293485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,1024,0.0031582222630580268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,10240,0.19045155578189424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,7168,3072,0.0748586654663086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,8192,0.18103378348880342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,64,0.019116444720162284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,5120,0.022999111149046157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,768,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,1536,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,2560,0.029683556821611192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,12288,0.18587556150224474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,5120,0.014436443646748861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,4096,0.011578666667143503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,5120,0.021271111236678228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,3584,0.010302222437328762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,12288,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,512,0.018918222851223417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,128,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,32,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,16384,0.033231112692091204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,3072,0.00741244438621733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,16384,0.11350399918026394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,7168,0.12803288300832114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,5120,0.02159822152720557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,51200,0.12800799475775823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,65536,0.7217377556694878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,512,0.013745778136783175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,7168,0.013172444370057849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,16384,0.020209777686331008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,128,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,128,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,3072,0.02240088913175795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,128,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,3584,0.018844443890783522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,1024,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,2560,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,5120,0.026160889201694067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,10240,0.0232622226079305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,2560,0.027439999911520217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,4096,0.01627111103799608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,64,0.008967111508051554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,51200,0.6479759746127659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,7168,0.013636444177892474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,8192,0.29971112145317924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,8192,0.461259577009413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,2048,0.01143022212717268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,2560,0.15786843829684788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,65536,0.07562933365503947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3072,12288,0.26025154855516225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,768,0.03995022177696228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,4096,0.02388444377316369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,7168,0.019129777948061626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,32,0.03291911217901442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,3584,0.007684444387753804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,6144,0.02569777766863505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,512,0.07452799876530965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,64,0.016639999217457242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,768,0.2653004328409831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,12288,0.018027555611398485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,16384,0.08890844715966119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,64,0.001810666587617662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,12288,1.4837235344780817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,10240,0.08505777517954509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,768,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,256,0.13667466905381945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,8192,0.01479644411140018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,64,8192,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,51200,0.01315822203954061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,64,0.0042088888585567474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,128,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,4096,0.040037333965301514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,256,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,16384,0.8082079887390137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,32,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,12288,0.01926577753490872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,3584,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,16384,0.02836266822285122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,256,0.01900800069173177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,10240,0.1782311068640815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,51200,4.8616587320963545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,512,0.01847644481394026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,256,0.002313777804374695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,2560,0.01018666641579734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,1536,0.019369777705934312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,6144,0.07892533143361409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,2560,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,8192,0.07225244575076632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,3072,0.03177422285079956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,2560,0.031622221072514854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,16384,0.1440231137805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,6144,0.011525332927703857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,10240,0.05884889099333021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,12288,0.05388977792527941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,2048,0.01000355515215132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,7168,0.3351297908359104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,512,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,8192,0.05841155846913656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,512,0.014360888136757744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,512,0.06964177555508085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,16384,0.11797422832912868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,8192,0.019621334142155118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,64,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,32,0.0019031111150979996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,8192,0.012634666429625617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,768,0.013047110703256396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,512,0.015955555770132277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,3072,0.02721333338154687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,65536,128,0.00519466648499171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1536,51200,0.05033066537645128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,3072,0.021612443857722815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,8192,0.015574221809705099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,3584,0.19537156158023408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,256,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,4096,0.03440533412827386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,256,0.015225778023401896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,16384,0.02371911042266422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,12288,0.31541866726345485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,128,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,5120,0.027650665905740526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,12288,0.28191910849677193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,51200,0.05203288793563843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,3584,0.1548524432712131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,51200,0.916320906745063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,16384,0.17131377591027153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,4096,0.004530666602982415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,3584,0.07384800248675875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,768,0.01633777717749278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,1024,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,2560,0.022340445054901972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,512,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,6144,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,3072,0.007236444287829929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,5120,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,3584,0.024307555622524683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,3584,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,3584,0.08115733332104154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,6144,0.5079795519510905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,51200,0.7363822195265027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,51200,3072,0.19400621785057914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,16384,0.09906844298044841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,12288,0.08585333161883885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,10240,0.015598222613334656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,5120,0.020249777370029025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,4096,0.0151582227812873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,65536,0.47093868255615234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,32,0.001526222253839175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,51200,0.13874133427937826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,2048,0.020063999626371596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,512,0.008079999850855933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,2048,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,1024,0.01518311103185018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,3584,0.05481333202785916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,64,0.002797333316670524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,51200,256,0.020273778173658583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,256,0.014837332897716098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,32,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,1024,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,256,0.026847999956872728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,65536,3072,0.03287555442916022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,5120,0.02642400066057841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,2560,0.01149511088927587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,6144,0.010110222631030613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,2560,0.012105777859687805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,65536,0.6087279849582249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,4096,1.5043715371025934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,3072,0.020465777979956735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,64,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,8192,0.014069333672523499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,7168,0.011036444041464063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,51200,0.045773333973354764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,10240,0.0414319998688168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,128,10240,0.013457777599493662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,51200,1.3735937542385523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,256,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,3584,0.006568888823191325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,1024,0.030518223841985066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,512,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,8192,0.01128799964984258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,6144,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,6144,0.01721333298418257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,1024,0.010086221827401055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,7168,0.010000000397364298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,12288,0.219596438937717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,51200,2.8718346489800344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,16384,0.04176622298028734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,6144,0.05286844571431478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,1024,0.026758222116364375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,2560,0.010696888797812991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,10240,0.034282667769326106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,16384,0.03697333402103848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,8192,0.031531555785073176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,16384,0.025940444734361436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,3072,0.013438222308953604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,256,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,8192,0.497486220465766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,4096,0.18477333916558158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,12288,0.016791999340057373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,1024,0.012489777472284106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,2560,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,3072,0.04092533389727274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,768,0.028157333532969158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,2560,0.006948444578382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,2048,0.00666844430896971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,16384,0.08941155672073364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,1536,0.04774666825930277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,10240,0.03566311134232415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,51200,0.11469066143035889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,256,0.005381333331267039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,256,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,64,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,51200,0.09573155641555786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,128,0.013411555853154926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,128,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,2560,0.014117333624098035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,1536,0.05899644560284085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,4096,0.02737422287464142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,128,2560,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,64,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,8192,0.024608888559871252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,3072,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,5120,0.00849333322710461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,5120,0.023617777559492324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,768,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,65536,0.3217208915286594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,128,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,6144,0.0554622213045756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,128,64,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,7168,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,3072,0.5956640243530273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,12288,0.01795644395881229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,5120,0.014621333943472968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,768,0.033575110965304904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,768,0.028363555669784546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,512,0.006983111302057902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,2048,0.02604177759753333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,12288,0.0391591125064426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,2560,0.27312710550096303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,65536,0.06464266777038574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,2560,0.033017777734332614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,10240,0.012696888711717395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,128,0.05746755335066053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,32,0.0021066665649414062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,3072,0.07843466599782307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,4096,0.023153778579499986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,65536,2.5301706526014542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,7168,0.019273777802785236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,1536,0.008345777789751688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,768,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,5120,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,12288,0.004910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,2048,0.021926222576035395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,10240,0.3519582218594021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,768,0.01960177719593048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,8192,0.015856888559129503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,10240,0.1328399976094564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,8192,0.6726533571879069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,32,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,5120,0.009338666995366415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,1536,0.03346577617857192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,512,0.009358222285906473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1024,8192,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,51200,0.040417777167426214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,6144,0.01221422188811832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,5120,0.04227644536230299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,128,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,64,0.034527109728919134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,1536,0.018935110833909776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,32,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,3584,0.007575111256705389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,3584,0.023754666248957317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,65536,0.09211911095513238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,2560,0.052814222044414945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,5120,2048,0.05981155236562093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,4096,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,65536,0.0548257761531406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,8192,0.020163555939992268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,2048,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,768,0.013262222210566202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,5120,0.042450666427612305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,3072,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1536,512,0.02069155540731218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,2048,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3584,65536,0.07097599903742473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,64,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,1536,0.3450924555460612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,4096,0.014759999182489185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,3584,0.030012445317374334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,768,0.031166222360399034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,256,0.014537778165605334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,65536,0.38857422934638125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,8192,0.04747733473777771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,2560,0.02943555514017741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,1024,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,3584,0.021990221407678392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,3072,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,4096,0.021017778250906203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,51200,0.26585600111219615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,128,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,1536,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,5120,0.05959910816616482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3584,1536,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,64,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,1536,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,7168,0.025528000460730657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,3072,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,65536,5120,0.1217271089553833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,8192,0.01779822177357144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,3584,0.007123555574152205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,32,0.014927999840842353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,12288,0.031939556201299034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,7168,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,16384,0.04326755470699734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,5120,0.6598089006212022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,512,256,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,5120,0.018887110882335238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,3584,0.015802666544914246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,2048,0.053780443138546415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,256,0.01497244503762987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3584,5120,0.0251226673523585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,16384,0.037181334363089666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,2560,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,51200,2.0774194929334855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,7168,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,2048,0.01960444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,768,0.0221368885702557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,256,0.027898665931489732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,1536,0.05110488997565376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,3584,0.021133333444595337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,1536,0.059446222252315946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,7168,0.015390222271283468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,3072,0.02428177826934391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,65536,0.18417599466111925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,768,0.012409778104888068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,7168,1.097751087612576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,512,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,16384,0.026120889517996047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,3072,0.017271111408869427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,2048,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,10240,0.04702311091952854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3584,51200,0.05389155613051521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,1536,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,1024,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,5120,0.0064195555945237475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,5120,0.02678577767478095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,12288,0.03618399964438544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,768,0.015511110424995422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,256,0.012421333127551608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,2048,0.0070159998204973005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,1536,0.019876443677478366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,32,0.007772444850868649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,7168,0.09925244251887004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,65536,0.06023910972807142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,64,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,12288,0.04509511258867052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,6144,0.09828622473610772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,8192,16384,0.4917759895324707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,6144,10240,0.20938311682807076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,51200,0.040785776244269475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,1536,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,12288,0.12130488289727105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,768,0.29033509890238446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,3072,0.016555555992656283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,16384,0.12025066216786702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,51200,0.05059822069274055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,12288,0.1938871145248413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,51200,0.09760622183481853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,7168,0.10932799842622544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,256,0.0058453331391016645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,1536,0.008519111408127679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,768,0.02513155506716834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,6144,0.03044711218939887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,1024,0.04610400067435371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,16384,0.4871066411336263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,256,0.026008889079093933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,8192,0.022640888889630634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,6144,0.02011999984582265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,65536,0.0938897795147366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1024,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,12288,0.04661866691377428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,128,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,8192,0.030954665607876245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,5120,0.02963199880388048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,1536,0.011979555918110741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,10240,0.6405937936570909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,65536,0.06294577651553683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,6144,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,768,0.02254044512907664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,8192,0.07096266746520996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,256,0.0041759999261962045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,128,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,51200,0.0476399991247389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,3072,0.14902755949232313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,8192,0.009526222116417354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,8192,0.026874666412671406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,64,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,3584,0.036850667662090726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,4096,2560,0.015786666009161208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,12288,0.024886222349272832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,5120,0.017320000463061862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,768,0.003181333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,256,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,2048,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,12288,0.03521600034501817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,12288,0.016679111454221938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,2560,0.020625778370433383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,16384,0.013120888835854001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,3072,0.008181333541870117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,768,0.04084800018204583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,128,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,2560,0.004869333157936732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,1024,0.01316177762216992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,1024,0.014130666851997375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,10240,0.015435554915004306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,768,0.002814222127199173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,7168,0.0647653341293335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,65536,0.028074665202034846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,8192,256,0.021662222014533147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,4096,1.5327271355523004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,65536,0.6892666816711426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,4096,0.021871111459202234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,1024,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,8192,0.036534223291609026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,6144,0.009431111315886179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,4096,0.027814222706688776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,3072,0.8573306931389703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,64,0.017459554804695975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,4096,0.07680000199211968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,16384,0.44778045018513996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,10240,0.19233955277336967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,256,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,512,0.1852008899052938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,12288,0.02186222208870782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1024,6144,0.030485332012176514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,128,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,64,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,512,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,768,0.0033946666452619764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,1536,0.07755021916495429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,512,0.021520889467663232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,3072,0.0406640006436242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,16384,1.2255137761433919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,1024,0.0123831108212471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,6144,0.01681600014368693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1536,51200,0.027535999814669292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,4096,0.008114666574531132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,1024,0.019865777757432725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,2048,0.02055199940999349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,1536,0.057780444622039795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,4096,0.053364445765813194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,16384,0.033253331979115806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,64,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,128,0.02682311170630985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,2048,0.010354667074150508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,16384,0.04289866818322075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,65536,0.06686222553253174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,16384,2.8327564663357205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,16384,0.12633332941267225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2048,5120,0.05628355344136556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,6144,0.013781333135233985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,1024,0.014577777849303352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3584,2560,0.01796800063716041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,6144,0.020370667179425556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,51200,0.5500657823350694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,51200,0.0675777792930603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,1024,0.021341333786646526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,12288,256,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,512,0.00871288859181934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,5120,0.021887999441888597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,51200,0.0471093323495653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,6144,0.04682755470275879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,3584,0.016004444824324716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,16384,0.05840622054206001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,8192,0.010626667075686984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,2048,0.010072888599501716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,32,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,7168,0.02016444504261017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,2560,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,256,0.006685333533419504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,16384,0.03417955504523383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,6144,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,4096,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,6144,0.11407377984788682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,64,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,7168,0.3623653252919515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,5120,0.017283555534150865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,256,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,65536,0.07299911313586764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,3072,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,51200,0.027811557054519653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,32,0.006637333167923822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,3584,0.012332444389661154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,12288,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,512,0.0028604444944196274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,51200,12288,4.861921098497179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2560,51200,0.23709511756896973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,512,0.007876444194051955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,128,0.01703733371363746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,51200,0.1174337731467353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,4096,0.011748444702890186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,2560,0.02273688879277971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,6144,0.02113688819938236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,12288,0.03458222084575229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,6144,0.018564444449212816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,10240,0.12436710463629828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,16384,0.016117332710160148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,3584,0.09403733412424724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,768,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,1536,0.01757244434621599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,1536,0.015112888481881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,6144,0.10428888930214776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,768,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,1024,0.012715555727481842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,4096,0.024884444144037034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,3072,0.021951110826598272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,5120,0.004518222063779831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,4096,0.4527626567416721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,12288,0.11696799596150716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,6144,0.025848888688617285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,3584,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,768,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,7168,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,51200,1.1949164072672527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,4096,0.06879378027386136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,16384,0.039792001247406006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,768,7168,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,51200,7168,0.29242666562398273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,7168,0.09986844327714707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,128,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,7168,0.07634933127297296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,65536,0.058733330832587353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,65536,32,0.012448000411192576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,7168,0.03228711088498434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,7168,0.05374222331576877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,51200,2.2207404242621527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,2048,0.008466666771305932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,7168,0.1878133349948459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,1536,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,5120,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,128,512,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,128,0.003144888828198115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,3584,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,8192,0.029870222012201946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,3584,0.13262755341000027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,256,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,512,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,65536,10240,1.2782408396402996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,7168,0.12574311097462973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,3072,0.013346666263209449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,32,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,3072,0.12467732694413926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,65536,3072,0.6337386767069498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,128,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,2048,0.06775466601053874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,1024,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,8192,0.05198577708668179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,10240,32,0.0022044444663657085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,4096,0.056940442985958524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,2048,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,16384,0.04153955645031399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,64,0.0018773333479960759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,6144,3072,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,256,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,5120,0.9009102715386285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,4096,0.008382221890820397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,1024,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,3072,0.06992622216542561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,65536,1.057847128974067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,5120,0.09306489096747504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,8192,0.024999111890792847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,768,0.009377777576446533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,3072,0.02737688852681054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,51200,1.0612026850382488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,768,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,2560,0.041305777099397444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,1024,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,1024,0.003531555748648114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,3584,0.010828444527255164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,2048,0.004874666531880696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,2048,0.032073779238594904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,6144,1.8298640780978734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,6144,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,16384,65536,0.24052267604404023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,3584,0.02371999952528212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,256,0.013432888521088494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,12288,0.0867991116311815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,256,0.0069546666410234236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,128,0.015508444772826301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,7168,512,0.016846223009957206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,4096,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,51200,0.07086577680375841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,64,0.002161777681774563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,3584,0.007627555893527136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,51200,0.36516088909573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,65536,1.1518239974975586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,2048,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,8192,0.02367199957370758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,256,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,768,0.024328889118300542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,3584,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,768,0.021350221501456365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,5120,0.035895999934938215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,2560,0.02146666745344798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,51200,0.8465075492858887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,128,0.017557332913080852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,256,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,1536,0.018254222141371835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,16384,0.045660446087519325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,768,0.01792266633775499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,1536,0.004546666724814309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,8192,0.016036444240146212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,4096,0.009402666654851701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,256,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,5120,0.017892445127169292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,64,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,51200,5.745215098063151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,1536,0.01792888840039571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,51200,5120,0.08881866931915283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,5120,0.1771786610285441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,512,0.01309511148267322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,768,0.02077155477470822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,1536,0.02694844371742672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,512,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,256,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,12288,0.05258311165703667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,3584,0.09362844626108806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,1024,0.016008888681729633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,3072,0.027323555615213182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,32,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,5120,0.04032622112168206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,7168,0.05929333633846707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,10240,0.029955555995305378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,10240,0.015758223003811307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,2560,0.013864888913101621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,3072,0.8042382134331597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,4096,0.024375999967257183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,32,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,5120,0.28213156594170463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,2048,0.009974222216341231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,512,0.003583999971548716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,32,0.011525332927703857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,128,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,8192,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,12288,0.0195795562532213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,32,0.03287555442916022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,512,0.02230577833122677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,4096,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,51200,0.06029778056674533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,5120,0.0064488889442549805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,7168,0.06141422192255656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,10240,0.006312888943486744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,1024,0.020679111282030743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,2560,0.1832817792892456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,2560,0.030743112166722614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,5120,0.004724444614516364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,32,6144,0.023660444551044043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3584,3072,0.024681778417693243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,16384,0.09606755442089504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,65536,2.233934190538194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,2560,0.005575110928879843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,1024,0.003859555555714501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,1536,0.00925422211488088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,768,0.02058577703105079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,64,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,8192,0.3080764346652561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,16384,0.03058755397796631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,1536,0.0063075555695427795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,1536,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,128,0.016895999511082966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,1024,0.03849511014090644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,65536,1.0828417672051323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2048,12288,0.1620186699761285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,51200,0.06736266613006592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,3072,0.12203910615709092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,12288,0.019976889093716938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,2560,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,7168,0.07829688655005561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,32,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,768,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,3584,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,2048,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,768,0.016717332932684157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,12288,0.4417866600884332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,64,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,1536,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,16384,0.060124443637000195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,1024,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,51200,0.09889155626296997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,51200,2.10316530863444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,10240,0.014811555544535318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,768,0.009696000152164036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,51200,0.03831555445988973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,8192,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,8192,0.007728888756699032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,51200,512,0.24252978960673013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,65536,0.03785689009560479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,12288,0.013736888766288757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,128,0.010802666346232096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,2560,0.0070497774415545994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,32,0.008015111088752747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,3584,0.015798222687509324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2560,1536,0.012854222622182635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,10240,0.9446036020914713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,256,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1536,3072,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,1024,0.029507554239696924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,256,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,12288,0.2865368790096707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,51200,0.8728702333238391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,2048,0.004514666481150521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,1024,0.025177778469191656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,32,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,32,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,3584,0.0088755554623074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,3584,0.011534222298198275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,12288,0.014084445105658637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,3584,0.024889777104059856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,5120,0.011992000043392181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,6144,1.9665387471516926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,65536,5.248482598198785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,512,0.014939554863505893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,10240,0.12093155913882786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,256,0.002475555572244856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,4096,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,128,0.005886222339338726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,512,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,12288,12288,0.16026043891906738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,2048,0.010264889233642155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,256,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,1024,0.014212444424629211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,768,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,10240,0.053025778796937734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,7168,0.24599822362263998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,51200,0.05440800057517158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,1024,0.029417776399188574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,1024,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,128,0.016020443704393175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,10240,10240,0.19611644744873047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,2560,0.029600001043743555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,1536,0.005135111096832487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,12288,0.04090399874581231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,5120,0.009139555195967356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,16384,0.006081777728266186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,65536,0.10542844401465522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,64,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,51200,7.215265062120225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,1536,0.024664888779322307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,51200,0.22303289837307402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,6144,0.00941244430012173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,16384,0.18860977225833467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,16384,0.0391813317934672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,256,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,3072,0.003882666842805015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,768,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,3072,0.016725333200560678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,12288,0.6953644222683377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,1536,0.05720622009701199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2560,7168,0.12669421566857234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,8192,0.022642667094866436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,6144,0.01311466677321328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,8192,0.25493155585394967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,5120,0.11763021681043838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,8192,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,256,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,2560,0.005224888937340842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,3584,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,3584,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,3584,0.03273155623012119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,2048,0.0372924440436893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,1024,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,12288,0.036350223753187395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,51200,0.044350223408805005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,65536,4.160331726074219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,51200,0.23053420914543998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,1536,0.04409689042303297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,6144,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,51200,0.015807999504937064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,5120,0.1194408867094252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,1536,0.02334755493534936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,5120,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,128,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,512,0.0027991111079851785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,3584,0.07563910881678264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,3072,0.08780088689592148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,8192,7168,0.054437332683139376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,7168,0.015909334023793537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,3584,0.06448888778686523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,5120,0.12015733453962539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,64,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,128,8192,0.06426133049858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,10240,0.21547378434075248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,16384,0.08389421966340806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,768,0.01942755613062117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,8192,0.01594311164485084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,8192,12288,0.6769982443915473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,10240,2560,0.032167110178205706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,51200,0.09596177604463364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,12288,0.24336889055040148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,3584,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,12288,0.02072177827358246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,12288,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,4096,0.8043413162231445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,128,0.006400888992680444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,4096,0.02332088847955068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,7168,0.12665155198838976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,1024,0.008438222110271454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,7168,0.07896977663040161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,3584,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,1024,0.003141333245568805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,32,0.0041804446114434134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,8192,12288,0.021014221840434607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,2048,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,3584,0.023783110909991797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,10240,0.444992012447781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,768,0.018255111243989732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,256,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,2048,0.06008177995681763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,12288,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,3072,0.008607111043400234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,6144,0.05717955695258247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,64,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,4096,0.4945315784878201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,256,0.008340444829728868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,3584,0.09948622518115574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,65536,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,51200,2.398355484008789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,3072,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,12288,0.019911999503771465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,51200,0.18330578009287515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,16384,0.04186488853560554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,7168,0.0233297778500451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,65536,1.3109813266330295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,128,0.04505599869622124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,3584,0.023804444405767653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,12288,0.31596533457438153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,3072,0.024840000602934096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,3072,0.02336888843112522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,512,0.10052266385820176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,7168,0.028171555863486394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,8192,0.3436720106336806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,7168,0.012773333324326409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,4096,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,65536,4.265168084038629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,7168,0.021440888444582622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,10240,0.16655822594960532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,7168,0.12790044148763022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,2560,0.022337777747048274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,51200,7168,0.06367199950748019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,64,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,512,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,5120,0.008799110849698385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,51200,5120,0.185080885887146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,2560,0.013792888985739814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,2560,0.030223111311594646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,65536,0.3692026668124729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,10240,0.04917422268125746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,6144,0.03191377719243368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,16384,0.027756444282001917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,768,0.0038888889054457345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,8192,0.03146933184729682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,128,0.015140444040298462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,1536,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,128,0.006050666587220297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,1024,0.012135999898115793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,1536,0.03087555699878269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,10240,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,3584,0.5295519828796387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,7168,0.012522666818565793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,64,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,4096,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,1536,0.006364444477690591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,1024,0.044235554006364614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,10240,3584,0.06948533323076037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,5120,0.026406221919589575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,5120,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,6144,0.019761777586407132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,64,0.0022533333135975734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,8192,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,51200,0.07528089152442084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,7168,0.03373510970009698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,65536,0.06558310985565186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,2560,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,512,0.005167110926575131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,8192,0.19132444593641493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,1536,0.027069333526823256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,512,0.015445333388116626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,1024,0.019187556372748483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,32,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,768,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,16384,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,5120,0.009022222624884712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,10240,0.0269804447889328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,65536,0.08493600289026897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,1536,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,32,0.0018746666610240936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,10240,0.015206222732861837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,2560,0.007459555235173967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,12288,0.046000000503328115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,4096,0.03434755404790243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,2560,0.008610666626029545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,10240,51200,0.45789511998494464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,768,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,65536,3072,0.25448534223768443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,768,0.022335110439194575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,3072,0.07594311237335205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,4096,0.016516443755891588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,1536,0.0041537777417235905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,12288,0.030638221237394545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,32,0.07847288582060072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,2048,0.028873778051800195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,10240,0.09683110978868271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,7168,0.017839110559887357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,512,64,0.001834666563404931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,51200,0.07685866620805529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,12288,0.056165330939822726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,128,0.010339555641015371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,2560,0.013109332985348172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,768,0.004224888980388641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,51200,0.35603290134006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,512,0.00250133333934678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,256,2048,0.013403555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,256,0.021294222937689886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,3072,0.00739911115831799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,128,0.015178667174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,2048,0.019688889384269714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,5120,0.038663112454944186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,2560,0.004923555586073133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,8192,0.024673778149816725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,4096,2.3630292680528426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,10240,0.018228444788191054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,4096,0.2347102165222168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,3072,0.023870221442646448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,10240,0.012068444656001197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,2560,0.009326222042242685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,512,0.0299893319606781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,7168,0.005336888962321811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,1536,0.015098666151364645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,1536,0.0969440009858873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,3072,0.014871110518773397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,65536,0.42146844334072536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,1024,0.02330844435426924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,768,0.004837333328194088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,768,0.015501333607567681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,2560,0.017494221528371174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,768,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,768,0.020087111327383254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,1536,0.005521777603361342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,12288,0.022732444935374792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,2048,0.037463999456829496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,51200,0.22265156110127768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,4096,0.12851022349463567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,128,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,3584,0.03368533319897122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,4096,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,5120,0.02490933405028449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,6144,0.012120889292822944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,6144,0.08253422048356798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,256,0.02196177840232849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,2048,0.04313866628540886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,6144,0.10768533415264553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,8192,0.019641778535313077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,5120,0.03476266728507148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,2048,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,3072,0.007685333490371704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,1024,0.00871022211180793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,3584,0.067885332637363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,51200,0.312678231133355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,6144,0.07283555799060397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,51200,0.1350542174445258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,128,0.9896106719970703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,1024,0.007409777906205919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,4096,0.038726223839653864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,8192,0.17544178167978922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,51200,0.25115378697713214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,3584,0.057208889060550265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,3072,0.10798311233520508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,8192,0.02630311085118188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,256,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,16384,0.34979110293918186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,7168,0.029838220940695867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,12288,0.014121777481502958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,5120,0.03852799865934584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,8192,0.022320000661744013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,12288,0.03570400012864007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,1024,0.015216888652907478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,16384,3584,0.12458577420976426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,12288,0.18570400608910453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,1536,0.01071733319097095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,512,1536,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,7168,0.013266666895813413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,3584,0.022336888644430373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,6144,0.250237332450019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,7168,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,51200,1.0747298134697807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,51200,16.29091559516059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,1536,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,64,0.009343999955389235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,51200,5.339169820149739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,2048,0.09475911325878567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,3072,0.04536799920929802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,10240,0.21340266863505045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,512,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,6144,0.020311110549502902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,1536,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,32,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,5120,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,3584,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,16384,0.7100106875101725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,8192,0.09828711218304104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,128,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,65536,0.08978222476111518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,1536,0.010962666736708747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,3584,0.062183108594682485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,3584,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,1536,0.026250667042202417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,7168,0.032666666640175715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,1536,0.003564444267087512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,2560,0.020847999387317236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,7168,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,7168,0.3649866580963135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,1024,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,1024,0.054903109868367515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,51200,1.4560444090101454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,64,0.005555555638339784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,16384,2560,0.06740978029039171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,10240,0.04454133245680067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,5120,0.009135111338562435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,1536,0.10849955346849229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,512,0.012491555677519904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,768,0.00563466673096021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,65536,0.05617688761817085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,6144,0.16017866134643555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,65536,0.3278791109720866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,4096,0.063028441535102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,2048,0.028051555156707764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,6144,0.008763555851247575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,1536,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,8192,0.04936977889802721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,1024,0.04022488991419474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,51200,0.10617422395282322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,5120,0.028128001425001357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,32,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,51200,0.5059839884440104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,128,0.0024417777442269856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,4096,0.09159821934170193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,3584,0.031297776434156634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,51200,2560,0.027895109521018133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,12288,0.11798932817247178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,8192,0.12730577256944445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,2048,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,4096,0.054175999429490834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,256,0.0439573327700297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,7168,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,8192,0.3557751178741455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,3584,0.022808000445365906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,3072,0.05346488952636719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,7168,0.022670222653283015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,128,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,1024,0.21407822767893472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,4096,0.010960000256697336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,2048,0.023038221730126276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,5120,0.1139306624730428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,1536,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,256,0.006724444528420766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,2048,0.024547555380397375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,128,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,2560,0.004228444563017951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,16384,0.02163822286658817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,1024,0.009526222116417354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,51200,0.08996178044213189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,8192,0.059338668982187905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,2048,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,512,0.004168888760937585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,2048,0.07173689206441243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,6144,0.09116977453231812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,128,0.021274665991465252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,5120,16384,0.10340088605880737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,10240,8.276502821180555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,512,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,8192,0.05406044589148628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,51200,0.46520445081922746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,768,0.00286311118139161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,65536,0.11773778332604302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,128,0.012437333663304647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,8192,0.2694728904300266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,7168,0.05259199937184652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,65536,0.12117600440979004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,5120,0.010011555420027839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,7168,0.009379555781682333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,128,2048,0.013094222380055321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,51200,0.06825866964128283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,10240,2048,0.009667555491129557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,10240,0.01680799987581041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,2048,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,16384,2560,0.09584355354309082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,8192,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,65536,3.223728815714518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,768,0.02809244394302368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,1024,0.004950222041871813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,2560,0.025099555651346844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,16384,0.06552888949712117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,12288,0.04420800010363261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,512,0.0038453332251972626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,7168,0.012367111113336353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,4096,0.03216622273127238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,512,0.02603200078010559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,2560,0.015469332536061605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,51200,0.2068319982952542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,4096,1.0952435599433052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,512,0.06313155757056342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,3584,0.4683689011467828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,256,0.0026319999661710528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,4096,0.024953777591387432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,1024,0.02389777700106303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,1536,0.013547555440002017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,10240,0.3074177900950114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,65536,9.965750800238714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,10240,0.02752711044417487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,5120,0.23599910736083984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,2560,0.07678488890329997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,7168,0.4751946661207411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,10240,0.029482665989134047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,768,256,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,10240,0.05939110782411364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,3072,0.0369395547442966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2048,768,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,16384,0.021205332544114854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,768,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,512,0.006224000205596288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,3584,0.09001688824759589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,51200,0.40026132265726727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,2048,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3584,768,0.004302222281694412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,65536,0.4160444471571181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,10240,0.030533333619435627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,65536,0.09045777718226115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,6144,0.08049333095550537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,10240,0.015936889582210116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,1536,0.017966222431924608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,256,0.006257777826653586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,2560,0.032418668270111084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,4096,0.008328888979223039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,4096,0.048047112094031445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,5120,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,7168,0.021850667066044275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,3072,0.03489511211713155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,4096,0.030441777573691473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,10240,12288,0.02981333269013299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,10240,1.8848550584581163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,51200,0.05357244279649523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,4096,0.03932888971434699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,65536,0.3600871033138699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,512,0.08081422249476115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,1024,0.00998044427898195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,10240,0.21307199531131318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,4096,0.03255911005867852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,256,0.009872888525327047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,2560,0.00741511086622874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,51200,0.04307466745376587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,65536,0.20702489217122397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,10240,0.025654221574465435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,65536,3584,0.34401244587368435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,16384,26.715408325195312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,10240,2.942285325792101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,3584,0.04222311245070564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,128,0.02123911182085673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,6144,0.10587200191285874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,3072,0.016415110892719693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,512,0.014502222339312235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,10240,0.03394133183691237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,1536,0.0069857777820693115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,10240,0.07040088706546359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,2048,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,10240,0.022682666778564453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,256,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,128,0.029873778422673542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,1024,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,4096,0.03003644280963474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3072,5120,0.008630221916569604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,8192,0.05776088767581516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,1536,0.009619555539555019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,128,0.016926222377353244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,3072,0.01956977778010898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3584,10240,0.044996443721983165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,128,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,3072,0.02164977788925171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,12288,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,10240,0.015849777393870883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,1024,0.013131555583741931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,4096,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,16384,0.06162133481767443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,7168,0.1514079968134562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,5120,0.046535998582839966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,7168,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,2048,0.27229155434502494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,256,0.005632888939645555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,7168,3584,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,3584,0.017222222354676988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,7168,0.04847288793987698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,65536,0.06593422094980876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,512,0.048878222703933716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,2048,0.0070168889231152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,512,0.006690666493442323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,3072,0.014315555493036905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,128,0.022942221826977197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,64,0.014932443698247274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,3584,0.06718933582305908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,64,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,2560,0.007195555501514011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,512,0.012070222033394707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,64,0.0022355556074115965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,3072,0.06414755847718981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,65536,0.552328003777398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,16384,0.02368533280160692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,16384,0.10569155878490871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,7168,0.010772444307804108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,65536,0.1519884400897556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,12288,0.23418755001491973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,12288,8192,0.02338933282428318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,4096,0.01661066710948944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,3584,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,6144,0.017809778451919556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,6144,0.014273777604103088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,8192,0.032440000110202365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,1536,0.4081164466010199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,8192,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2048,5120,0.009736888938479954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,12288,0.23005334536234537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,65536,0.17912622292836508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,8192,0.01071999967098236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,1024,0.0038808890514903595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,128,0.0062977779242727495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,16384,0.0188746667570538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,8192,0.037816001309288874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,51200,0.15441244178348118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,512,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,3584,0.030772444274690416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,512,0.020693333612547982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,512,0.012687111066447364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,2560,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,5120,0.16685333516862658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,3584,0.02233155568440755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,10240,32,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,256,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,12288,0.13248266114128962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,51200,0.24969866540696886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,512,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,128,0.002464888824356927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,8192,1.2745946248372395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,64,0.01595288846227858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,4096,0.029520889123280842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,2560,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,512,0.014150222142537435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,8192,0.07480800151824951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,10240,0.2894453207651774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,3584,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,7168,0.012176888684431711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,256,0.015096889601813423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,51200,0.051979555024041064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,128,512,0.007221333682537079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,16384,0.01868266695075565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,51200,0.5050364600287544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,12288,0.023794665932655334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,2048,0.021017778250906203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,65536,0.09190311034520467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,7168,0.033879998657438494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,2560,0.013917333549923368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,12288,0.26269955105251735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,768,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,128,0.0021502222451898786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,1536,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,256,0.004879111217127906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,2048,0.004591111093759537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,64,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,5120,0.029994666576385498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,7168,0.029412445094850328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,128,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,512,0.01723022262255351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,2560,0.021033777130974665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,1024,0.017107556263605755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,3584,0.004324444466167026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,1536,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,65536,0.0676693320274353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,8192,0.01899377836121453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,64,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,1024,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,6144,0.03065511253145006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,8192,0.016349333855840895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,51200,0.047055999437967934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,2560,0.02751022246148851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,64,0.00452622233165635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,512,2048,0.009520889156394536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,256,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,2048,0.006558222075303395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,51200,1.450974252488878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,256,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,10240,0.1737768914964464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,8192,0.06266755527920194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,1536,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,10240,0.021941334009170532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,51200,2048,0.053504887554380626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,8192,0.24535910288492838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,512,0.01253600004646513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,2560,0.06264888577991061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,16384,0.030512889226277668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,32,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,12288,0.03696444299485948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,6144,7.4448962741427955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,5120,0.054142223464118115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,2048,0.03019022279315525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,8192,0.15519200430976018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,16384,0.044953776730431445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,5120,0.07575200001398723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,5120,0.29593245188395184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,5120,0.0214720004134708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,256,1.3782435523139105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,5120,0.009389333426952362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,512,0.003212444484233856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,6144,0.4794035487704807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,12288,768,0.05330311258633932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,512,0.00905688852071762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,2048,0.01812622282240126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,1024,0.0031973332580592898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,12288,1.4503066804673936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,2048,0.01961599952644772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,768,0.023137778043746948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,768,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,6144,0.1289164490169949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,12288,0.03562488820817735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,16384,0.18390044901106092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,16384,0.020289777053727042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,12288,0.4253848923577203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,8192,0.15907466411590576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,768,10240,0.03418133325046963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,10240,0.29393688837687176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,8192,2560,0.032038221756617226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,3072,0.050985776715808444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,3072,0.027092445227834914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,128,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,8192,0.02166844407717387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,3072,0.014896000425020853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,64,0.006324444380071428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,7168,2560,0.27551200654771596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,8192,0.03725511166784499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,12288,0.03531288769510057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,128,0.012210666305489011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,6144,0.045533332559797496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,128,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,10240,0.04051733348104689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,3584,0.04001422060860528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,51200,5.465836418999566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,65536,0.2824817763434516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,64,0.19597955544789633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,51200,0.1276924477683173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1024,5120,0.018554665976100497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,3072,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,64,0.00590133335855272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,16384,4.017255147298177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,5120,0.11679822868771023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,4096,1.0814879735310872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,768,0.04593955477078756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,512,0.008783111141787635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,256,0.005547555370463266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,256,0.03209866748915778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,6144,0.2387119928995768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,16384,0.08903911378648545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,5120,0.0388702220386929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,5120,0.013330666555298699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,3584,0.022827555735905964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,2048,0.023750222391552393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,1536,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,32,0.002465777720014254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,768,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,2560,0.008522666990756989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,8192,0.07880800300174289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,3584,0.032299555010265775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,256,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,256,0.0021084443562560612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,256,1024,0.0034826666944556763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,7168,0.06013511286841499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,256,2048,0.025803556044896443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,3584,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,64,0.0024622221373849446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,65536,1.2626043955485027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,32,0.020498666498396132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,8192,0.039213332864973284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,1024,0.00907199995385276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,6144,0.012506666282812754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,6144,0.03219377663400438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,3072,0.00996088898844189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,512,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,3584,0.028488000233968098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,128,0.0052888890107472735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,768,0.023334221707450017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,7168,0.011773332953453064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,768,1024,0.013463999662134381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,1024,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,2560,0.2638000117407905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,3584,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,512,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,128,0.014119999276267158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1024,8192,0.02649333410792881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,1536,0.023642667465739783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,65536,0.05838221973843045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,4096,0.05897688865661621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,51200,0.23570667372809517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,8192,0.05146400133768717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,128,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,8192,0.07671733034981622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,65536,0.41240443123711484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,65536,512,0.1309760014216105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,32,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,1024,0.014087999860445658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,8192,0.31539999114142525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,16384,0.06867200136184692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,768,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,6144,0.015776000089115567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,3072,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,1536,0.010017777482668558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,256,0.0021760000122918025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,512,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,512,0.01626311077011956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,7168,0.004877333425813251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,768,0.005963555640644497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,65536,0.05560977591408623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,32,0.02438844409253862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,65536,0.20536088943481445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,256,0.015331554743978711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,51200,2.8907750447591147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,512,0.013453333742088742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,12288,0.06458489100138347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,51200,0.061546663443247475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,6144,4.841017829047309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,16384,0.12488888369666205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,768,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,16384,0.025792888469166223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,5120,0.0418906675444709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,128,0.006647999915811751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,3072,0.02608355548646715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,32,0.006968888971540663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,2560,0.016759999924235873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,6144,0.11471022499932183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,256,0.018240888913472492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,6144,0.29135910669962567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,1536,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,2048,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,5120,0.024266666836208765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,2560,0.07492533657285902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,7168,51200,0.3675804403093126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,512,0.157151116265191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,1024,0.05961600277158949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,1024,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,1536,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,3072,0.030799110730489094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,6144,0.018181333939234417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,8192,10240,0.07732444339328341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,12288,0.02051377793153127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,10240,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,2560,0.010151999692122141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,65536,1.0304479598999023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,512,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,7168,0.18458578321668836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,768,0.018600889378123812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,51200,0.05019911130269369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,5120,0.03009866674741109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,2560,0.009099555512269339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,6144,0.015450666348139444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,256,51200,0.0510079993142022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,2048,0.35820534494188094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,7168,0.07345777750015259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,5120,0.011865778101815118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,3072,0.5941013230217828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,16384,0.036355555057525635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,16384,3.1646160549587674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,64,0.006262222097979651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,512,0.010649777948856354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,32,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,6144,0.05730399820539686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,2048,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,1536,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,256,0.019680889116393197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,512,0.09159466955396865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,7168,0.2625902228885227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,128,0.00434755575325754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,2560,0.020784889658292133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,2560,0.02686133318477207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,2048,0.00980000032318963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,256,0.01578044394652049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,7168,0.06886400116814508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,512,0.003128888913326793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,768,0.0124195557501581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,12288,0.15196443928612605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,256,0.03722755445374383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,65536,0.515334235297309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,8192,0.10502488745583428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,8192,0.3400248951382107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,3072,0.03498311175240411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,128,0.01538755496342977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,256,2560,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,3072,0.00741511086622874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,8192,0.13632978333367243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,2560,0.0900497767660353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,8192,0.10675022337171768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,5120,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,3584,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,4096,1536,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,12288,0.09209599759843613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,768,0.04373244444529215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,128,0.006798222247097228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,1536,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,64,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,2560,0.011495999991893768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,7168,0.0740977790620592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,3584,0.009384888741705153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,6144,0.023839111129442852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,7168,0.010133333504199982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,2048,0.013422222601042854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,4096,0.012079111403889127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,8192,0.09992000129487778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,2048,0.019395555059115093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,128,0.0018684443914228016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,2048,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,51200,4.658165401882596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,3072,0.9551386303371854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,3584,0.02854666776127285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,12288,0.07327644692526923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,128,0.016174222032229107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,256,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,51200,0.20350044303470186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,16384,0.09803022278679742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,3072,0.01127022256453832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,128,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,2560,0.011511111425028907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,2048,0.015672889020707872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,2048,0.008242666721343994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,256,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,768,0.012140444583363004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,2560,0.008596444295512306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,1024,0.02997244397799174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,6144,0.09209599759843613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,5120,0.13751199510362414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,2560,0.0506551100148095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,3584,0.013228444589508904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,51200,0.3551955487993028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,10240,0.012958221965365939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,6144,0.585594654083252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,3584,0.047913776503668896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,7168,0.08016355832417806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,2560,0.017690666847758822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,768,0.022319111559126113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,51200,0.22366844283209908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,2560,0.018537777993414137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,6144,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,128,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,1024,0.003583111282851961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,3072,0.01554666625128852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,2048,0.013046222428480784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,8192,0.11291910542382134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,16384,0.07761955261230469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,768,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,16384,0.0365626679526435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,512,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,2048,0.022291556000709534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,4096,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,16384,0.060567114088270396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,1024,0.0028124445428450904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,1024,0.00795111142926746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,768,0.044734223021401294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,16384,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,512,0.01349244432316886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,6144,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,12288,0.04462933209207323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,768,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,4096,0.018519111805491976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,64,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,5120,0.04745599958631727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,65536,0.12999644544389513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,1536,0.00388711111413108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,65536,0.9119200176662869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,1024,0.026682666606373254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,5120,0.2649244467417399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,65536,0.08063111040327284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,2048,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,65536,1024,1.747511969672309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,51200,0.07273066706127591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,3072,0.28790222273932564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,16384,0.04872266782654656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,3584,0.01348622226052814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,2560,0.004920000003443824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,768,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,768,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,2560,0.04226933254135979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,2048,0.02458933326933119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,16384,0.020290666156344943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,1024,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,1024,0.04239733351601494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,5120,0.012196444802814059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,1024,0.009983110758993361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,51200,0.05640088849597507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,128,0.04707644383112589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,51200,0.10633866654502021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,2048,0.10238489177491929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,6144,0.01295555548535453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,1536,0.05994488795598348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,2560,0.015095111396577625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,2048,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,1024,0.027601778507232666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,51200,0.10762755738364325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,768,0.034817778401904635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,1536,0.01900622248649597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,64,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,5120,0.03255733185344272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,4096,0.012987555729018318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,51200,0.3780657715267605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,10240,0.020266667008399963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,32,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,32,1536,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,3072,0.005391110976537068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,256,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,64,0.0018684443914228016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,51200,0.03861422340075175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,16384,0.02425777746571435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,7168,0.01349422252840466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,3072,0.019951999187469482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,65536,2.2449234856499567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,6144,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,4096,0.049867557154761426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,1024,0.017851556340853374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,2560,0.0810826685693529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,1536,0.014081777797804939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,51200,0.4237182140350342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3584,10240,0.01479200025399526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,1024,0.022495999932289124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,51200,1536,0.09889155626296997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,1024,0.004208000169859992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,65536,0.49453777737087673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,768,7168,0.021692444880803425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,5120,0.03436977664629618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,65536,0.05846222241719564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,4096,0.02128711177243127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,2048,0.029214223225911457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,32,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,3072,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,2560,0.005048888838953442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,8192,0.05667466587490506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,1024,0.01276088919904497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,1536,0.03501333461867438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,16384,0.21352977222866484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,16384,0.028797331783506606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,1536,0.038303110334608294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,32,512,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,5120,0.02251288957066006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,3584,0.0789324442545573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,768,0.014472888575659858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,256,0.049672888384925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,10240,512,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,10240,0.09448355436325073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,8192,0.05766933494144016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,16384,0.29374578264024526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,6144,0.027349332968393963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,768,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,12288,256,0.0032293332947625052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,7168,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,1536,0.01648622254530589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,768,0.005872000008821487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,1536,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,512,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,768,0.0034186666210492453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,64,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,128,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,8192,0.005682666682534748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,8192,0.06774666574266222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,3584,0.04593422346644931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,8192,0.023574221465322707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,2048,0.06643466817008124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,16384,0.0243511117166943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,128,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,128,0.0018115554832749898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,8192,0.011326221956147088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,5120,0.014562666416168213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,16384,0.048581331968307495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,512,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,4096,0.013545778062608508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,2048,0.010796444283591377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,8192,0.07629066705703735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1024,10240,0.009591999981138442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,64,0.0018577778504954444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,16384,16384,0.9290604061550565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,512,0.008415111237102086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,5120,0.0517502228418986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,5120,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,65536,0.5356204244825575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,3072,0.018811555372344125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,8192,1.8876444498697917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,3584,0.015863110621770222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,5120,0.019175110591782462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,256,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,1024,0.03375022278891669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,768,0.01440000037352244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,3072,0.22265511088901094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,8192,256,0.029783109823862713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,4096,0.29527465502421063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,10240,0.025786666406525507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,7168,0.02749600013097127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,2048,0.026756443911128577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,256,6144,0.021351110604074266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,2048,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,16384,0.016200888488027785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,4096,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,10240,0.16180622577667236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,65536,0.4388497670491536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,64,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,64,0.011576000187132092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,64,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,768,0.013897778259383308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,7168,0.08133688900205824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,7168,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,1024,0.01350844403107961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,6144,0.09636355770958795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,6144,0.026361778378486633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,1536,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,1536,0.07549066676033868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,16384,0.052757332722345986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,128,0.01312888910373052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,6144,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,6144,0.011072888970375061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,1536,0.030455999904208716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,3072,0.47664891348944766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,7168,0.0324924455748664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,1536,0.019870221614837646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,3072,0.054343998432159424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,128,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,8192,0.11223733425140381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,32,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,3584,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,1536,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,512,0.02221155497762892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,1536,0.006250666661394968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,3072,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,1024,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,512,6144,0.022237333986494277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,12288,0.08516889148288304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,768,0.0026515554636716843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,2048,0.05623822079764473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,64,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,7168,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,7168,0.01444533301724328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,5120,0.009305777649084726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,8192,0.07107110818227132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,32,0.03932799895604452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,10240,65536,3.4239590962727866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,8192,0.012179555164443122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,12288,0.4097235467698839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,65536,5120,0.5989191267225478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,1024,0.011334222224023608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,1536,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,2048,0.014290667242474027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,64,0.025797333982255723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,1536,0.005190222213665645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,7168,0.048327111535602145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,16384,256,0.06281422244177924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,65536,0.6370808813307021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,3072,0.03361777795685662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,1024,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,65536,2048,0.04858577913708157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,768,0.0243857784403695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,65536,0.02981511089536879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,5120,0.02494488822089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,51200,0.7301928732130262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,1024,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,256,0.01459022197458479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,3072,0.05295377638604906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,128,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,128,0.0018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,4096,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,12288,0.01752355529202355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,3584,0.09153155485788982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,256,0.008068444828192392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,256,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,7168,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,51200,0.07842666572994657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,32,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,6144,0.07825866672727796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3072,16384,0.21189688311682808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,2048,0.022953778505325317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,10240,0.01996799972322252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,32,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,65536,0.4673768679300944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,3072,0.10362222459581162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,2048,0.007231111327807109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,3584,0.03324266605907016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,32,1536,0.0035439998739295532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,768,0.013391999734772576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,32,1536,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,6144,512,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,51200,12288,1.205534193250868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,3584,0.07003644439909193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,51200,0.06495466497209337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,6144,0.1257964505089654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,1536,0.022647110952271357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,3584,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,4096,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,16384,0.01645866698688931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,256,0.01638488968213399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,7168,0.2957644462585449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,4096,0.022291556000709534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,1536,0.020250666472646926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,128,0.02236533330546485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,512,0.02027555637889438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,7168,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,65536,0.058113780286577016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,6144,16384,0.1956231064266629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,12288,0.04742044541570875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,512,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,8192,0.07182221942477755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,6144,0.45542044109768337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,512,0.003162666741344664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,51200,0.47876797782050234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,2048,0.06174577607048882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,10240,0.015576000014940897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,8192,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2560,1536,0.05875911315282186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,2048,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,3072,0.008332444561852349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,1536,0.18531555599636504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,1024,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,12288,0.01722400055991279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,5120,0.012439111040698158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,512,0.026546667019526165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,5120,0.00833511104186376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,8192,0.011042666931947073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,51200,0.6707822481791178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,51200,0.7266835636562772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,3584,0.013412444127930535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,12288,0.029559999704360962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,2048,0.026328888204362657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,51200,1.7089502546522353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,512,0.05362933211856418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,256,0.0026417778184016547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,128,0.01754755609565311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,64,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,16384,0.09929333130518596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,1536,0.0034968890249729156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,6144,1.466843605041504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,768,0.033160001039505005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,4096,0.03087644444571601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,32,0.008044444852405125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,3584,0.022679999470710754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,1536,0.07298488749398126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,1024,0.0029484445436133277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,16384,0.19083644284142387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,768,0.03018222252527873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,512,0.005294222384691238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,4096,4096,0.022287110487620037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,10240,7168,0.021584888299306233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,1024,0.0038480001191298165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,768,0.011796444654464722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,5120,0.030540443129009668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,10240,256,0.009398221969604492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,6144,5120,0.031397332747777305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,16384,128,0.015614221493403116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,2560,0.006976000136799282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,51200,0.9851297802395291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,65536,1.3900221718682184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,256,0.0768337779574924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,12288,0.019348444210158456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,4096,0.029515554507573444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,51200,0.015626667274369132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,768,0.012454222473833295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,512,0.011146666275130378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,1536,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,3072,0.015947555502255756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,3584,0.024566221568319533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,16384,0.12544177638159856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,3072,0.07038311163584392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,4096,0.10104178057776557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,1536,0.004855111241340637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,10240,0.30407823456658256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,8192,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,768,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,1024,0.027502222193611994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,256,0.011714666253990598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,51200,0.07548889186647204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,12288,0.05138399865892199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,2048,0.049754665957556836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,51200,12288,0.22065156035953096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,65536,2560,0.9146391550699869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,5120,0.011046222514576383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,3072,0.01735555628935496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,3584,0.03436977664629618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,6144,12288,0.2291475666893853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,51200,0.05640977621078491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,16384,0.022281777527597215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,3072,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,32,0.011382222175598145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,1024,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,4096,0.006899555524190267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,65536,0.33510398864746094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,2560,0.007253333098358578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,7168,0.00959288908375634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,2560,0.011394666300879585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,32,0.049976888630125255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,65536,0.49234310785929364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,7168,0.07198488712310791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,2048,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,10240,0.140538665983412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,2560,0.0882951087421841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,8192,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,5120,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,8192,0.057426664564344615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,768,0.004295111116435793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,1024,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,4096,0.03330399923854404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,3072,0.012796444197495779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,2560,0.007343110938866933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,256,0.007089777953094906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,2560,0.004906666775544484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,2048,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,256,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,3584,0.010612444745169746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,1536,0.01717333330048455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,768,0.015804444750150044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,512,0.006224888894293044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,5120,0.038894222842322454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,2560,0.007671111159854465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,64,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,51200,0.16954844527774385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,512,10240,0.07920000288221571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,7168,3584,0.013434666726324292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,6144,0.04910222358173794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,64,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,256,256,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,5120,0.019487111104859244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,6144,0.13924088742997912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,2048,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,3584,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,6144,0.012420444852775998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,256,0.13414043850368923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,64,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,4096,0.21481333838568795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,10240,0.18254666858249238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,4096,0.013764444324705334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3072,51200,0.6929599973890516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,12288,0.15545422501034206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,128,0.0032035555276605818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,64,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,6144,0.16777777671813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,7168,0.014928888943460254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,256,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,128,0.031161778502994116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,768,0.013618666264745923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,3072,0.05411822266048855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,6144,0.01218488895230823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,8192,0.02576711111598545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,3072,0.013163555827405719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,12288,0.03233244352870517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,1024,0.003212444484233856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,3072,0.016889777448442247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,5120,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,3584,0.021737777524524268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,4096,0.017496888836224873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,16384,0.1801617807812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,3072,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,512,0.011606222225560082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,1536,6144,0.009405333134863112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,512,0.011348444554540845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,7168,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,6144,0.07933333184983996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,12288,0.02401422295305464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2048,5120,0.00869688888390859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,6144,0.05383822321891785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,768,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,8192,0.06995999813079834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,512,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,8192,7168,0.01954844428433312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,3072,0.011297778122954898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,12288,0.016145777371194627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,65536,0.36267466015285915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,2048,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,3072,0.015799999237060547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,256,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,4096,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,10240,0.01641688909795549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,8192,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,7168,0.019352000620630052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,128,2048,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,8192,0.010888000329335531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,16384,0.03392533461252848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,256,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,51200,0.12040978007846409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,7168,0.0391813317934672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,5120,0.06815377871195476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,1536,0.030255999830034044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,128,0.03050222330623203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,12288,0.02985422147644891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,16384,0.020336000455750358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,12288,15.3418697781033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,4096,0.03045155604680379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,8192,0.01037333326207267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,3584,0.05481688843833076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,4096,0.07582488987180921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,1536,0.2196284400092231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,1536,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,2048,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,1024,0.09887733062108357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,256,0.025693333811230127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,16384,0.4030622111426459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,8192,0.029430223835839167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,10240,0.014141332772043018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,8192,0.5844800207349989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,10240,0.06148978074391683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,12288,0.21532000435723198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3072,768,0.0041742221348815495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,3072,0.33885688251919216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,1536,0.07558933231565687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,4096,0.019367999500698514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,5120,0.013075555364290873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,768,0.0041706665522522396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,51200,1.1300800111558702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,1024,0.012094222009181976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,6144,0.013175999952687157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,51200,0.86572175555759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,51200,0.48450575934516055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,65536,0.1053404476907518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,64,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,5120,0.13313955730862087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,4096,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,65536,0.16092888514200845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,51200,0.029155555698606703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,51200,0.07779377698898315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,8192,0.06630222002665202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,8192,0.015594666202863058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,1536,0.013916444447305469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,1024,0.01939022209909227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,256,0.015121777852376303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,7168,0.1375217835108439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,256,0.0033582221302721235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,512,0.03468711177508036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,2048,0.010655110908879174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,768,0.0037288889288902283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,256,0.010503999888896942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,4096,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,2048,0.2944835556877984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,16384,0.01982666717635261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,5120,0.049489776293436684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,768,0.013774221969975365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,6144,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,4096,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,3584,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,3584,0.03233066532346938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,2560,0.01032977799574534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,1024,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,12288,3072,0.027612444427278306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,2048,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,51200,0.14128266440497503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,2560,0.008451555338170793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,3072,0.017518222332000732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,2048,0.07391466697057088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,8192,0.022520888182852004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,6144,0.012332444389661154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,1536,0.07193955447938707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,8192,0.026478222674793665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,1024,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,2048,0.01218933363755544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,65536,0.22065333525339761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,2048,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,12288,0.05687644746568468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,128,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,8192,0.025278222229745653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,32,0.004221333397759331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,4096,0.035775999228159584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,2048,0.05686222182379829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,3072,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,3072,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,512,0.00872444444232517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,8192,0.7186089091830783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,65536,0.5830782254536947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,51200,0.11107289128833348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,128,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,128,0.006309333360857434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,10240,1.4098488489786785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,6144,256,0.01108000013563368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,12288,0.017608889275126986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,256,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,5120,0.6257635752360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,1024,0.007338666253619724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,768,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,51200,0.08758755524953206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,128,0.005211555709441503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,32,0.006344888773229387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,3584,0.02987288766437107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,8192,8192,0.04619377851486206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,3072,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,16384,2048,0.06587200032340156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,512,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,1536,0.007274666594134436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,256,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,10240,5120,0.01605422298113505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,768,0.006267555471923616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,768,0.01581333412064446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,7168,0.48666932847764754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,768,0.018605333235528734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,12288,0.03122044603029887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,128,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,8192,0.008906666603353288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,2560,0.023783110909991797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,16384,0.1468542284435696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,3072,0.054309331708484225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,256,0.025443555580245122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,16384,1536,0.6662889056735569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,12288,0.0767093300819397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,10240,0.09326489104164971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,128,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1024,5120,0.008373333348168267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,65536,0.8537128766377767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,2048,0.052073776721954346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,256,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,1536,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,3072,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,256,0.024475556280877855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,16384,0.23617511325412324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,51200,0.28630932172139484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,1536,0.09496177567376031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,3072,0.02516888909869724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,3072,0.07565066549513075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,256,0.004242666479614046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,2048,0.00675733346078131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,512,0.0028400001012616684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,51200,0.06536710924572416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,64,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,5120,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,128,0.012823111481136747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,65536,0.05643288956748115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,512,0.02093066606256697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,3584,0.006587555425034628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,6144,0.02053866618209415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,3072,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,12288,0.09416444434059991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,128,0.012387555506494312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,7168,0.06734489070044623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2048,6144,0.020263110597928364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,512,0.002495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,65536,0.006789333290523953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,2560,0.0166657782263226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,8192,0.05730577972200182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,2560,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,6144,0.04088888896836175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,2048,0.023784000012609694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,5120,0.02811644474665324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,3072,0.02021511064635383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,6144,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,1536,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,64,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,2048,0.06191733148362902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,128,0.0018435555199782054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,4096,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,64,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,4096,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,16384,0.020067556036843192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,8192,0.027580445011456806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3584,12288,0.021503999829292297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,16384,0.643155574798584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,3584,0.014171555638313293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,4096,10240,0.015444444285498725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,4096,0.03764355513784621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,1024,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,128,0.0052391112678580815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,51200,0.1996471087137858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,2048,0.188570671611362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,128,0.015577778220176697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,2048,0.018271999226676095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,2048,0.030472887886895075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,12288,0.12852088610331217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,1024,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,12288,768,0.022330666581789654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,512,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,32,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,768,0.1667262183295356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,2560,0.02942044536272685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,128,0.00835111074977451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,1536,0.011898666620254517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,3584,0.0310479998588562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,5120,0.01310400002532535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,3072,0.03740977909829881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,8192,65536,0.16983289188808867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,4096,0.02181244393189748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,32,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,7168,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,65536,0.8756977717081705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,128,0.011062222222487131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,512,0.007368889119890001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,51200,0.058223108450571694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,5120,0.037300444311565824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,8192,0.054530666934119336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,2560,0.12271732754177517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,4096,0.011769777370823754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,7168,0.017133333616786532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,1024,0.004464888738261329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,128,0.0162782222032547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,7168,0.03750577900144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,3072,0.028049776951471966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,3072,0.01682133310370975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,5120,0.022626666559113398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,4096,0.05651022328270806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,128,0.005375111268626318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,4096,0.06662755542331271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,2560,0.00850133349498113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,5120,0.010111110905806223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,2560,0.02683911058637831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,768,0.02369600037733714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,65536,0.014742222097184924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,4096,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,7168,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,32,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,1024,0.006993778049945831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,4096,0.00998755544424057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,768,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,768,0.005671999934646819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,128,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,3584,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1024,4096,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,16384,0.5369946691724989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,512,0.014331556028789945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,3584,0.08076977729797363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,5120,0.02583822276857164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,128,0.014825777875052558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,32,0.020045333438449435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,10240,0.024120888776249353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,6144,0.24869955910576713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,4096,0.1264302200741238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,128,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,16384,0.07065511412090726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,32,0.004703999807437261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,32,0.11078221930397882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,12288,0.18589066134558785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,768,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,1536,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,128,0.21909777323404947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,16384,0.06161599689059787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,8192,0.1428791152106391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,1536,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,4096,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,3584,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,65536,0.06984711355633207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,3072,0.014540443817774454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,6144,0.008990222381220924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,7168,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,4096,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,8192,0.018988443745507132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,16384,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,128,0.012738666600651212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,512,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,256,0.021731555461883545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,6144,0.005593777944644292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,2560,0.0147733340660731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,16384,0.050945778687795006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,4096,65536,0.08352710803349812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2560,65536,0.06364444229337904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1024,8192,0.010299555957317352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,16384,512,0.05996266338560316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,5120,10240,0.017691555950376723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,1024,0.03808355662557814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,1536,0.01311822235584259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,2048,0.06592800219853719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,65536,0.04161422120200263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,1536,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,65536,0.10834666755464341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,256,0.15425154897901747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,6144,0.096160888671875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,16384,0.016240888171725802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,5120,0.015810666812790763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,16384,0.024576000041431848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,6144,0.18800532817840576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,768,0.02296177711751726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,3072,0.06032533115810818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,512,0.015607110328144498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,512,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,1536,0.018111111389266122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,128,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,6144,0.02904711167017619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,2048,0.004179555508825514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,128,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,5120,12288,0.025068443682458665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,65536,0.08511200216081406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,2560,0.014448889427714877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,16384,0.02312799957063463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,12288,0.009319999979601966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,16384,0.028775999943415325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,16384,0.045588443676630654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,128,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,512,0.01238844460911221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,10240,0.8524284362792969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,51200,0.046167999505996704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,512,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,128,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,10240,0.02307022280163235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,3072,0.007986666427718269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,3072,0.005341333233647876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,51200,0.051727109485202365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,65536,7168,0.24602932400173613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,8192,0.010411555568377176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,6144,0.31088444921705455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,3584,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,256,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,6144,0.36700622240702313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,4096,10240,0.020232889387342665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,12288,0.6509928703308105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,16384,0.01647111111217075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,7168,0.09802310996585423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,10240,0.050937778419918485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,256,0.04832888974083794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,768,0.01793599956565433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,4096,0.022467555271254644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,6144,0.13383466667599148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,1024,0.011381333072980246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,5120,0.08875022331873576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,256,0.002830222249031067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,4096,0.0185253338681327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,7168,0.35761600070529515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,128,0.012744888663291931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,7168,0.030777778890397813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,2560,0.07413244247436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,4096,0.3519297705756293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,256,0.010966222319338055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,12288,0.05844089057710436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,4096,0.03278133273124695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,2560,0.04054844379425049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,16384,0.032219555642869734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,6144,0.3956577777862549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,8192,0.14857510725657144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,65536,0.06390577554702759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,12288,0.03807466559939914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,4096,0.198634664217631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,128,0.005912888795137405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,10240,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,1024,0.030819555123647053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,4096,0.09338133202658759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,1024,0.008998221821255153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,2048,0.006030222194062338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,1536,0.030478222502602473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,10240,0.02203377750184801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,12288,0.1069599986076355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,5120,0.025066667132907446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,2560,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,1536,0.011072888970375061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,1024,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,768,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,128,0.00350577798154619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,8192,0.028646220763524372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,51200,1.1624293857150607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,7168,0.031833777825037636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,16384,0.058803558349609375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,8192,0.07586310969458686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,8192,0.03764977720048692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,32,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,65536,0.34580087661743164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,65536,1.0953644646538627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,256,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,7168,0.057414220439063184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,3584,0.2567431131998698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,256,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,64,0.01516266663869222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,7168,0.5248595343695747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,6144,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,64,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,2048,0.0107022225856781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,8192,0.03566222058402167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,51200,3.275027592976888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3584,5120,0.028804444604449805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,10240,0.5666942066616482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,7168,0.06520622306399874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,3584,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,128,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,512,0.02495999965402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,65536,20.433603922526043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,3584,1.7468409008449977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,768,0.011201777391963534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,768,0.01822933389080895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,65536,0.24868888325161406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,32,0.006335999816656113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,6144,0.01107911103301578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,2560,0.016906667086813185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2048,2560,0.005505777895450592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,3072,0.007711999946170383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,3584,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,32,0.001839111041691568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,3072,0.008479110896587372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,1024,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,2048,0.006602666858169768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,65536,3.504335191514757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,1024,0.030392001072565716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,4096,0.03875822159979079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,12288,0.013561777770519257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,7168,0.029903110530641343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,8192,0.01459022197458479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,2560,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,6144,0.04933244321081373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,3072,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,65536,0.06707021925184461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,8192,0.04313866628540886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,65536,128,0.051705777645111084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,51200,16384,0.11523555384741889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,2048,0.02643377747800615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,16384,0.03558933403756883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,768,0.01812799937195248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,768,0.029743111795849268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,3072,0.012968888713253869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,65536,0.2876977655622694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,1536,0.4062648879157172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,12288,0.02459822263982561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,512,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3072,1536,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,768,0.00299822228650252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,3072,0.019193778435389202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,768,0.009477333062224919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,4096,0.01852799952030182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,3072,0.010667555862002902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,51200,9.520968119303385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,51200,0.24368000030517578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,10240,0.11554666360219319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,51200,0.12832178009880915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,512,0.003552000141806073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,2048,0.006648889018429651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,64,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,64,0.0017875555074877208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,2048,0.014767110347747803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,16384,0.020625778370433383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,5120,0.013760888742076026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,16384,0.2942897743648953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,32,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,768,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,4096,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,512,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,3072,0.019928889142142404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,1024,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,65536,0.05679821968078613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3072,3072,0.00776533368561003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,1024,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,65536,1.400728013780382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,3072,0.033032000064849854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,3072,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,32,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,2560,0.020427554845809937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,128,0.0021884443445338146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,256,0.03850311040878296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,12288,0.10751110977596706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,16384,1.1706657409667969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,64,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,2560,0.008356444537639618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,10240,2.5444649590386286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,1536,0.009646221995353699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,16384,0.12311289045545791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,3072,0.032270222902297974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,6144,0.02831288841035631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,12288,256,0.01718933383623759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,7168,0.02852177619934082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,3072,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,256,10240,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,1536,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,65536,0.05916621949937609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,768,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,1536,0.016705777910020616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,64,1536,0.010206222534179688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,512,0.00868266655339135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,3072,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,51200,0.8131910959879557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,16384,0.2840079996320936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,3584,0.008336000144481659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,1536,0.0068808889223469635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,2048,0.010054222411579555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,128,0.018783110711309645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,10240,0.29684445593092174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,6144,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,128,0.009185777770148383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,6144,0.010968888799349466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,128,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2560,3584,0.015836444165971544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,5120,1.3144666883680556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,7168,0.04722311099370321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,7168,0.039000001218583845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,256,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,12288,0.06360444757673475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,7168,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,7168,0.507156425052219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,7168,0.08094666401545207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,1024,0.02331999937693278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,1536,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,3072,0.03793155484729343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,12288,0.12508889039357504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,256,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,51200,128,0.11686933040618896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,128,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,4096,0.04034311241573758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,8192,0.018930666976504855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,3584,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,7168,512,0.009958222508430481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,65536,0.15353422694736057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,6144,0.010087110930018956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,1536,0.06806577576531304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,7168,0.24597154723273384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,5120,0.014279110564125909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,32,0.05348711212476095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,512,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1024,7168,0.048379557000266186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,3072,0.34301598866780597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,1536,0.022913777165942725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,7168,0.020004444652133517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,10240,0.011697777443461947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,1536,0.008189332981904348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,3584,0.011108444796668159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,8192,0.14699555767907038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,8192,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,10240,2.444978713989258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,2048,0.034698665142059326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,4096,0.04713955521583557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,12288,0.025384889708624944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,1536,0.0038542221817705366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,128,0.021225778592957392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,10240,0.12563199467129177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1024,128,0.006311999840868845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,4096,64,0.02494488822089301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,12288,0.02889333168665568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,2560,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,128,0.005208000126812193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,4096,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,6144,0.046441776884926685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,512,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,128,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,3584,0.008258666429254744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,8192,0.04408533374468485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,768,0.016494222813182406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,65536,0.10649777783287896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,512,0.011804444094498953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,3584,0.031277332040998675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,16384,3584,0.0682142244444953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,51200,4096,0.20856711599561903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,1024,0.0075502221783002215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,128,256,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,51200,0.05593244234720866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,16384,0.04065955678621928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,12288,0.20799377229478624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,2560,0.01480977733929952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,2048,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,512,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,3072,0.03071199854214986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,3584,0.015468445089128284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,1536,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,51200,0.05309066507551405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,3072,0.04713244570626152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,5120,0.060806221432156034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,65536,0.47758844163682723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,10240,0.22570843166775176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,6144,0.46554046207004124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,256,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,7168,0.05662844578425089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,8192,0.03184088733461168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,256,0.015471999843915304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,7168,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,16384,0.025035555164019268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,12288,0.047107554144329496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,512,0.0031715554909573663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,768,0.012030222349696688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,3584,0.017297777864668105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,128,0.01790488925245073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,65536,0.21857955720689562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,51200,0.017827555537223816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,768,4096,0.008981333010726506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,768,0.021988444858127173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,10240,0.0315688881609175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,64,0.021615111165576514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,51200,0.32812976837158203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,256,0.014675555957688225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,51200,1.1204328536987305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,6144,0.015620445211728414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,6144,0.12046933174133301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,8192,0.4909671147664388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,10240,0.05065955718358358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2048,6144,0.012063999970753988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,768,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,2560,0.026776888304286536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,1024,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,7168,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,4096,0.0381991101635827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,3584,0.47270753648546004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,1024,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,7168,0.05456799930996365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,128,0.022648000054889258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,12288,256,0.017214222086800467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,12288,0.12864443990919325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,1024,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,3584,0.00793866647614373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,2560,0.06448000007205539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,12288,1536,0.03404622276624044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,3072,0.03601955705218845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,768,0.00554666668176651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,2560,0.007137777904669444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,2560,0.027777777777777776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,2560,0.07765422264734904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,256,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,128,0.005679111099905438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,10240,0.2105111148622301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,1536,0.006713777780532837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,2560,0.00489066665371259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,1536,0.006496888895829518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,10240,0.02201777696609497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,5120,0.7113679779900445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,12288,0.12993510564168295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,2560,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,8192,0.040876444843080305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,256,0.006334222025341458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,4096,0.023058666123284235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,512,0.009759110709031424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,1536,0.01731911136044396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,65536,0.07700977722803752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,768,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,7168,0.010680889089902243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,3072,0.037288000186284385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,51200,0.05957421991560194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,128,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,65536,0.29962489340040416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,12288,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,256,0.0021688888470331826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,1536,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,65536,0.21494932969411215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,768,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,10240,0.17161244816250273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,65536,1.2591671413845487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,256,0.002458666761716207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,6144,0.020445333586798776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,64,0.029095109966066148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,3584,0.018239999810854595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,256,0.07490577962663439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,8192,0.18632977538638643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,3072,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,4096,0.009564444422721863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,8192,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,1536,0.07870399951934814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,3072,0.05728444125917223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,64,0.005648889061477449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,8192,1.3202248679267037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,2048,0.009215110705958473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,1536,0.01996888882584042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,4096,0.01038933296998342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,128,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,128,0.02510044475396474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,1536,0.02897955642806159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,768,0.006223111102978389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,2048,0.09970311323801677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,12288,0.6322506798638238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,32,3072,0.03545244534810384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,8192,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2048,1024,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,51200,2.0597616831461587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,2560,0.022614222433831956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,1024,0.05939110782411364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,1024,0.013355555633703867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,256,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,2048,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,8192,0.01389600005414751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,128,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,2560,0.010768888725174798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,2560,0.3966080082787408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,5120,0.009921777579519484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,6144,0.008743999732865227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,512,0.006032889087994893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,65536,1.3232907189263237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,10240,0.07936177651087443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,32,0.0018808889306253856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,4096,0.01776000029510922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,3584,0.015213333898120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1536,12288,0.05204711026615567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,128,0.008087111016114553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,4096,0.01683288812637329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,1024,0.0041253334946102565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,512,0.003162666741344664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,51200,8.129724290635851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,2048,0.37810932265387637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,6144,0.020561777883105807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,4096,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,51200,0.05224355392985874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,2560,0.007290666302045186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,12288,0.06507022513283624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,5120,2.0180977715386286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,256,0.19420088662041557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,12288,0.5465057690938314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,7168,10240,0.036481777826944985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,3072,0.013071111506885953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,1536,0.010051555931568146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,1024,0.018518222702874076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,256,0.019854222734769184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,3584,0.015464888678656684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,128,0.036531554328070745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,2560,0.03803199860784743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,512,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,4096,0.041123555766211614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,2560,0.013834666874673633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,3584,0.007920888562997183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,16384,0.05286488930384318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,3072,0.013048888908492194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,768,0.01440355512830946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,2560,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,10240,0.028766221470303003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,1024,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,8192,0.013971555564138623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,51200,0.39774311913384336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,3584,0.06953599717881945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,512,65536,0.15404710504743788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,3584,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,1024,0.03397688931889004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,32,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,3072,0.01626844373014238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,3072,0.005708444449636672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,128,0.028004444307751123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,64,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,2560,0.04901599884033203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,4096,0.036179555786980525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,8192,0.04018933243221707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,2048,0.19594400458865693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,2048,0.023757333556811016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,1536,0.039172444078657366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,51200,0.06622222397062513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,7168,0.017866666118303936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,3584,0.03492533498340183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,7168,0.02164711058139801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,768,0.003846222327815162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,1536,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,512,0.11687200599246556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,16384,4096,1.5887617535061305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,51200,0.05348177750905355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,51200,0.036335110664367676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,1024,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,3072,0.016198222835858662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,51200,0.023518222901556227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,10240,0.08011733161078559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,1024,0.022050667140218947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,1024,0.057224002149369985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,51200,0.05004088746176826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,768,0.0028115556471877625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,3584,0.02975644336806403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,7168,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,51200,0.05106222298410204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,256,0.014797333214018079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,3072,0.01275644451379776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,768,0.057351112365722656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,6144,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,256,0.014914666612943014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,2048,0.03825511203871833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,256,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,768,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,65536,0.8424888716803657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,65536,0.06686666938993666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,51200,768,0.6230737898084853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,16384,6.708676656087239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,65536,0.3638666735755072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,2048,0.7388453483581543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,2048,0.020792000823550753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,8192,0.031930665175120033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,16384,0.11084089014265273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,65536,5120,0.560732470618354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,12288,0.05922489033805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,2048,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,32,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,7168,0.018914666440751817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,32,0.007841777470376756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,8192,0.030972444348865088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,8192,0.055254220962524414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,512,0.00996888925631841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,2048,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,12288,0.06537510951360066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,8192,0.024804444776640996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,2048,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,1536,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,64,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,128,0.002469333302643564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,256,0.017512000269360013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,8192,0.01384888920519087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,4096,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,128,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,7168,7168,0.01661777827474806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,7168,0.031145777967241075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,3072,0.028991109795040552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,10240,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,512,0.002847111059559716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,128,0.012453333371215396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,2048,0.006496888895829518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,51200,4.211606343587239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3072,768,0.05427822139528063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,3072,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,6144,0.01834488908449809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,1536,0.05366488960054186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,768,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,2048,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,51200,10.480083889431423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,3072,0.03314400050375197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,1536,0.01936533384852939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,65536,0.1270044379764133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,3072,0.08235110839207967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,1536,0.012786666552225748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,16384,0.07760088973575167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,16384,0.2903173234727648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,128,0.03482666611671448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,32,0.004187555362780889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,7168,0.013735111388895245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,768,0.06048355499903361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,4096,512,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,256,0.014023999373118082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,2560,0.09316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,65536,0.19455822308858237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,256,0.01702222228050232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,3072,0.025808889004919264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,5120,0.04451910985840691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,51200,14.005419413248697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,2048,0.005046222358942032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,10240,0.02576622201336755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,4096,0.007000000112586551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,2560,0.01275466630856196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,8192,0.03458222084575229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,10240,0.043896890348858304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,3584,0.04198222359021505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,51200,0.24918222427368164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,768,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,8192,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,7168,5120,0.011086222198274402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,128,0.007842666572994655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,6144,10240,0.43480979071723086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,4096,0.013608000344700284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,51200,0.1598657766977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,8192,0.013164444102181328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,256,0.016178665889634024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,3584,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,32,0.014434667097197639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,768,0.0038622220357259116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,12288,0.033187554942237005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,4096,0.03226844469706217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,768,0.043064888980653554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,65536,0.3130648930867513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,16384,4.065758175320096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,65536,2.4581076304117837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,12288,0.04876888791720072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,65536,0.07896177636252509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,32,0.01089333328935835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,6144,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,768,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,8192,0.032904889848497175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,128,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,8192,0.19006755616929796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,51200,0.4275875621371799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,5120,0.07842666572994657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,1536,0.05346133311589559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,10240,0.052292442984051175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,1536,0.0069208890199661255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,4096,0.020616888999938965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,1536,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,64,0.0041457778877682155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,8192,8192,0.31350933180914986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,768,0.012303999728626676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,3072,0.11055466863844131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,5120,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,51200,0.06258310874303182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,8192,0.01092977739042706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,256,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,16384,0.044842667049831815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,5120,0.041540443897247314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,512,10240,0.0417137775156233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,1536,0.004875555634498596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,10240,0.016488000750541687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,65536,0.1514657735824585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,2048,0.050907555553648204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,51200,0.05473866727617052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,5120,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,8192,0.011369778050316704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,64,5120,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,7168,0.02351555559370253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,3584,0.02249777813752492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,8192,0.005665777872006099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,16384,0.03303111261791653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2560,12288,0.017495110630989075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1536,8192,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,3584,0.043040888177024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,8192,0.11072711149851482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,512,51200,0.13611111376020643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,5120,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,10240,0.020582222276263766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,1024,0.017688888642523024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,12288,51200,1.5138613382975261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,7168,0.025974222355418738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,65536,0.10386666986677383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,7168,0.11180178324381511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,128,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,65536,10240,3.493519041273329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,5120,0.05106222298410204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,8192,0.15931110911899143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,4096,0.18771466943952772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,16384,1.5366747114393446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,8192,0.23611821068657768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,768,0.022224000758594934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,1536,0.005791999813583162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,12288,0.19515732924143472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,1024,0.007640889121426477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,128,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,128,0.01552888916598426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,8192,16384,0.08306844366921319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,2560,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,16384,0.04467466804716322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,256,5120,0.009090666969617208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,128,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,7168,0.34908178117540145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,16384,0.009311999711725447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,51200,0.016020443704393175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,1024,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,7168,1.1954844792683919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,12288,0.017254221770498488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,7168,0.10739199982749091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,10240,0.049599998527103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,3584,0.010699555277824402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,65536,0.08170311318503486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,7168,0.022619555393854778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,7168,0.01923199991385142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,65536,0.5415413114759657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,16384,0.02455288834042019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,32,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,8192,0.025645333859655593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,2560,0.058796445528666176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,8192,0.07502666446897718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,10240,0.23644532097710502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,12288,0.22622309790717232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,512,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1536,65536,0.21255466673109266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,3584,0.319576899210612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,256,0.01789777808719211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,3072,0.036008887820773654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,16384,0.05072089036305746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,1536,0.023269333773189124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,3584,0.019900444481107924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,512,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,12288,0.16914489534166124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,2048,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,8192,3584,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,16384,10240,0.08345866865581936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,768,0.13456888993581137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,1536,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,5120,0.009691555466916826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,1024,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,51200,0.11327466699812148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,768,0.01069866700304879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,10240,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,5120,0.009126221968068017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,8192,1.0906960169474285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,8192,0.010411555568377176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,3072,0.007364444434642792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,32,0.010652444428867765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,4096,0.09053511089748806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,768,0.02264444364441766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,2048,0.12564978334638807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,256,0.016511110795868766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,5120,0.5145386589898003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,128,65536,0.14659289518992105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,7168,0.01629688839117686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,51200,128,0.040656887822681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,16384,0.010503999888896942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,3072,0.03678222166167365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,6144,0.35615377955966526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,6144,0.012455110748608908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,65536,0.048002666897243924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,1024,0.004518222063779831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,3584,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,7168,0.3658497863345676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,12288,0.048787554105122886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,12288,0.18166489071316191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,6144,10240,0.08886400196287368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,12288,0.037963555918799505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,1536,0.05788355403476291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,3584,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,8192,0.17098044024573433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,2560,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,2560,0.007299555672539606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,8192,0.3185315661960178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,1536,0.048657778236601085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,7168,7168,0.13620977931552464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,128,0.057893335819244385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,8192,0.03110044532352024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,3584,0.17350310749477812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,8192,0.03941511114438375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,6144,0.027730666928821143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,12288,0.032383110788133405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,1536,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,3072,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,3072,0.015311110350820752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,512,0.024567999773555334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,2048,0.02055911057525211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,16384,0.06764888763427734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,768,0.01998933321899838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,16384,0.04614399870236715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,512,0.019474666979577806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,3072,0.038458665211995445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,3072,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,5120,0.02942844563060337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,2048,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,4096,0.011970666547616323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,2560,0.009401777552233802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,16384,0.022032888399230108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,4096,12288,0.13487911224365234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,51200,2.547775904337565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,256,0.008374222450786168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,768,0.015115555789735583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,12288,0.02904711167017619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,65536,0.15037688944074842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,4096,0.03470755616823832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,51200,0.13911110825008816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,2560,0.2515377733442518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,32,0.026556443836953905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,512,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,51200,0.3786462148030599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,3584,0.01586577792962392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,64,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,1536,0.007079111205206976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,512,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,65536,0.10736888647079468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,256,0.017070222232076857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,5120,256,0.026039999392297532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,1536,0.0786000026596917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,12288,0.25205421447753906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,3584,0.010748444332016839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,1536,0.009863999982674917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,7168,0.4145413239796956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,7168,0.17775821685791016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,128,0.0021582223061058256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,3584,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,2048,0.01124622258875105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,2560,0.043755554490619235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,3072,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,128,0.011428444749779172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,65536,0.05398133397102356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,128,0.0019964445382356644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,6144,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,10240,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,1536,0.006240889016124938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,7168,0.016318221886952717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1024,4096,0.020371556282043457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,4096,0.04070933328734504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,3584,0.024735111329290602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1024,256,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,128,0.0038524443904558816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,3584,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,7168,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,51200,0.5060293409559462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,7168,0.10374310943815444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,4096,0.02029244436158074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,512,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,12288,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,256,0.030096888542175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,512,0.014493332968817817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,1024,0.01812977757718828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,64,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,4096,0.022270222504933674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,3072,0.013139555851618448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,12288,0.049970666567484535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,768,0.01391999920209249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,16384,0.02809422214825948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,51200,0.4392373296949599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,1024,0.008232889076073965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,65536,0.06309600008858575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,7168,0.03124533428086175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,16384,0.03457599878311157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,512,0.012062222593360476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,1024,0.0031662223239739737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,2560,0.034280889564090304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,3584,0.036551998721228704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,512,0.01479822231663598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,4096,0.016240888171725802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,4096,0.01204533295498954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,7168,0.0387662218676673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,65536,0.23352445496453178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,6144,0.12452088461981879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,6144,0.03709955679045783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,128,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,1024,0.024712888730896845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,6144,0.027843554814656574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,3072,0.04731288883421156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,32,0.002132444539003902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,1024,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,256,16384,0.04341422186957466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,12288,0.01836266616980235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,12288,0.8930390675862631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,7168,0.03267466690805223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,3584,0.011375111010339526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,32,0.06435110833909777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,1024,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,32,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,1536,0.0038702223036024305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,2560,0.07261600097020467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,3072,0.06950222121344672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,5120,4096,0.008062221937709385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,65536,0.7067715326944987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,2560,0.0151591118839052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,32,0.014298665854665967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,3072,0.029659556017981634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,10240,65536,0.14866756068335638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1024,128,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,12288,0.06958666774961683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,12288,0.582089795006646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,512,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,6144,0.11512977547115749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,6144,0.4751822153727214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,3584,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,3072,0.023053333163261414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,3584,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,65536,6.309694925944011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,3584,0.029616001579496596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,1024,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,12288,0.07668799824184842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,65536,0.06378133429421319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,16384,0.027036445008383855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,2048,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,7168,0.07889511187871297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,8192,0.010677333507272931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,768,2048,0.01481599940194024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,12288,0.038831998904546104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,2048,0.004487110922733943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,256,0.016365332735909354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,5120,0.05644355879889595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,64,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,768,0.01460800071557363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,3072,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,10240,0.020592000749376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,256,0.010777778095669217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,512,0.008360889222886827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,2560,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,10240,3072,0.0122079998254776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,768,0.0081386665503184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,768,12288,0.0566542214817471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,1536,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,51200,0.12793955538007948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,5120,0.023717333873112995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,512,0.011349333657158745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,10240,0.0844764444563124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,128,0.0018924443672100704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,5120,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,65536,0.06940444310506184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,12288,0.055340445703930326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,10240,0.06474933359358045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,32,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,51200,3072,0.2583928902943929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,65536,0.5386640230814616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,2560,0.010639111200968424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2048,4096,0.04624888963169522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,2048,0.12785066498650446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,51200,0.04591466652022468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,3584,0.06274488899442884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,8192,0.0314631097846561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,3584,0.016720889343155753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,2560,0.07177778085072835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,8192,0.17262844244639078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,7168,2048,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,1536,0.018957333432303537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,768,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,1024,0.024458666642506916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,5120,0.018696000178654987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,2560,0.034028444025251604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,3584,0.04023733403947618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,128,0.009405333134863112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,12288,0.06518933508131239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,128,0.006980444822046492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,65536,0.2039333317014906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,4096,0.014891554911931356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,51200,1.2423822614881728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,6144,0.019876443677478366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,512,0.014136888914638095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,512,0.017831999394628737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1024,2560,0.009888889061080085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,2048,0.07350399759080675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3072,2560,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,16384,0.0957404441303677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,2048,0.09141155746248032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,4096,0.029818667305840388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,256,0.014505777094099255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,10240,0.5785528818766276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,8192,0.013772444592581855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,16384,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,10240,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,768,0.00887111077706019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,7168,0.2751377688513862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,6144,0.05798133214314779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,12288,0.5706613328721788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,51200,0.04617599977387322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,8192,0.05508088734414843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,8192,1.5445519553290472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,4096,0.4437146716647678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,3072,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,768,0.016176889340082806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,3584,0.018128888474570382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,128,0.0021706666383478376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,16384,0.043338666359583534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,4096,0.011516444385051727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,5120,6144,0.08835911088519627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,12288,0.1415075593524509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,768,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,2560,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,1024,0.014489778214030795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,5120,1024,0.004543111142185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,1536,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,128,0.0031537777847713898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,1024,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,7168,0.0466195543607076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,16384,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,1024,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,12288,0.02234399980968899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,1024,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,16384,0.16038400597042507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,4096,0.010834666589895884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,768,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,10240,0.031530665026770696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,4096,0.03209422363175286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,12288,0.02198666665289137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,512,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,5120,8192,0.014861333701345654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,65536,0.12519110573662653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,2560,0.0788524415757921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,3072,0.015506666567590503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,2560,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,16384,0.027453333139419556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,6144,0.00925688859489229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,2048,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,6144,0.19532444741990831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,1536,0.00416088890698221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,4096,7168,0.13672178321414524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,8192,0.02699288891421424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,3584,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,7168,6144,0.015636444091796875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,64,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,7168,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,65536,0.8567093213399252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,128,0.031154665682050917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,3072,0.0478168891535865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,128,0.006074666563007567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,512,7168,0.02164622147878011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,51200,0.13338400257958308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,16384,0.01994222237004174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,1024,0.020246222615242004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,5120,0.011209777659840055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,65536,0.05338399940066867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,51200,0.10878488752577041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,16384,3584,0.016562667157914903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,8192,0.02095555596881443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,768,0.017958222164048087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,6144,1024,0.025462221768167283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,6144,0.10708177751964992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,256,0.11430933740403916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,16384,0.030266665750079687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,51200,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,2560,0.015503999259736804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3584,7168,0.03661777906947666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2048,3584,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,7168,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,1024,0.005921777751710679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,16384,0.3003182146284315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,6144,0.059458666377597384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,8192,1536,0.019208000765906442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,7168,0.02678044471475813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,2560,0.016527111331621807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,12288,0.05445955528153313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,512,65536,0.1482951111263699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,10240,0.1931635538736979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2560,3584,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,12288,0.2085768911573622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,12288,0.25105423397488064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,64,0.05758844481574165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,65536,0.0679911110136244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,3072,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,4096,0.024728889266649883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,16384,0.056777775287628174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,5120,0.023065777288542852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,51200,0.11587822437286377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,128,0.010724444356229572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,5120,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,8192,1536,0.021942221456103857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,768,0.0028080000645584534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,10240,512,0.024511999554104272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,512,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,3584,0.02388622197839949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,8192,0.04389422138532003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1024,4096,0.0457911127143436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,7168,0.02681333323319753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,12288,0.05051022105746799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,7168,1536,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,512,768,0.01888799998495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,256,0.0058906666106647914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,51200,0.21532089180416533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,16384,0.1754080057144165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,16384,3.7305518256293406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,6144,0.045816888411839805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,1024,256,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,65536,0.07640444570117526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,51200,0.9434418148464627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,7168,0.10457688570022583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,256,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,10240,0.2958222230275472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,1536,0.005980444451173146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,5120,0.008678221868144141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,768,0.08462488651275635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,4096,3072,0.04362844427426656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,8192,0.013069333301650153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,768,0.15486221843295628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2560,2560,0.14713422457377115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,256,0.022122666239738464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,2048,0.07666133509741889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,768,0.0406275557147132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,7168,0.03874311182234023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,128,0.015200000670221118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,65536,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,32,0.006618666566080517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,7168,0.11453778213924831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,10240,0.11991288926866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,16384,0.023960888385772705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,2560,0.01680266691578759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,768,0.003145777723855443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,1536,0.049441777997546725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1536,1024,0.009289777941173976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,6144,0.2881351047092014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,1024,0.00926399976015091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,1536,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1024,2048,0.011911110745535957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,6144,0.42284001244439023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,512,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,1536,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,12288,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,256,0.02092622220516205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,8192,0.014352889524565803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,64,0.01038577738735411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,5120,0.08008888694975111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,768,0.019555555449591745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,8192,0.10582933160993789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,2560,0.026750221848487854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,256,0.06485955582724677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,3584,0.033352000845803156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,512,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,6144,0.02115999990039402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,51200,0.329111999935574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,3584,0.26166311899820965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,51200,12.132112291124132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,16384,0.30804623497856987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,3072,0.021959111094474792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,128,2560,0.009013333254390294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,3072,0.47870577706231016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,32,2048,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,51200,0.09226133426030476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,1536,0.010024000373151567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,6144,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,64,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,10240,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1536,10240,0.045181334018707275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,2048,0.030647112263573542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,6144,2560,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,65536,0.09202755822075738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,1024,0.003547555456558863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,65536,0.3892844518025716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,768,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,6144,0.22980178727043998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,512,0.1645804378721449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,5120,0.15562933021121556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,128,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,256,0.006191111273235745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,256,0.0677164461877611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,128,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,51200,0.46660354402330184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,2048,0.01924444403913286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,512,0.024663999676704407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,51200,0.04876533481809828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,16384,0.07311911053127713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,5120,0.00721955547730128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,2560,0.0603004429075453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,768,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,16384,0.017335111896197002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,3584,0.011396444506115384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,5120,0.059121780925326876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,3584,0.018908444378111098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3072,128,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,3072,0.010831111007266574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,1024,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,2560,0.019034667147530448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,12288,0.017836444907718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,8192,1.1350044674343533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,6144,0.025145777397685583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2048,12288,0.01367377738157908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,32,0.013754666679435305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,6144,6144,0.014384888940387301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,12288,0.05089777708053589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,2560,0.007263110743628607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,768,0.23958844608730742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,64,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,1536,0.009380444056457942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,5120,0.015524443652894763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,2048,0.02922044528855218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,768,128,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,65536,0.12468088997734918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,8192,0.15722844335767958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,7168,0.017853332890404593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,3072,0.03684799869855245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,4096,0.03232533401913113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,3584,0.11356977621714275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,3072,0.007605333295133378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,4096,0.010967111421955956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,7168,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,8192,0.011018666956159802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,512,0.0032053333189752367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,6144,0.05799733267890083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,8192,2.6877945793999563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2048,10240,0.09362844626108806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,2048,0.10917777485317653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3584,16384,0.2140737771987915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,6144,0.04654133319854736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,1536,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,4096,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,64,16384,0.1046817766295539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,16384,0.018565333551830716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,8192,0.025789333714379206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,4096,0.04519111249181959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,768,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,3072,0.019880889190567862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,2048,0.08855644199583267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,7168,0.010012444522645738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,8192,0.037639111280441284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,64,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,1024,0.028038223584493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,12288,0.20918667316436768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,16384,0.0243520008193122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,6144,0.12867466608683267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,1536,0.0038897775941424896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,4096,0.07897955841488309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,512,0.04256977637608846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,16384,0.03997155692842271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,256,0.020396444532606337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3584,8192,0.05763910876380073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,12288,0.038498666551378034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,512,0.035226666265063815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,512,0.021968000464969214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,1024,0.0032275555034478507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,10240,0.010435555544164447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,8192,32,0.0052168890833854675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,2560,0.03054222133424547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,16384,0.02092444399992625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,2560,0.020576000213623047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,4096,0.4362951119740804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,3584,0.022983110613293115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,2048,0.0654275549782647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,32,0.0020373333245515823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,512,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,12288,0.2671119901869032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,12288,0.1108035577668084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,16384,0.06239466534720527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,1024,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,65536,10240,0.2114595572153727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,1536,0.008727110922336578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,4096,1.5158462524414062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,8192,0.044736888673570424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,2560,0.46534935633341473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,8192,0.011209777659840055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,5120,0.1548373301823934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,16384,0.04545422063933479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,2560,0.02536800007025401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,12288,0.1487671136856079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,1024,0.0320720010333591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,2560,0.08089244365692139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,6144,1024,0.00685422205262714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,32,0.007584888901975419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,51200,0.7356497976515027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,512,0.06313600142796834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,51200,0.05589422252443102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,3584,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,51200,0.1273075607087877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,5120,0.02905511193805271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,16384,0.09609066777759129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,128,0.011067555182509951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,51200,0.21092800299326578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,2560,0.1037635538313124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,1024,0.004429333325889376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3584,768,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,6144,0.04388799932267931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,16384,0.02978488802909851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,768,0.02608088817861345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,3072,0.02750933335887061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,5120,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,8192,0.01070755554570092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,2560,0.1600088940726386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,512,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,8192,0.0609511137008667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,2560,0.0074826669361856245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,8192,0.03571910990609063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,3584,0.021930666433440313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,10240,0.027577777703603108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,16384,0.6678382025824653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,7168,0.025615110993385315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,1536,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,128,0.017631111873520743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,256,0.04421955678198072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,1536,0.010715555813577441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,51200,3584,0.23762843343946669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,4096,0.04296355446179708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,65536,0.07976355817582872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,2560,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,51200,0.04967200093799167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,768,0.01960444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,16384,0.021594666772418555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,3072,0.9332141876220703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,3584,0.026163554853863184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,64,0.0030497778207063675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,3584,0.0070168889231152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,128,0.003548444559176763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,256,0.03443822264671326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,128,0.07323822047975329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,51200,0.8895840115017362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,1024,0.02706577711635166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,51200,1.7273546854654949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,3072,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,3584,0.047071112526787647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,1536,0.016869333055284288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,512,0.0028035555862718155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,5120,0.08292177650663587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,7168,0.009695111049546136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,7168,0.02086044516828325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,8192,0.022312000393867493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,10240,0.016846223009957206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,10240,0.028508444627126057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,3584,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,512,0.03347200155258179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,4096,0.03741866681310866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,12288,0.22613244586520723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,128,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,2560,0.02488533324665493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,32,0.09975644614961411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,16384,0.27656356493632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,256,51200,0.040039112170537315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,32,0.003552888830502828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,512,0.011374221907721626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,65536,0.2602693239847819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,1536,0.03090755475891961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,768,0.00942488925324546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,7168,0.18678311506907144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,7168,1.512288941277398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,1536,0.005271111097600725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,7168,0.01806400054030948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,2560,0.03559289044804043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,65536,0.17303110493554008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,4096,3072,0.007315555380450354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,128,0.005202666752868229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,3072,0.15904800097147623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,32,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,32,0.020428443948427837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,1024,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,10240,0.008540444076061249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,65536,128,0.07039022445678711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,16384,0.14537599351671007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,768,0.024953777591387432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,1536,0.05461066630151537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,8192,0.07524444659550984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2560,16384,0.1449893315633138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,51200,0.49982579549153644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,10240,0.19391377766927084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,6144,0.0328435566690233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,16384,0.053787555959489614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,1024,0.0069360000391801195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,256,2048,0.017456000049908955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,768,0.01220355596807268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,7168,0.013759999639458127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,65536,0.15952711635165745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,6144,0.03222044308980306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,7168,0.024180443750487432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,3072,1.770057784186469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,3584,0.11794133981068929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,1536,0.2984702322218153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,3584,0.03397688931889004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,4096,0.020177778270509508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3072,512,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,3584,0.03672622309790717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,1024,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,1536,0.018375110295083787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,8192,0.06972444719738431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,16384,65536,3.056959999932183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,7168,0.060749332110087075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,5120,0.02287111183007558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,256,0.025480000509156123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,5120,0.01240799989965227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,12288,0.029812445243199665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,51200,0.12277777989705403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,7168,0.010729778144094678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,768,0.004118222329351637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,16384,0.04634666774008009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,5120,0.008945778012275696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,51200,3584,0.06341511011123657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,64,0.16278310616811117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,128,0.012358222570684222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,3072,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,16384,32,0.009039999710188972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,5120,0.010007999837398529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,64,0.007975999679830339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,6144,0.028543111350801256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,6144,0.12011910809410943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,16384,0.016157334049542744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,6144,0.009962666365835402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,3584,0.012770666844315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,2560,0.014152889450391134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,768,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,16384,2.9038043551974826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,12288,0.02255288925435808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,1536,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,512,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2560,2048,0.008440000315507254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,6144,2560,0.026778666509522334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2048,65536,0.06641778018739489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,16384,0.39641510115729434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,4096,0.03865155577659607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,6144,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,5120,0.00924622184700436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,64,0.0025235555238193935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,5120,0.11210666762457953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,12288,5120,0.018681777848137748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,6144,0.010636444720957013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,8192,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,1024,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,16384,0.18428444862365723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,128,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,1024,0.007924444145626491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,7168,0.020654223031467862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,256,0.016189333465364244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,512,3072,0.023522666758961145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,12288,0.27974756558736164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,51200,0.07521066400739881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1536,768,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,256,65536,0.31190488073560924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,128,0.14692266782124838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,1024,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,128,0.014367111855083041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,512,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,2048,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,256,0.0052968888647026485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,256,0.0021404443929592767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,768,0.03417955504523383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,2048,0.02858755654758877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,5120,0.11710133817460801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,51200,0.4430835511949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,256,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,1024,0.010223111344708337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,512,0.014318222800890604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,4096,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,512,0.019395555059115093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,10240,0.014486221803559197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,8192,0.01536355581548479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,12288,0.5353333155314127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,7168,0.08142755428949992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,16384,5.566376156277126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,7168,0.0952755543920729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,51200,4.859164343939887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,1536,0.045800887876086764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,2048,0.02715733316209581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,128,256,0.0028275555620590844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,65536,0.700710243648953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,51200,0.17395910951826307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,16384,0.058711111545562744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,4096,0.06458666589525011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,2560,0.02330044408639272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,2048,0.010786666638321347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,2560,0.07234666744867961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,6144,16384,0.043671111265818276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,16384,0.1879111131032308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,3584,0.10330488946702744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3072,16384,0.09828533066643609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2560,8192,0.016415999995337594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,4096,0.07595555649863349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,256,0.024819556209776137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,12288,0.5861866739061143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,12288,0.017141333884663053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,5120,0.05335288908746508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,64,0.006056889063782162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,256,1024,0.01758044461409251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2048,1024,0.005701333284378052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,128,0.04584711127811008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,10240,2048,0.01075111081202825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,3072,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,128,0.007735999921957652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1536,2048,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,2048,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,2048,0.07894755734337701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,1024,0.04596799943182203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,65536,0.2020639975865682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,1536,0.0038648889296584656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,256,0.002412444394495752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,16384,0.05078311098946465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,6144,0.06896622313393487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,4096,0.029846221208572388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1024,3072,0.02436711059676276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,3072,0.12320621808369954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,4096,0.009018667042255402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,10240,0.015437333120240105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,1536,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,10240,0.9765218098958334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,7168,0.034155554241604276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,2048,0.10029244422912598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,65536,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1536,7168,0.010332444475756751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,12288,0.10698844326866998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,5120,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2560,2560,0.014336888988812765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,6144,0.29845333099365234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,3072,0.08211644490559895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,768,0.01458666721979777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,8192,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,256,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,65536,0.29844533072577584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,51200,0.19827377796173096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,65536,0.02344888945420583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,5120,0.009672000176376766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,1536,0.32757332589891225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,768,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,3072,0.026155556241671245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,4096,0.012226666841242047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,512,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,1536,0.058918224440680615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,65536,0.6375457975599501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,7168,128,0.0021137777302000257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,512,0.013481777575280933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,1024,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,32,0.0441599984963735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,512,0.0031848889258172777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,1024,0.23073156674702963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,32,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,65536,0.07940177785025702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,1024,0.019529778096410964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,5120,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,2560,0.008658666577604082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,1536,0.009844444692134857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,512,0.12395732932620579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3584,256,0.002128888956374592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,3584,0.024995555480321247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,2560,0.009999111294746399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,16384,0.13492622640397814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,128,0.002144888871245914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,4096,0.0178195552693473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,8192,768,0.0273982220225864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,128,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,4096,0.013935999737845527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,5120,0.010377777947319878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,3072,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,256,0.011457777685589261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,1536,0.015881778465376962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,1536,5120,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,12288,0.045744889312320285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,7168,0.010302222437328762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,1024,0.020970667401949566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,3072,0.022016889519161646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,1536,0.01774222155412038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,8192,2.727284537421333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,2560,0.02072266737620036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,51200,0.7356737984551324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,7168,0.06276977724499173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,3584,0.011671110987663269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,6144,0.018632888793945312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,4096,0.015472888946533203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,16384,0.050604446066750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,3072,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,64,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,256,0.005533333453867171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,256,0.01240711079703437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,128,0.06109244293636746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,12288,0.2134417825274997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,51200,0.42192800839742023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,1024,0.006285333385070165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,1536,0.013436444103717804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,4096,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,768,0.00927022182279163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,4096,0.017889777819315594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,128,0.005299555758635203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,512,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,16384,0.03173333406448364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,10240,12288,0.44121244218614364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1024,128,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,65536,0.2979324393802219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,2560,0.020358221398459542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,65536,0.28045333756340873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,51200,0.5391902393764919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,16384,0.03523644473817613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,64,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2048,3072,0.03821422325240241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,3584,0.057095110416412354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,256,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,128,6144,0.0075911109646161394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,65536,0.015294222368134392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,5120,0.17960000038146973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,51200,0.05726310941908094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,65536,2.2173298729790583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,5120,0.06199466519885593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,128,0.005642666584915585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,2048,0.04090755515628391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,32,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,128,0.01703999936580658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,51200,0.09417777591281468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,128,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,7168,0.0358897778722975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,8192,0.03316355413860745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,12288,0.015208888385030957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,768,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,16384,7168,0.03727644350793626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,2560,0.09214755561616685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,10240,0.05663822094599406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,3584,0.058519999186197914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,6144,0.04284800092379252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,2048,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,10240,0.12877599398295084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,8192,0.8149929046630859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,128,0.015963556038008798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,256,3072,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,32,0.01791466606987847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,10240,0.0704666641023424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,1024,0.01794755624400245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,1024,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,6144,0.032813333802753024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,32,0.011750222080283694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,5120,0.014094221923086377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,1024,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,2560,0.014092443717850579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,2560,0.004872888740566042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,5120,0.1278453270594279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3072,4096,0.07104711400138007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,65536,0.07215733660591973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,4096,0.016533333394262526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,1536,0.006979555719428592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,16384,0.02570933269129859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,32,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,512,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,768,0.005619555711746216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,6144,768,0.03298666742112901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,3584,0.07866044176949395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,128,0.011315555208259158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,65536,0.09560622109307183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,12288,0.03730400072203742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,8192,0.06937599844402738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,12288,0.9407510757446289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,2560,0.011982222398122152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,6144,0.02146400014559428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,768,0.02197422252760993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,256,0.00996800015370051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,1024,0.0313351121213701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,2048,0.024711110525661047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,256,0.0041724443435668945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,8192,0.8576506508721246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,16384,0.07427466577953763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,768,1536,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,2560,0.023948444260491267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3072,3584,0.020577778418858845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,128,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,2560,0.013015111287434896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,128,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2048,8192,0.013554666605260638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,32,0.007970666719807519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,3072,0.020262221495310467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,2048,0.004889777965015835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,32,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,16384,0.02293511066171858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,2048,0.019893333315849304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,10240,0.02683288852373759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2560,6144,0.083588441212972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,4096,0.03183822168244256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,51200,0.5014764467875162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,32,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,16384,0.36881332927280003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,51200,1.1574711269802518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,16384,3.517196443345812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,3584,0.05314755439758301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,10240,0.15156355169084337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,2048,0.055996444490220815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,256,0.01483022173245748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,64,0.021320889393488567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,7168,0.08700711197323269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,128,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,3072,0.02793777651256985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,51200,1.1275413301255968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,768,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,51200,0.5558604664272732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3072,4096,0.008952000074916417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,32,2560,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,256,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,768,0.013061333033773633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,7168,0.2067991097768148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2560,51200,0.45651912689208984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,3584,4096,0.06442844205432467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,8192,4096,0.029506666792763606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,128,0.0148044443792767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,51200,0.3039795557657878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,4096,3.291912078857422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,64,0.0031786666562159858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,6144,0.22967110739813912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,10240,768,0.006806222101052602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,5120,0.008707555631796518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,2560,0.048615111245049365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,3584,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,8192,0.0336097776889801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,64,0.0034453332838084963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,3584,0.0491226679748959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,6144,0.15285689300960964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,51200,2560,2.538191901312934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,65536,0.701016002231174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,128,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,8192,0.2831795480516222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,1536,0.020392888122134738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,128,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,64,0.01110666659143236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,65536,0.19297422303093803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,2560,0.017550221747822232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,51200,0.06458133459091187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,65536,0.14103021886613634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,2560,0.022727110319667395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,16384,0.43433600001864964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,8192,0.004602666530344221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,65536,256,0.8409306738111707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,4096,51200,0.055564443270365395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,51200,0.07056088580025567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,768,0.03195110956827799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,10240,0.3944684399498834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,1536,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,2048,0.19540888733334014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,51200,51200,0.4865111245049371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,12288,0.013449778159459433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,51200,0.26600445641411674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,3584,0.03933777742915683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,6144,0.019928000039524503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,12288,0.11529510551028782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,10240,0.011004444625642566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,2048,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,12288,0.9747538036770291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,768,12288,0.014471999473041959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,2560,0.29825954967074925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,128,3072,0.019205333458052743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,7168,10240,0.020255999432669747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,16384,0.07341511381997003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,128,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,7168,0.22002489036983916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,128,0.015056000815497505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,768,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,10240,0.34243912167019314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,512,0.005777777896987067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,1536,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,3584,0.007154666715198093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,64,0.008055110772450766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,256,0.01477866702609592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,512,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,6144,0.021496888664033677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,2560,0.017887110511461895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,65536,0.08450400167041355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,1024,0.02023555503951179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,128,0.04355377621120877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,8192,0.01646577815214793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,8192,3584,0.011171555353535546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,768,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,51200,0.0652408864763048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,6144,0.009292444421185387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,4096,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,3072,0.007371555599901412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,10240,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,8192,0.007622222105662028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2560,12288,0.07304888963699341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,3072,0.09415644407272339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,65536,0.49762757619222003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,6144,0.1484266652001275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,4096,0.04924355612860786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,8192,0.027263111538357202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,10240,0.04786400000254313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,6144,128,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,128,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,65536,1.4751155641343858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,51200,0.0731875565316942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,5120,0.07011555300818549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,768,0.020093333390023973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,3584,0.07308355304929945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,10240,0.04308888978428311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,768,0.006183111005359226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,12288,0.016872000363137987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,16384,0.06292888853285047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,3072,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,256,0.0032559999575217566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,6144,51200,0.06041066514121162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,1536,0.016128889388508268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,10240,0.23109155231051973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,32,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,256,0.009322666459613377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,8192,0.025935999221271936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,128,0.00999644481473499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,2560,0.020218667056825426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,128,0.0051857779423395796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,512,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,65536,768,0.0844871136877272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,8192,0.13971732722388372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,16384,0.1752693388197157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,128,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,8192,0.023352889551056757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,2560,0.012487110992272695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,8192,0.028452446063359577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,7168,1.2934444215562608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,3584,0.011719110939237805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,2048,0.009346666435400644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,8192,0.055030223396089345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,6144,0.0440631111462911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,51200,0.12289244598812527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,5120,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,4096,0.008294222255547842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,128,0.07837866412268744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,12288,0.14299110571543375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,5120,0.029973334736294214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,10240,0.020063999626371596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,1024,0.028138667345046997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,65536,4096,0.5172524452209473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,1024,0.00665155549844106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,5120,0.5418177710639106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,10240,0.056891553931766085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,64,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,7168,0.06458044714397855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,512,51200,0.050607111718919545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,51200,0.13078133265177408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,512,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,10240,0.02028888960679372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,2048,0.015222221612930298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,12288,7168,4.191392686631945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,3584,0.26951556735568577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,768,0.01816444430086348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,8192,0.07284088929494222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,2048,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,10240,0.02676711148685879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,1536,0.023447111248970032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,768,0.02353688908947839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,8192,0.18880621592203775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,768,0.0031866667171319327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,1024,0.015658666690190632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,3072,0.025124443901909724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,10240,0.018232888645595975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,1024,0.013460444079505073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,8192,0.01461155547036065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,4096,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,2048,0.018744000130229525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,10240,0.5497288703918457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,512,6144,0.03166666626930237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,6144,0.04944444364971585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,8192,0.025240000751283433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,768,0.00944088896115621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,256,0.014879999889267815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,128,0.018178666631380718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,12288,0.10198399755689834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,1536,0.008369777765538957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,51200,0.11401955286661784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,1536,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,51200,0.03685955537690057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,512,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,51200,0.12062488661872016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,512,0.02570488883389367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,2048,0.013064000341627332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,128,0.012039110892348819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,8192,0.3984657658471002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,8192,0.024561777710914612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,1536,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,4096,5120,0.09346399704615276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,8192,0.018512000640233357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,512,0.01424711114830441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,1536,0.006986666884687211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,51200,0.052299555804994374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,3584,0.013850666582584381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,32,12288,0.010031999813185798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,128,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,8192,0.03256888853179084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,1024,0.00832622249921163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,1024,0.010489777558379702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,12288,0.03461600012249417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,12288,0.023341332872708637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,768,0.003871110992299186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,6144,0.012529777983824411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,3584,0.03268977668550279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,1024,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,7168,0.010486221975750394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,51200,0.3857777648501926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,8192,0.016753777861595154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,128,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,3584,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,10240,0.06699555450015597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,12288,0.039473778671688504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,768,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,256,0.0042133331298828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,768,0.014693333042992486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,7168,0.05974577532874214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,64,5120,0.009321778184837764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,32,0.010400888820489248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,6144,768,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,4096,0.04760266674889458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,65536,0.20741777949863008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,32,0.007040888898902469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,128,0.08881955676608616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,256,0.014808000789748298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,51200,0.2634826766120063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,1024,0.016493333710564505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,3584,0.014474666780895658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,256,1024,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,7168,0.03775733378198411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,12288,0.027100443840026855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,768,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,256,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,16384,32,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,128,0.005969777703285217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,3584,0.07428977886835735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,768,0.018167111608717177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,128,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,1536,0.31392354435390896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,8192,0.0759813321961297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,4096,0.07441777653164335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,128,0.002193777718477779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,512,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,51200,6144,0.5244142214457194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,4096,0.09801333480411106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,2560,0.010480889015727572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,768,0.004908444566859139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,3584,0.021500445074505273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,256,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,10240,0.04336799846755134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,1536,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,2560,0.022191999687088862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,128,0.021778666310840186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,10240,128,0.02404977712366316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,2560,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,128,0.009003555609120263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,65536,1.6198968887329102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,768,0.004552000098758274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,2560,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,12288,0.02446311049991184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,3072,0.00904088881280687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,12288,0.041639112763934664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,3072,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,3072,0.014986667368147107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,2560,0.06530933247672187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,1024,0.017344888713624742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,4096,0.043776889642079674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,32,0.008545777863926357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,3584,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,12288,0.082933333184984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,512,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,8192,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,32,0.01428533262676663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,8192,0.04958666695488823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,512,1024,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,6144,768,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,3072,0.012434666355450949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,3072,3072,0.06855733527077569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,12288,0.8797458012898763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,8192,0.28010310067070854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,12288,0.1508302158779568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,128,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,5120,0.025052444802390203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,6144,0.09943999846776326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,8192,0.0404151115152571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,1536,0.03056266572740343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,6144,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,65536,0.1520355542500814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,4096,0.009090666969617208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,3072,0.012290666500727335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,1536,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,7168,0.010427555276287926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,7168,0.045569777488708496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,2048,0.004854222138722737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,65536,0.06420799758699205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,2048,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,2560,0.07943644126256307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,3072,0.018840889135996502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,2560,0.06377600298987494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,6144,0.10396177901162042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,1536,0.026068444053332012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,2560,0.06391200092103747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,256,1024,0.008101333346631791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,256,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,10240,65536,0.13462932904561362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,1536,0.007081777685218387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1024,3072,0.016415999995337594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,4096,0.0417848891682095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,8192,0.022316444251272414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,32,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,256,0.007239110767841339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,768,0.017272000511487324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,3072,0.01696888936890496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,256,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,1536,0.01647911138004727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,12288,0.016872000363137987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,32,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,2560,0.004583111239804162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,768,0.0031511110977994073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,7168,0.014577777849303352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,16384,0.03939022289382087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,32,0.00267999991774559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,2048,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,1024,0.014789332946141561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,3072,0.09573600027296279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,3072,0.03481244378619724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,8192,0.07413599888483684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,2048,0.02980088856485155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,64,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,51200,1.6972968843248155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,3584,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,2560,0.0463031099902259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,6144,0.028708444701300725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,512,0.017179555363125272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,7168,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,6144,0.015498666299713982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,64,0.0032213332338465583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,4096,1536,0.006260444306664997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,256,0.010999999940395355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,12288,0.014315555493036905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,2048,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,4096,3072,0.010400888820489248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,65536,0.006999111009968652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,3072,0.04006933172543844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,512,1024,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,8192,0.06774310933219062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,6144,0.30514221721225315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,768,0.021740444832377966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,768,0.020317332612143624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,4096,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,6144,3584,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,10240,2560,0.014380445082982382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,5120,7168,0.028258668051825628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,512,0.0027866665687825945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,2560,0.0125928885406918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,128,0.0052177777720822235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,51200,0.2756933371225993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,128,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,5120,0.09834400150511001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,128,0.017231111725171406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,16384,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,256,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,3072,0.05753778086768257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,6144,0.2727582189771864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,10240,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,16384,0.17238222228156197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,128,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,64,0.008082666330867344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,7168,0.01779111060831282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2048,16384,0.03781777620315552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,4096,0.039992888768514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,4096,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,1024,0.003216000066863166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,2560,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,65536,1536,0.08278133471806844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,8192,0.35473065906100804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,2560,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,32,0.0021182222084866627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,2560,0.1255964438120524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,128,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,5120,1536,0.019261333677503798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,1536,0.02456888887617323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,64,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,6144,0.037964443365732826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,8192,0.03431111243036058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,2048,0.023640000157886084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,4096,0.1224106682671441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,768,0.003945777813593547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,128,0.024703111913469102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,5120,8192,0.07550577984915839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,7168,0.01789955629242791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,16384,0.025620443953408137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,6144,0.15704445044199625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,512,0.016487111647923786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,768,0.005151111218664381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,7168,0.009723555710580613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,10240,0.049118220806121826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,2048,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,3072,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,4096,128,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,128,0.010732444624106089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,4096,0.01405244403415256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,3072,0.08743555678261651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,256,0.01942044496536255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,16384,0.042818665504455566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,3584,0.38584266768561465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,256,12288,0.02782400117980109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,16384,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,10240,0.014731556177139282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,5120,0.011023999916182624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,12288,0.10763111379411484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,3584,0.013739555246300168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,3072,0.007333333293596904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,64,1536,0.0034959999223550162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,7168,0.042633778519100614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,16384,0.03781955440839132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,1024,0.02302933401531643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,1536,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,768,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,2048,0.017832888497246634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,51200,0.15533867147233751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,6144,0.007141333487298753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,6144,0.08210933208465576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,12288,0.032355556885401406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,16384,0.05110488997565376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,51200,65536,0.6330889066060384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,8192,0.03167377909024557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,51200,0.06321689155366686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,3584,0.07678666379716662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,2048,0.028014222780863445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,2048,0.03289866778585646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,8192,3584,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,12288,0.0665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,512,0.027848889430363972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,1536,0.07263644536336263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,3072,0.5309333271450466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,65536,0.6571146647135416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,768,0.0035439998739295532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,10240,0.032764444748560585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,65536,0.1572293308046129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,7168,0.40486576822068954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,65536,0.8576178020901151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,8192,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,256,12288,0.023471110396915015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,7168,0.019877332780096266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,65536,1.894424862331814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,10240,0.22159822781880698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,32,0.012712889247470431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,6144,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,16384,0.027060444156328838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,128,0.006268444574541516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,128,0.03306844499376085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,4096,0.021213332811991375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,65536,0.08129333125220405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,16384,0.015802666544914246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2048,51200,0.0681697792477078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,256,0.004214222232500712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,16384,3072,0.014803555276658801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,768,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,8192,0.004912888838185204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,3072,0.08649688959121704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,51200,0.3499644332461887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,128,0.005277333160241445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,64,0.005974222388532426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,8192,0.05190933412975735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,3072,0.038314667012956403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,2048,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,3072,0.29281510247124565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,1024,0.017843556072976854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,10240,0.03896444373660617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,4096,0.00830577810605367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,3584,0.016813332835833233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,6144,0.028775999943415325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,768,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,1024,0.052376889520221286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,8192,0.026260443859630164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,1536,0.02091377807988061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,65536,1.0081404580010307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,8192,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,1536,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,768,0.01919911139541202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,10240,0.04092266824510362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,1024,0.02752888864941067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,7168,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,3072,0.024548444483015273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,1536,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,64,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,128,0.005195555587609609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,6144,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,128,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,1536,0.005481777919663324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,16384,0.2135617733001709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,2048,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,256,4096,0.033131556378470525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,5120,10240,0.025968000292778015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,256,0.00902133352226681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,1536,0.008253333469231924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,256,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,51200,10.26478746202257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,7168,0.020098666350046795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,768,0.018245332770877414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,128,0.022712000542216833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,6144,0.07779822084638807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,16384,64,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2048,1024,0.037263999382654824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1024,64,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,1024,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,768,0.023518222901556227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,256,1536,0.004552888787455029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,256,0.020230222079488967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,128,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,51200,0.5608622233072916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,5120,0.07945866717232598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,32,0.0021893332401911416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,128,0.04669422242376539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,2560,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,1024,0.0061919999619325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,6144,0.24708710776435006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,16384,0.016136889656384785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,64,0.001814222170246972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,7168,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,768,3072,0.03754399882422553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,12288,0.0665102203687032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,2048,0.007484444313579136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,2560,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,1536,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,12288,0.08299466636445788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,4096,0.068122665087382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,128,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,65536,0.8989102045694987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3072,512,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,7168,0.016743110285864938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1536,4096,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,256,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,3072,0.012330667012267642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,6144,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,7168,0.12560444407992893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,6144,0.5393039915296767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,3584,0.07207822137408786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,64,0.0053022222386466126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,7168,0.03173155585924784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,4096,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,3072,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,128,0.015455111861228943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,3072,0.08913866678873698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,12288,0.017484444710943434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,8192,0.15561332967546251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,3072,0.009887999958462184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,2560,65536,0.07034311029646131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,6144,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,256,0.12367822064293756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,16384,0.03355288836691115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,512,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,64,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,3584,0.06288711229960124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,64,2560,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,12288,0.032050665881898664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,6144,0.04996533195177714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,10240,0.028472890456517536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,6144,0.13003377119700113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,1024,0.013902222116788229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,64,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,2560,8192,0.011367999845080905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,51200,0.04532622297604879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,10240,0.022971555590629578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,512,0.007154666715198093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,65536,0.8460737864176432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,65536,0.10906755261951023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,12288,0.03363555669784546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,16384,0.05662755833731758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,1536,0.012083555261294046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,16384,0.018582221534517076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,3072,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3584,4096,0.008632889224423302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,51200,0.3482480049133301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,256,0.025705777936511572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,65536,12288,0.7439617580837674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,12288,0.030311998393800523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,32,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,768,0.005945777727497949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,2560,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,64,0.011009777585665384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,8192,10240,0.2411608960893419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,5120,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,2048,0.006150222072998683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,4096,0.06626222531000774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,1024,0.01333688861793942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,4096,0.022152889106008742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,2560,0.02715377840730879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,65536,2560,0.11481332778930664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,4096,2560,0.007124444676770105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,1536,0.007653333246707916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,2048,0.013914666242069669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,16384,0.0361697773138682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,512,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,16384,0.0498888889948527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,2048,0.011970666547616323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,4096,0.07168355253007677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,32,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,65536,0.04403466648525662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1024,12288,0.13011022408803305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,6144,0.04794133371777005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,4096,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,3584,0.021847999758190576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,3072,0.034613334470325045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,5120,0.017806222041447956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,3584,0.10102578004201253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,128,0.027365333504146997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,8192,0.02643377747800615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,6144,0.020050666398472257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,5120,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,1536,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,1024,0.013117333253224691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,128,7168,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,65536,0.0731840001212226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,3072,0.019091554813914828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,512,0.0033191111352708605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,1024,0.0075128889746136135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,2560,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,768,0.023944000403086346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,64,0.007638221813572778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,256,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,8192,768,0.0041697778635554844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,768,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,128,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,12288,0.014209777116775513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,3072,0.015584889385435315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,1536,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,16384,512,0.0042071110672420925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,4096,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,7168,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,1024,0.09617333279715644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,512,1024,0.003167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,51200,10240,0.1727208826276991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,6144,8192,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,3584,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2048,8192,0.10894311136669582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,7168,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,7168,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,7168,0.20440533426072863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,128,0.0053137776752312975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,64,5120,0.024344000551435683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,16384,0.07765333520041572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3584,10240,0.01774666706720988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,4096,0.02032355633046892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,6144,0.6149511337280273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,2048,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,1536,0.11688178115420872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,2048,0.4560124609205458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,128,0.024365334047211543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,512,0.030743112166722614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,6144,0.020749333831999037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,512,0.003177777760558658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,256,0.026714666022194758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,32,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,768,0.0269768883784612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,7168,768,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,2560,0.02346311178472307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,128,0.012063110868136087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,12288,0.028804444604449805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,16384,0.05330400003327263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,2560,0.04551644457711113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,5120,0.10398577981524998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,4096,0.10885333352618748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,768,0.01680799987581041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,10240,0.09710755613115098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,51200,1.0389315287272136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,12288,0.023808000816239253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,128,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,2048,0.004591111093759537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,8192,256,0.005978666659858491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,8192,1.4950568940904405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,65536,0.07768622371885511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,10240,12288,0.052879111634360425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,3072,0.06066222323311699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,1024,0.007890666524569193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,51200,0.12974577479892307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,512,4096,0.02057422200838725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,4096,0.009988444546858469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,16384,0.02158222264713711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2048,8192,0.013742222554153867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,7168,1536,0.054808888170454234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,512,0.15955021646287706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,12288,768,0.04493066668510437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,2560,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,65536,0.14254755443996853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,768,0.021672000487645466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3584,7168,0.0795795586374071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,2560,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,32,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,65536,1536,1.683484395345052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,12288,0.02667733364635044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,128,0.02329866588115692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,1024,0.007268444531493717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,128,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,5120,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,512,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,4096,0.0967715581258138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,8192,0.01130311108297772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,3072,0.05607733461591932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,1024,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,10240,0.01737155516942342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,6144,0.018231110440360177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,10240,0.12160178025563557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,6144,0.019666666785875957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,1536,0.018180444836616516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,3072,0.020169778002632987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,12288,0.16269066598680285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,3072,0.010080888867378235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,6144,0.010032888915803697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,1536,8192,0.00777333312564426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,32,3072,0.10075110859341091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,51200,0.252804438273112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1536,16384,0.13522756099700928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,7168,0.07004177570343018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,4096,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,256,0.002530666689078013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,128,0.010649777948856354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,64,0.0024702221983008915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,512,0.013722666435771518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,12288,1.6290541754828558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,16384,0.027639110883076985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,1536,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,8192,1536,0.055467555920283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,16384,0.14503199524349636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,5120,0.03091644578509861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,51200,0.05746933486726549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,12288,64,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,16384,0.045680890480677284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,512,0.00926755534278022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,1536,0.015087111128701104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,2560,0.08375733428531223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,3584,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,6144,4096,0.024344889654053584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,128,0.013423999978436364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,2048,0.013059555656380124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,1024,0.005525333185990651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,65536,0.03980177640914917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,128,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,128,0.0079297779334916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,3584,0.15669777658250597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,12288,0.0373653339015113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,6144,0.11430933740403916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,3072,0.011407111254003314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,16384,0.0369395547442966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,6144,2048,0.01889155473974016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,1536,0.06127644247478909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,7168,0.013531555732091268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,7168,0.00981422182586458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,4096,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,768,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3072,6144,0.0633902218606737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,7168,0.02870933214823405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,6144,0.022099556194411382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,10240,0.1782462199529012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,512,0.015591111448076038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,3072,0.09075822432835896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,4096,0.5287777582804362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,4096,0.04745689034461975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,2560,0.02234577801492479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,128,1024,0.003535110917356279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,1536,0.019949333535300363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,2048,0.5502310858832465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,6144,0.017430222696728177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,5120,0.01811822255452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,12288,0.04413155714670817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,2560,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,16384,0.021385777327749465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,51200,0.06423999865849812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,65536,4.075499640570746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,51200,0.21721066368950737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,256,0.014650666051440768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,3584,0.013215111361609565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,128,0.009640889035330879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,3584,0.7574319839477539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,12288,0.04484711090723673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,5120,0.13745510578155518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,768,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,51200,0.06579733557171293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,10240,0.032834665642844305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,4096,0.024539555112520855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,4096,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,8192,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,2048,0.004204444587230682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,1536,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,16384,0.027496889233589172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,4096,0.07614488734139337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,128,0.0101724440852801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,32,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,128,0.1600382195578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3072,3584,0.031903998719321355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,7168,0.017649778061442904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,2560,0.01719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,3584,0.015849777393870883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,768,2048,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,2048,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,4096,0.004518222063779831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,64,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,5120,0.00720266666677263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,768,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,128,0.005990222096443176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,12288,0.021309332715140447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,2560,0.06165688567691379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,10240,0.03169599837727017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,2048,0.004570666700601578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,128,0.009352000223265754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,3072,0.011321778098742167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,51200,1.7263431549072266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,7168,0.10249511400858562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,1024,0.0226248883538776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,768,0.01610311037964291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,51200,0.25872000058492023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,768,0.01129066695769628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,256,12288,0.010797333386209277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,51200,0.22557600339253744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,7168,0.014974221587181091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,512,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,2048,0.023779556155204773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,3584,0.01187999960449007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,6144,0.034959110948774554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,4096,0.015228443675571017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,512,0.007246221933099959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,65536,0.992826673719618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,512,0.005647999958859549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,2048,0.004514666481150521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,128,32,0.0021804444905784395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,12288,768,0.0073902226156658595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,768,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,2560,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,1024,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,16384,0.23786399099561903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,10240,0.05906399753358629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,32,0.006518222391605377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,128,0.003128888913326793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,6144,0.009638222555319468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,5120,0.022727999422285292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,12288,2.895680957370334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,7168,0.05318222112125821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,4096,0.01349511080318027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,128,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,512,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,1024,0.005937777873542573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,8192,0.011062222222487131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,128,0.058535999721950956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,7168,0.011114666859308878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,1536,0.025693333811230127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,65536,0.07620622052086724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,64,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,3072,0.013562666873137156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,4096,0.022775999373859827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,2560,0.06890666484832764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,3584,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,2048,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,256,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,7168,0.009711110757456886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,128,0.0069679998689227635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,6144,0.004866666677925321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,8192,0.01239999963177575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,5120,0.03272622161441379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,64,0.0058497778243488735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,7168,0.2860106627146403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,12288,0.18738667170206705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,10240,0.09619110822677612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,3584,0.01882222294807434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,7168,2560,0.02481777800454034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,51200,0.05310577816433377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,2048,0.010650667051474253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,512,0.04008622301949395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,5120,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,256,0.005285333428117964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,256,4096,0.02382044494152069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,6144,12288,0.13900266753302679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,1024,768,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2048,512,0.013408000270525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2048,5120,0.018460444278187223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,512,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,5120,0.23819732666015625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,4096,6144,0.03456799851523505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,65536,4.854804568820529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,256,0.018034666776657104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,7168,0.6533342467414008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,768,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,6144,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,256,0.027407111393080816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,512,0.022503111097547743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,1024,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,12288,0.06935911046134101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,256,0.023415111833148535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,2048,0.03312622176276313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,3072,0.011926222178671094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3072,51200,3.216397391425239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,8192,12288,0.027109333210521277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,128,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,16384,0.01886488828394148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,2048,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3584,51200,0.31243822309705943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,768,0.006119110931952794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,1024,0.028780443800820246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,65536,7168,0.7414186795552572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,3584,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,3072,0.006741333338949416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,1536,0.0580879979663425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,51200,0.1137804455227322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,16384,2.4162347581651473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,6144,0.02163733376397027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,8192,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,65536,0.63173156314426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,3072,1.0225742128160265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,32,0.01720711092154185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,10240,0.026613333159022864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,1536,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,3072,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,10240,0.02403555479314592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,256,0.15052266915639242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,16384,1.3656702041625977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,16384,0.15614222155676946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,128,0.007280000381999546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,12288,0.030091553926467896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,16384,0.0478168891535865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,3072,0.048042668236626514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,256,0.009730666875839233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,2048,0.10007111231486003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,1536,0.0038319999972979226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,7168,0.15876800484127468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,512,0.04743999905056424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,7168,12288,0.20192000601026747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,128,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,5120,1.3317688835991752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,512,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,1536,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,10240,0.004878222114510006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,512,0.014152000347773233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,12288,0.04025333457522922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,2560,0.011701333026091257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,51200,0.09614400068918864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,768,0.0070426662762959795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,4096,0.019183999962276883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,512,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,32,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,2560,0.03261955579121908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,256,0.002504000026318762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,8192,0.015392000476519266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,2048,0.03074666526582506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,10240,0.012080888781282636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,10240,2.1601058112250433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,5120,128,0.019919999771647982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,768,0.017940445078743827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,1024,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,64,16384,0.07871911260816786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,51200,0.12257956133948432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,32,0.036702222294277616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,10240,12288,0.029750221305423315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,1536,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,12288,0.02991822361946106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,256,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,2560,0.06702666812472872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2048,7168,0.025028443998760645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,6144,0.09605689181221856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,128,0.02751111156410641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,10240,0.027460444304678176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,5120,0.00776977754301495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,6144,0.05420088768005371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,32,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,12288,0.2107635604010688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,7168,0.02149155570401086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,256,0.014858666393491956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,1536,0.4317973454793294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,51200,0.24964798821343315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,64,0.002514666567246119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,65536,0.13656889067755804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,8192,0.4618799951341417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,512,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,16384,0.34604443444146055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,512,0.03087200058831109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,1024,0.2161324421564738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,65536,0.11910577615102132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,64,0.07963022258546618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,3072,0.19623288843366835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,64,12288,0.0628711117638482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,2560,0.01072088877360026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,512,0.1705920033984714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,65536,1.4329475826687281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,5120,0.0396453340848287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,256,0.002505777817633417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,16384,0.04047466648949517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,16384,0.056624001926845975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,1536,0.016559999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,2048,0.018923555811246235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,512,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,5120,0.045309331681993276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,3584,0.09429422352049087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,64,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,65536,0.37359200583563906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,4096,0.06806133190790813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,4096,0.030386666456858318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,8192,51200,1.479045338100857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,4096,0.010728889041476779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,2560,0.017272000511487324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,4096,0.009146666361225976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,768,0.026360000173250835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,5120,0.023716444770495098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,6144,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,7168,0.007096889118353526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,128,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,1536,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,3584,0.013785777820481194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,10240,0.026125333375400964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,65536,0.2977280086941189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,2560,0.058449778291914195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,64,0.014214222629865011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,51200,0.1251573297712538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,12288,0.10247644450929429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,3584,0.03864444295565287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1536,6144,0.048581331968307495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,2560,0.04185511006249321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,7168,0.12120621734195286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,768,7168,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,2560,12288,0.04606133368280199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,128,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,12288,128,0.015159999330838522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,4096,0.020269332660569083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,512,0.0033813334173626373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,7168,0.13591555754343668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,1536,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,51200,0.05357066790262858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,6144,0.04919999837875366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,16384,0.1634248892466227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,3584,0.02479022244612376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,512,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,51200,0.05919022030300564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2560,12288,0.0192577772670322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,8192,0.0226275556617313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,5120,0.05221066541141934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,768,0.012025777664449481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,4096,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,128,0.005537777725193236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,51200,16.481618245442707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,256,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,10240,0.016199111938476562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,4096,0.060257779227362744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,10240,2.0138782925075955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,512,512,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,5120,0.7680693202548556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,7168,0.013035555680592855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,1024,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,2560,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,256,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,12288,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,51200,0.3874488936530219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2560,256,0.019256000717480976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,1024,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,1024,0.0029146667155954572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,3584,0.06295022037294176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,10240,0.09261155790752834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,4096,0.03269066744380527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,6144,0.034895999564064875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,1024,0.04338488976160685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,51200,1024,0.22770489586724174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,3584,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,256,0.08626933230294122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,12288,0.01943822205066681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,2048,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,7168,256,0.011756444142924415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,3072,0.016749334004190233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2048,3584,0.00795733349190818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,4096,0.026289777623282537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,12288,0.04218577676349216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,65536,0.6027599970499674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,2048,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,16384,0.16884266005622017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,256,0.005664000080691443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1024,512,0.0028648889727062653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,256,0.009507555928495195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,1536,0.003887999802827835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,7168,0.10995733075671726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,4096,0.34480267100863987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,64,0.01478844384352366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,32,0.004252444538805219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,768,0.014831110835075378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,2560,0.011686222420798408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,512,0.00655733338660664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,12288,0.7933502197265625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,1024,0.00945955514907837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,256,256,0.012436444560686747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,3584,0.046943998999065824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,3072,0.03566044569015503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,256,0.020063999626371596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,512,0.022315555148654517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,768,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,3584,0.027036445008383855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,3584,0.025519111090236243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,5120,3584,0.07937244574228923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,7168,0.14095821645524767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,5120,0.00906755526860555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,1024,0.01827022267712487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,128,0.005617777920431561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,3072,0.34829065534803605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,128,0.006254222244024277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,2560,0.5393679936726888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,2560,0.011008888483047485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,5120,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,8192,0.01129155523247189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,4096,0.08895022339291042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,10240,0.08554933468500774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,5120,0.7660746574401855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,16384,0.020623111062579684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,768,0.003101333354910215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,5120,10240,0.04380177789264255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,5120,0.020575111111005146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,4096,0.018251554833518136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,6144,0.03908711009555393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,32,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,16384,128,0.06255111429426405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,51200,51200,1.5034942626953125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,512,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,8192,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,64,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,16384,0.014541332920392355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,5120,0.07298044363657634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,256,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,65536,0.2548453278011746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,768,0.00318755561278926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,51200,0.34282578362358945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,3072,0.04993244343333774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,7168,0.017044444878896076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,3584,0.21964177820417616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,12288,0.07256622446907891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,2048,0.023006222314304773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2048,3584,0.03463022245301141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,12288,0.11513688829210068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,65536,0.7475573221842448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,6144,0.09946133030785455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,65536,3.121249728732639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,128,0.17137955294715035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,256,0.008193777667151557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,7168,0.01313688854376475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,65536,0.07088977760738797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,6144,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,1024,0.011694221860832639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,6144,0.040398223532570735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,768,512,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,10240,0.016544888416926067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,3072,0.05276977684762743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,5120,0.02250755495495266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,2560,0.01589866644806332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2560,256,0.012763555679056378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,768,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,512,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,65536,0.10193333360883926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,768,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,256,0.027098667290475633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,128,0.012523555921183692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,5120,0.35203199916415745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,7168,0.02868000004026625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,64,0.0018613334331247543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,51200,0.10435822274949814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,512,0.10490310854381985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,3072,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3584,64,0.001977777729431788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,256,0.02365422248840332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,2048,0.028592000404993694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,512,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,7168,0.009837333526876237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,8192,51200,0.5319857597351074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,1024,0.007162666983074612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,65536,0.19900533888075086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,10240,0.11067289113998413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,16384,0.022464000516467627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,16384,8192,0.024534222152498033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,5120,0.08941066927380031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,1024,0.00868800034125646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,768,0.030827555391523574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,1536,0.0035004446076022256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1536,1024,0.003600888782077365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,128,51200,0.05711555480957031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,3584,0.020247111717859905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,4096,0.017857778403494094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,4096,0.033709334002600774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,256,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,12288,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,7168,0.0444160004456838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,128,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,2048,0.012746666868527731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,256,0.0021226666867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,768,0.01587022178702884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,16384,0.01681866745154063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,65536,0.2643582291073269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,768,0.01670222149954902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,1536,1536,0.017183999220530193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,3072,0.043024000194337636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,7168,0.04176444477505154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,65536,3584,0.08993155426449245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,10240,32,0.01770755483044518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,3072,0.01553244392077128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,2560,0.054309331708484225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,512,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,4096,0.01688088807794783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,6144,0.32925600475735134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,10240,5120,0.03751111030578613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,10240,0.025528889563348558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,5120,0.40529243151346844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,5120,0.010001777774757808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,6144,0.019911999503771465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,128,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,7168,0.011404444773991903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2048,1024,0.011749332977665795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,12288,0.017684444785118103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,512,0.00599644457300504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,6144,0.061117331186930336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,65536,0.04732444551255968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,16384,0.11477688948313396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,512,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,7168,0.016907556189431083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,128,0.015784000356992085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,512,0.1489866707060072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,512,8192,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,7168,0.012786666552225748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,128,0.0027964444210131965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,2560,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,8192,64,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,4096,512,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,2560,0.01706044375896454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,16384,0.022389334109094407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1536,1536,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,8192,0.145215113957723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,12288,51200,0.840277353922526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,1536,0.020690666304694284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,128,0.00553688903649648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,16384,0.04348533352216085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,16384,0.09758222103118896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,4096,0.013051555388503604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,10240,0.032812443044450544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,2560,0.0226239992512597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,6144,0.09720000293519761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,64,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,5120,0.012063110868136087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,16384,0.42685778935750324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,8192,0.0144213338692983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,12288,3072,0.013739555246300168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,65536,2048,0.022984888818528917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,51200,0.0759235554271274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,65536,0.07768444220225017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,2048,0.034508443540996976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,65536,0.17440356148613823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,128,0.0277137772904502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,1536,0.003847111016511917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,2560,0.007592889169851939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,128,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,4096,0.011373333632946014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,6144,0.015610666738616096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,64,0.010564444793595208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,7168,0.020957332518365648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,10240,0.021547555923461914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,512,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,128,51200,0.06179022126727634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,51200,0.009575111170609793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,12288,0.03663466705216302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,4096,0.04377777708901299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,1024,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,768,0.006063111126422882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,7168,0.03406577640109592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,3072,0.03945777813593546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,3584,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,8192,0.104567997985416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,65536,1.107357343037923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,1024,0.01349066694577535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,6144,0.01686755485004849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,7168,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,10240,0.013554666605260638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,512,0.002862222285734283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,7168,0.31450043784247506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,1536,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,1024,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,256,0.011745777395036487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,3584,0.0068195557428730865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,3584,0.9411670896742078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,16384,0.09632088740666707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,10240,0.014377777775128683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,128,0.0026942222482628296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,4096,0.03887110948562622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,4096,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,512,65536,0.2371759944491916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,32,0.005647999958859549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,256,0.002218666589922375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,768,0.004023999803596073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,16384,16384,0.2194462219874064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,1536,0.006302222195598815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,6144,0.021290666527218286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,256,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,1536,0.008977777428097194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3072,51200,0.04932889011171129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,6144,0.13553510771857366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,6144,0.14867644839816624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,10240,0.012372444073359171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3072,128,0.04189688960711161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,768,0.009014222357008193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,10240,0.004888888862397936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,65536,1536,0.023360000716315374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,6144,1.1285386615329318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,128,0.002834666727317704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,12288,0.05647822221120199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,512,0.024303999212053087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,7168,0.020766221814685397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2048,5120,0.02421866688463423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,4096,0.09981777932908799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,768,0.015407999356587728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,6144,0.009349333743254343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,2560,0.006600889066855113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,7168,0.030255999830034044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,256,0.0053075556125905775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,7168,0.12903200255499944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,32,5120,0.0053102220926019884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,1024,0.00360000009338061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,256,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,10240,0.034536888202031456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,7168,0.30962133407592773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,2560,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,256,0.0148026661740409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,16384,3584,0.016295111841625638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,256,256,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,768,0.00813155538505978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,8192,0.048899554544024997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,2048,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,4096,0.07095022333992852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,128,1024,0.019447111421161227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,12288,0.07574844360351562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,768,1536,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,256,0.009680888719028896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,10240,0.11276088820563422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,128,0.014860444598727755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,768,0.005979555762476391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,256,4096,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,32,0.007325333025720384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,16384,0.04516622092988756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,5120,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,512,0.015627556376987033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,2560,0.016322667400042217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,768,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,768,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,6144,1.8974649641248915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,3584,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,3072,0.007544889218277401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,32,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,3072,0.01830311119556427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,256,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,7168,0.007043555378913879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,1536,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,64,0.0036959999965296853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,2048,0.04018488857481215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,1024,0.027845333019892376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,7168,16384,0.07341600126690336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,3584,0.022865777214368183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,3584,0.008281778130266402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,2560,0.006497777584526274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,12288,2048,0.028756442997190688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,16384,0.0185244447655148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,32,0.002463111033042272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,2560,0.018943111101786297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,6144,0.05335288908746508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,1536,0.027461333407296076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,4096,6144,0.05012889040840996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,51200,0.06365778048833211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,16384,0.016833777228991192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,6144,0.0324248903327518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,2560,0.012527111503813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,16384,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,256,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,16384,0.007999111380841997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,4096,0.007947555846638149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,6144,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,12288,0.9201173782348633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,7168,0.020247111717859905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,8192,0.05298933386802673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,12288,0.02071555621094174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,8192,0.0587315559387207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,768,0.015471110741297403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,3584,0.029194666279686823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,51200,16384,0.22436711523267958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,128,0.005366222312053044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,256,7168,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,65536,1.2998488744099934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,16384,0.07322577635447185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,2048,0.031530665026770696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,1024,0.015599111715952555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,5120,0.017619555195172627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,512,0.014430221584108142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,7168,0.01754044493039449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,256,5120,0.016270221935378183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,2048,0.30250755945841473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,6144,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,8192,65536,1.9875111050075953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,512,0.0024764444679021835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,256,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,16384,1.4353182050916884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,3584,0.05292444427808126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,8192,0.2639831172095405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,12288,0.0745511121220059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,10240,0.013743110828929476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,128,0.06792266501320733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,2560,0.056982219219207764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,768,0.019145778483814664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,1536,0.009992000129487779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,3584,0.06337244643105401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,3072,0.02232088810867733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,2560,0.006559111177921295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,2560,0.02752711044417487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,7168,0.019554666346973844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,4096,0.10043466753429836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,4096,512,0.005631999837027655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,1024,0.007261333366235097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,8192,0.24214222696092394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,6144,0.17569511466556123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,3584,0.08168888754314847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,3584,0.15937333636813694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,2048,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,65536,51200,1.019338713751899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,512,0.09180000093248154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,3584,0.03589866558710734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,10240,0.02602133320437537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,7168,0.012816000315878125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,6144,0.10956533087624444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,2048,0.025451555848121643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,8192,0.15844533178541395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,16384,0.17946133348676893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,51200,49.740749782986114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,6144,0.024657777614063684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,2560,0.02525866693920559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,51200,0.04064800010787116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,3072,0.020021332634819877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,2560,0.011402666568756104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,10240,0.042064001162846885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,3584,0.008997333546479544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,768,2560,0.03759466608365377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,8192,0.010953777366214328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,8192,0.037965334124035306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,65536,1.286054187350803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,51200,0.20409866174062094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2560,3584,0.021945777866575453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,2560,0.055255999167760216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,10240,0.03986310958862305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,3072,0.007264888948864407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,5120,0.02402755618095398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,768,0.017843556072976854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,512,0.06369511286417644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,16384,0.06417510906855266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,16384,0.1036284433470832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,128,0.014163555370436775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,128,16384,0.04403733213742574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,32,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,3584,0.0788631108072069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,4096,0.2713529003991021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,2048,0.005532444351249271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,65536,0.06687288814120822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,512,0.009499555660618676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,12288,0.05494755506515503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3584,10240,0.05863733424080742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,768,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,7168,0.02715911136733161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,4096,0.01126844435930252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,128,16384,0.04851466748449537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,10240,0.015676443775494892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,12288,0.034838222795062594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,6144,1536,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,12288,0.06621155473921034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,6144,0.04810844527350532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,32,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,256,12288,0.02571288910177019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,16384,0.06788800160090129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,5120,0.06671910815768771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,10240,0.05172444383303324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,3072,0.013063111239009433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,1024,0.014152889450391134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,2560,0.023149333066410486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,7168,0.053005334403779775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,4096,0.00888888869020674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,7168,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,512,0.014124444789356656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,8192,0.018704889549149405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,10240,0.02738044493728214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,512,0.013517333401574029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,10240,0.015990222493807476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,768,0.007715555528799693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,16384,0.0897173351711697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,768,3072,0.021632888250880774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,8192,0.05970044268502129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,2560,0.03416533271471659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,7168,0.03029688861634996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,2048,0.01015466699997584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,3072,0.03144355614980062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,3584,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,768,0.005674666828579373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,3584,0.02160177793767717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,5120,0.02106933295726776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,1536,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1536,1536,0.03366133239534166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,65536,1.5127538045247395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,3584,0.020589333441522386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,65536,0.31123020913865834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,7168,0.07844977908664279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,3584,0.008719999757077958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,256,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,10240,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,51200,256,0.021566222111384075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,64,0.009292444421185387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,2560,0.05533688929345873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,2048,0.20946755674150255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,16384,0.009746666583749983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,2560,0.12160710493723552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,3584,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,65536,256,0.13037421968248156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1024,7168,0.0069351109365622205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,3072,0.011863999896579318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,51200,5.2818684048122835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,1536,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,12288,0.03050666716363695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,3072,0.016520000166363187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,256,6144,0.01825777855184343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2560,64,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,3584,0.017475555340449016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,51200,32,0.03879911038610671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,768,0.011393778026103973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,10240,0.1431235604816013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,4096,0.07530755466885038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,12288,0.027151111099455092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,51200,0.12098044819302028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,6144,0.012895999683274163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,2560,0.05319911241531372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,7168,0.012098666694429187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,5120,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,2560,0.02504355543189579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,51200,8192,1.3628639645046656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,128,3584,0.015563555889659457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,6144,0.02749866743882497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,128,0.0020657777786254883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,1536,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,5120,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,5120,0.07755377557542589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,256,0.025635555386543274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,7168,128,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,3072,0.013382222089502545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,2560,0.005254222287072076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,51200,0.8580551147460938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2560,5120,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,3072,0.014774221513006421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,3072,0.018248000078731112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,6144,0.14556177457173666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,3584,0.030990223089853924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,6144,51200,0.10395999749501546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,12288,0.1898346609539456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,51200,0.11258310741848415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,10240,10240,0.02642311155796051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,32,10240,0.012451555165979596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,7168,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,3072,0.13332088788350424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,10240,0.7321928872002496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,12288,256,0.0063235556913746735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,128,0.011438222395049201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,10240,0.01647644407219357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2048,1024,0.04843466811709934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,5120,65536,0.15027999877929688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,64,0.0018800000349680583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,3584,256,0.048919998937182956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,65536,0.08619288603464763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,4096,0.2345928880903456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,16384,0.01757866640885671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,51200,0.15453599558936224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,8192,0.02569688856601715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,8192,0.04781244529618157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,5120,0.00977244476477305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,5120,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,8192,0.04629066586494446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,2560,0.007610666255156199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,6144,0.11606132984161377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,6144,0.046725332736968994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,128,0.024043555061022442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,64,0.010397333237859938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,512,0.08222844203313191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,3072,0.016504888733228047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,512,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,128,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,8192,0.018897778458065458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,2560,0.03010933266745673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,2560,0.10205777486165364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,2048,0.0752755535973443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,768,0.04798400070932177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,2048,0.022620444496472675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,7168,0.033153778976864286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,12288,0.17127645015716553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,5120,0.016335111525323655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,512,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,10240,0.039918220705456205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,768,2560,0.006973333656787872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,512,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,5120,0.039098666773902044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,2560,0.018057778477668762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,10240,0.025960889127519395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,512,16384,0.0348364445898268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,12288,0.42238577206929523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,3072,0.04144800040456984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,3072,0.02030044462945726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,256,0.00702400008837382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,65536,0.0632737808757358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,64,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,8192,0.023523555861579046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,128,0.006575111299753189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,8192,0.03655111127429538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,128,5120,0.01664888858795166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,768,0.022887110710144043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,12288,32,0.002297777682542801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,51200,5.312015957302518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,256,0.013757333159446716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,5120,0.007266666326257918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,65536,0.5952764617072211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,4096,0.20021866427527535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,51200,64,0.2562826739417182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,16384,0.07628889216317071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,16384,256,0.019840889506869845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,64,51200,0.012096888489193387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,3584,0.08346044354968601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,16384,0.2088924381468031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,65536,0.0769760012626648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,51200,0.33163022994995117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,64,0.002141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,3072,0.17785244517856172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,128,0.02550044490231408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,2048,0.03327022327317132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,8192,0.04150844282574124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2560,65536,0.27328266037835014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,12288,0.017862222260899015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,2560,0.0041955556306574075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,2048,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,12288,0.03247822324434916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,128,0.014448000325096978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,6144,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,3584,0.005531555662552516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,12288,0.057745774586995445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,2560,128,0.012415111064910889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,51200,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,10240,0.025079111258188885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,6144,0.03461688756942749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,1536,0.02163644466135237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,1024,0.0734400020705329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,10240,0.021988444858127173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,65536,1.1105146408081055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,12288,0.014744889405038623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,2560,0.05149866806136238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,1024,0.009350222018029954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,10240,2048,0.010455999937322406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,65536,5.182359907362196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,12288,0.4243475596110026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,4096,768,0.024812445044517517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3584,128,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,128,0.019536889261669584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,3584,0.010703999963071613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,3072,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,12288,0.013185777597957186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,6144,0.008995555341243744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,512,0.00720711135201984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,7168,0.22604089313083223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1536,10240,0.21807911660936144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,7168,0.05877688858244154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,2560,0.010095111197895473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,3072,0.012737777498033313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,2048,0.007400888535711501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,256,0.007978666987684038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,5120,0.020208888583713107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,2048,0.013066666821638743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,128,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,768,0.00645955569214291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,65536,2.6762337154812283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,8192,0.011179555621412067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,16384,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,32,0.002458666761716207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,3584,0.033957332372665405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,4096,0.007755555212497711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,7168,2048,0.20561956034766304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,10240,0.05762844615512424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,12288,0.047929777039421924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,16384,0.013437333206335703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,1024,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,6144,0.48105600145128036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,4096,6144,0.08630844619539048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,256,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,12288,0.047247111797332764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,16384,0.054993778467178345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2560,3072,0.007900444169839224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,65536,0.086535996860928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,10240,0.03386577632692125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,65536,0.07209510935677423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,6144,0.013081777426931592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,512,512,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,2560,0.015875554747051664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,51200,0.10723111364576553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,4096,0.037900444534089826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,10240,0.01588355501492818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,7168,3072,0.023303111394246418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,3072,0.014288889037238227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,4096,0.035509334670172796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,6144,0.2501760058932834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,10240,0.11273422506120469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,256,0.015213333898120456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,4096,0.014886221951908536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,16384,0.10881955756081475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,3584,0.07774755689832899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,128,0.005236444373925527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,3072,0.01762488815519545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,12288,0.03602933221393161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,3584,0.6447564231024848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,12288,0.10513066583209568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,2048,0.008444444172912175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,7168,0.023870221442646448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,3584,0.01699199941423204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,6144,6144,0.06416089004940458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,5120,0.0107342220014996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,3072,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,10240,0.2766435676150852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,51200,1024,0.04042222102483114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,2560,0.011979555918110741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,65536,0.07513866821924846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,32,0.001840888833006223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,256,0.005283555636803309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,4096,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,1536,0.017679999272028606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,65536,0.07602844635645549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1024,10240,0.04844177762667338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,5120,0.18365866608089873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,16384,0.2994035614861382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,256,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,1024,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,7168,0.013744889034165276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,65536,51200,0.84097777472602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,1536,0.025460445218616064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,3072,0.010926222635640038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,1536,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,256,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,64,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3584,16384,0.06398755311965942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,3584,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,4096,1536,0.02090488870938619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,32,0.018187556001875136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,6144,0.02461333407296075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,8192,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,51200,0.14653155538770887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,4096,0.007843555675612556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,2048,0.011643555429246692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,2048,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3072,1024,0.010021333065297868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,512,0.018047110901938546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,5120,0.09330133597056071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,2560,0.1434088945388794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,256,0.057927999231550426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,51200,0.06241866615083483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,5120,0.019004444281260174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,3072,0.031643556223975286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,4096,0.11753955152299668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,1536,0.014381332529915703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,2560,0.22464177343580458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,65536,0.23604888386196562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3072,7168,0.08451910813649495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,2560,0.031178666485680476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,256,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,768,0.035727110173967146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,3072,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,2560,0.03405333227581448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,12288,0.030066665675905015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,16384,0.13580000400543213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2048,3072,0.007340444458855524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,51200,0.21985244750976562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,1536,0.0647146635585361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2048,1024,0.01239822225438224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,1536,0.020607110526826646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,2048,0.00463733325401942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,2560,0.007636444436179266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,2560,0.01128088848458396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,12288,0.041570666763517596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2560,2048,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,6144,0.028759111960728962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,32,0.011403555671374003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,32,0.018909333480728995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,512,0.033506668276256986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,5120,0.07212177912394206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,4096,1024,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,1536,0.14827999803755018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,12288,0.027002667387326557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,8192,0.022870222727457683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,10240,0.015188443991872998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,128,0.00564533347884814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,5120,0.008560889297061497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,2048,0.3165404531690809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,512,0.0030782222747802734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,1024,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,3584,0.04160177707672119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,128,0.005246222019195557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,6144,0.024988444315062627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,5120,0.039268443981806435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,3072,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,7168,0.06523822413550483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,4096,0.008206222620275285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,3584,0.01315199997689989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,16384,0.19207110669877794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2560,256,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,16384,0.009905777871608734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,1024,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,4096,0.022551111049122278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,2560,0.3675439887576633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,3072,0.020621332857343886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,5120,0.009637333452701569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,2560,0.02250755495495266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,5120,0.2517626550462511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,65536,0.5891964700486925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,512,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,256,0.02219733264711168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,1536,0.006949333681000604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,16384,0.013497778111033969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2048,3584,0.021239999267790053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,512,1536,0.006443555570311016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,5120,1.2937520345052083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,65536,0.0663964417245653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,12288,256,0.017699556218253244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,51200,512,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,128,0.03751911057366265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,65536,0.11062755849626328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,6144,4096,0.06883466906017728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,16384,0.8145893414815267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,65536,0.027450667487250432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,16384,0.021295999487241108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,16384,0.02048000031047397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,5120,0.009651555783218807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,768,0.009450666606426239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,5120,0.07132444116804335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,7168,65536,0.5452346801757812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,51200,5.391460418701172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,128,0.006528000036875407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,10240,12288,0.19740266270107695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,32,0.006795555353164673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,8192,0.18767733044094512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,65536,0.07160622543758817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,4096,0.014744889405038623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,128,0.019613333874278598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,5120,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,51200,0.06971022155549791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,768,0.0058666666348775225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,1536,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,5120,0.015683554940753512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,768,0.0041617775956789655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,32,0.006631111105283101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,6144,0.05400266581111484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,256,0.01238666640387641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,5120,0.008987555901209513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,12288,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,2560,0.03556888964441087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,128,0.0018568889548381169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,256,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,51200,0.26568622059292263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,512,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,12288,32,0.0028453332682450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,128,0.025906667113304138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,2048,0.08868622117572361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,32,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,7168,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,65536,0.6499679883321127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,256,128,0.011113777756690979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,65536,1.5853796005249023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,256,0.5207111040751139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,6144,512,0.006667555620272954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,768,0.0038453332251972626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,6144,1536,0.05244444476233589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,3584,0.00998755544424057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,16384,256,0.0029200000895394217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,1536,0.4446089002821181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,5120,0.028586665789286297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,2560,0.028668443361918133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,7168,768,0.13618489106496176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,768,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,1536,0.034287999073664345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,256,0.01961955593691932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,5120,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,128,0.012390221986505719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,12288,0.05816000037723117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,256,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,10240,0.01822933389080895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,4096,0.015823110938072205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,3584,0.00758133331934611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,5120,0.09957777791553074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,12288,0.023805333508385554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,64,4096,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,128,0.006004444426960415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,7168,0.018181333939234417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1024,51200,0.04952355557017856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,512,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,2048,0.10724444521798028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,51200,4096,0.07848799890942044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,10240,10240,0.5398880110846626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,1024,0.0041893331540955436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,1536,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,5120,0.020484444167878892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,10240,2560,0.11928266949123806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,512,0.007276444799370236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,2560,0.013756444056828817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,32,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,10240,0.025786666406525507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,3072,0.07750666803783841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,2048,0.06259022156397502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,2560,0.030209776428010728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,3584,0.007340444458855524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,10240,0.3198000060187446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,51200,2.349027633666992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,8192,0.03190933333502875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,1536,0.08214044570922852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,64,0.020177778270509508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,768,0.003936000168323517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,5120,0.25280621316697865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,3072,256,0.015400888191329109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,5120,0.3852391242980957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,6144,0.0709644423590766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,7168,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,64,0.001800000046690305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,3072,0.033477332856920027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,32,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,128,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,1536,0.015080889066060385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,6144,0.019113777412308585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,8192,0.041972445117102734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,2048,0.07523733377456665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,2560,0.009304888546466827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,512,0.010440888504187265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,1024,0.007088888850477006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,10240,0.04515733321507772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,65536,0.07779377698898315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3072,256,0.013843555417325763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2560,4096,0.07264800204171075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,16384,0.13277599546644422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,6144,0.07902844746907552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,1536,0.006412444429265128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,12288,0.06303377946217854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,7168,0.01072622256146537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,65536,0.673276424407959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,4096,0.019337778290112812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,1024,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,3072,51200,0.05221955643759834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,3072,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,4096,0.003935111065705617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,1024,0.009539555344316695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,6144,0.02472266720400916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,2048,0.004881777697139316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,1024,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,64,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,128,0.05631111065546671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,768,0.008637333081828224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,64,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,6144,0.017047999633683097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,5120,0.031030221117867365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,8192,0.013963555296262106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,512,0.01789866718981001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,5120,768,0.02572800053490533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,128,0.010709332923094431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,2560,0.021712000171343487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,8192,0.07933066950903998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,1024,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,256,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,2560,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,12288,0.033575110965304904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,32,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,128,256,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,7168,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,8192,0.018959111637539335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,6144,0.008989333278603023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,8192,0.42569510142008465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,51200,8192,0.11228177282545303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,12288,0.025591111845440332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,768,0.021958221991856892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,6144,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,1024,0.053745776414871216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,12288,2.2093351152208114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,2048,0.02032888929049174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,768,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,10240,0.014975999792416891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,51200,0.8773386743333605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,3072,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,256,0.022300443715519373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,4096,0.17593867248959014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,3584,0.009312888814343346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,1536,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,51200,0.8114844428168403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,512,0.006577777779764599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,1536,0.07597688833872478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,12288,0.01499466598033905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,2560,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,8192,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,12288,1024,0.05138488941722446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,64,6144,0.0674462252193027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,51200,0.0826524429851108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,1024,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,51200,0.10406488842434353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,1024,0.043712887499067515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,2560,0.22305422359042695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,16384,768,0.00619911112719112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,128,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,3072,0.03125244379043579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,2560,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,2048,0.06296444601482816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,512,0.017096888687875535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,4096,12288,0.020220445262061227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,65536,2.8524462381998696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,256,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,10240,0.017608889275126986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,4096,0.0350888901286655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,5120,0.01885688967174954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,256,0.0027386666172080566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,4096,0.013557333085272046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,512,0.019486222002241347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,3072,0.009362666971153682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,4096,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,6144,0.02602311140961117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,8192,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,32,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,65536,0.04273688793182373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,32,0.00628266649113761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,16384,0.04663377669122484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,256,0.004257777912749184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,128,0.004864888886610667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,16384,0.016359110673268635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,4096,0.015292444162898593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,12288,0.08755911058849758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,6144,0.06890400250752766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3072,128,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,51200,0.5859440167744955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,512,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,2048,0.034154666794670954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,65536,0.06863200002246432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,2560,0.01867822143766615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,2560,0.04725244310167101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,4096,0.03678311241997613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,3584,0.020625778370433383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3072,1536,0.011529777612951068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,256,0.033955554167429604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,16384,0.04158400164710151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,512,0.0028320000403457214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,3072,0.011035555766688453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,256,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,51200,0.213408006562127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,6144,0.2268693447113037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,10240,2048,0.012803555362754397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,3584,0.036622222926881574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,4096,0.026755554808510676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,12288,0.5111129018995497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,65536,0.07568711042404175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,4096,1024,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,1536,0.013807111316257052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,768,0.018366222580273945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,2048,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,256,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,6144,0.02125155594613817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,65536,1.0871671040852864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,1024,0.005727111051479976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2048,64,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,2048,0.008960000342792934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,51200,9.555550469292536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,32,0.006918222126033571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,3584,0.11014844311608209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,1536,768,0.007252444823582967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,3072,0.06541066699557833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,1024,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,5120,0.012416000167528788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,3072,0.03353777858946059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3584,3072,0.007310222420427535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,3584,0.05328355564011467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,7168,0.051242666112052075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,7168,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,256,7168,0.09391466776529948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,4096,0.40275200208028156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,2560,0.017151999804708693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,512,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,12288,0.006934222247865465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,7168,5120,0.271412452061971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,6144,32,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,7168,128,0.019871110717455547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3072,2560,0.01827822294500139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,3072,0.030461334519916113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,256,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,8192,0.01962488889694214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,16384,0.022882666852739122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2560,8192,0.02478222217824724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,512,0.0454053348965115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,32,3584,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,5120,0.009696889254781935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,16384,5120,0.056432002120547824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,128,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,5120,0.09679288996590509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,256,0.17808977762858072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,2048,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,768,0.023638221952650283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,5120,0.051925334665510386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,65536,8192,0.7531155480278863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,4096,10240,0.03465777635574341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,6144,0.041454222467210554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,65536,0.07057422399520874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,64,0.0018133332745896445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,10240,0.0391937792301178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,3072,0.008015111088752747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,7168,0.014509333504570855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,128,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,7168,0.08133333259158664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,3072,0.010113777385817634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,7168,0.045901334948009916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,1024,0.026268444127506677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,7168,0.01407999959256914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,1024,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,5120,0.039854221873813205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,64,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3072,1024,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,7168,0.06176977687411838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,10240,0.04307644565900167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,7168,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,16384,0.056468447049458824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,8192,0.07592178053326078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,512,0.007297777467303806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,512,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,128,0.020592888196309406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,512,0.002846222163902389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,64,0.00444088876247406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,1024,0.01145866678820716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3072,1024,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,5120,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,256,0.0028364445186323593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,1024,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,51200,0.750616020626492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,6144,0.35201067394680446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,5120,0.018908444378111098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,1024,0.06728000111050077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,1024,0.024326221810446844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,7168,0.014632888966136508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,2048,0.025968000292778015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,10240,0.10005333026250203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,8192,0.0957404441303677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,6144,0.1176791058646308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,12288,0.05260533425543043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,8192,0.0144195556640625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,768,0.02908888790342543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,1536,0.03970222340689765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,512,0.017940445078743827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,6144,0.011471110913488599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,512,0.01458755632241567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,5120,0.038141333394580416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,512,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,128,0.038007110357284546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,10240,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,2048,0.08011111285951403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,4096,6144,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,3584,0.009930666950013902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,16384,0.029831111431121826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,3072,0.011921777493423887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,65536,0.23120000627305773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,16384,16384,0.07891199986139934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,1024,0.0028862222615215513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,3584,0.014473777678277759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,1024,0.022260444031821355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,512,7168,0.009723555710580613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,2560,0.004594666676388847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,7168,0.07617600096596612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,256,768,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,4096,0.008665777742862701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,10240,0.3747893439398871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,8192,0.03678044345643785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,5120,0.03087377879354689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,7168,0.14204266336229113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,3072,0.026387555731667414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,4096,512,0.01928711103068458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,10240,0.3503519958919949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,7168,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,768,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,10240,0.03477955526775784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,7168,0.014840000205569796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,4096,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,4096,0.009012444151772393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,5120,0.05089333322313097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,51200,0.5491813553704156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,3584,0.1660239961412218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,10240,0.012558221817016602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2560,256,0.005193777796294955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,5120,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,4096,0.06813777817620172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,8192,0.13717956013149685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,3584,0.04394666684998406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,2048,0.06256533331341214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,128,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3584,5120,0.10606844557656182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,3072,0.024336000283559162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,3072,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,4096,0.30515911844041615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,10240,3584,0.12543377611372206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,6144,0.014462222655614218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,12288,0.2074648936589559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,3072,4096,0.059735112720065646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,1024,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,64,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,1536,0.033936000532574125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,128,0.006679111056857639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,16384,3584,0.03827644387880961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3584,3072,0.040135112073686384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,51200,0.03812088900142246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,10240,0.16977955235375297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,32,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,128,0.0921048919359843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,6144,0.015600888265503777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,7168,0.1588426695929633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,128,0.012992888689041138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,3072,1024,0.022096888886557683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,16384,0.020799999435742695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,12288,0.15465244981977674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,2560,0.00692622239391009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,5120,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,1024,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,1024,0.024081778195169237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,768,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,32,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,2560,0.05605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,16384,0.1412071122063531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,5120,0.027449778384632532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,2048,0.006228444476922353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,6144,256,0.015615999698638916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,1536,0.005879111174080107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,10240,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,7168,0.23221778869628906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,3072,0.07336355580223931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,256,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,1024,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,32,0.001779555653532346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,5120,1536,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,3072,0.018412444326612685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,3584,0.014378666877746582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,16384,0.08790489037831624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,65536,0.0626844432618883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,65536,512,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3072,7168,0.07794311311509874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,3072,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2560,16384,0.017494221528371174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,4096,0.017536888519922893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,3072,0.03182755576239692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,512,0.015427556302812366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,32,0.01107911103301578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,64,0.01037688884470198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,8192,0.02879200047916836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,8192,0.012416889270146688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,5120,0.01589866644806332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,3072,0.03544800149069892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,128,0.0052764444715446895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,7168,0.018915555543369718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,1024,0.2668151060740153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,768,1536,0.009032888544930352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,768,256,0.010407999985747868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,6144,0.01254488858911726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,7168,0.3107679949866401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,128,0.014677332507239448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,1024,0.0035208890007601846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,2560,0.031437334087159895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,4096,0.029056890143288508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,768,0.020994666549894545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,128,0.00870666652917862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,1024,0.009445333646403419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,2560,0.0329111119111379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3072,2560,0.022648889157507155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,10240,6144,0.01795644395881229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2560,12288,0.26745687590705025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3072,4096,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,3072,0.00776800016562144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,1024,0.006232888748248418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,128,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,256,0.005201777650250329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,5120,0.09471555550893147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,12288,0.2585119936201307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,5120,0.142502228418986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,65536,0.15451467037200928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,3072,0.018211555149820115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3072,16384,0.15143556065029568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,3584,0.050606220960617065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,8192,0.045722666713926524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,32,8192,0.034929778840806745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,3072,0.07321066988839044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,8192,0.05886489152908325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,512,0.014492443866199918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,1024,0.006892444358931647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,12288,0.01941155559486813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,64,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,1024,0.01849422189924452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,1536,0.30906489160325795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,10240,0.036794665786955096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,7168,0.02198933396074507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,12288,0.0543502238061693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,8192,2.7984657287597656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,128,128,0.013012444807423485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,1536,0.008663111262851292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,768,6144,0.009015111459626092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,4096,256,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,2560,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,768,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,8192,51200,0.5196080207824707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,16384,0.15634132756127253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3584,256,0.017328000730938382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,7168,0.04050844576623704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,12288,0.04974666568968031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,6144,0.02548888822396596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,8192,0.008254222571849823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,128,0.015835555063353646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,5120,2048,0.019498666127522785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,2560,0.007260444263617198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,65536,0.2576142152150472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,768,7168,0.04498222139146593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,1024,0.0031795555518733132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,128,16384,0.02571288910177019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,1024,0.017315554949972365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,128,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,3072,0.03609244359864129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,12288,12288,0.04302666584650675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,512,0.05951911211013794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,256,0.014762666490342883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,1024,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,7168,0.02481422159406874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,2560,0.21403822633955213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,5120,0.008613333106040955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,128,0.02123911182085673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,65536,0.07268444697062175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,2048,0.032405333386527166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,768,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,10240,0.22095821963416207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,8192,0.013006221916940479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,128,0.013767110804716745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,32,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,768,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,7168,0.011562666959232755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,512,51200,0.05160266823238797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,512,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,65536,256,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,1024,0.10357599788241917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,4096,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,6144,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,4096,0.19578666157192656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,8192,0.07651288641823663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,3072,0.037257777320014104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,128,0.0038399998512532976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,3072,0.07890666855706109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,5120,0.016157334049542744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,12288,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,10240,0.023984000086784363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,2560,0.28082044919331867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,2560,0.01983022193113963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,65536,0.1614560021294488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,1536,0.019291554888089497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,10240,256,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,51200,0.17696800496843126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,3584,0.017296888762050204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,4096,0.12683022022247314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,512,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,3584,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,7168,0.008634666601816813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,51200,7.3055674235026045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,10240,3584,0.016125332978036668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,3584,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,768,0.0029031110720502008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,1024,0.017246221502621967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,256,0.0606915553410848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,2048,0.05772888660430908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,1024,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,12288,0.37510045369466144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,7168,0.025609778033362493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,16384,0.10022399822870891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,4096,65536,0.7217813597785102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,65536,3584,0.3384702205657959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,16384,0.021733333667119343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,7168,0.013407111167907715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,12288,0.0162426663769616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,512,0.0181048893266254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1536,5120,0.02439733346303304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,65536,3584,0.11241777737935384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,512,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,4096,0.010945777926180096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,51200,0.035311109489864774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,51200,1.7885636223687067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,1024,0.009291555318567488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,8192,0.06332177586025663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,51200,0.14008711444007024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,3072,0.030797332525253296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,64,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,3584,0.04006577862633599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,7168,0.02129777769247691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,3072,0.014838222000333997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,3584,0.009581333233250512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,64,0.0019333333604865605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,4096,0.028148445818159316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,65536,0.08981421920988296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1024,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1024,51200,0.1184613307317098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,1536,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,3584,0.023016000787417095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,256,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,3584,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,1024,0.012379555238617791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,32,0.0025253333151340485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,1024,0.0069217777086628815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1536,768,0.005192000004980299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,128,0.005014222115278244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,8192,1024,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,1536,0.03371822171741062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,32,0.003740444365474913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,4096,0.008957333034939235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,16384,0.13674133353763157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,3584,0.0318177772892846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1024,12288,0.011075555450386472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,256,10240,0.028592887851927016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,5120,0.07483288976881239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,768,0.00611111107799742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,512,0.01719288859102461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,7168,0.013282666603724161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,6144,3072,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,3072,0.0074942219588491656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,64,12288,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2560,5120,0.046244445774290294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,51200,0.11274045043521458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,6144,0.023652444283167522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,256,256,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,1536,0.00508622204264005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,8192,0.04898311032189263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,768,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,10240,16384,0.056410663657718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,8192,0.13458400302463108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,2048,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,768,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,10240,0.17814489205678305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,3584,0.1128515534930759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,256,0.005969777703285217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,51200,0.21721777651045057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,65536,0.06523466772503324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,12288,0.02060622142420875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,64,0.006211555666393704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,4096,0.020584889584117465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,256,0.004884444591071871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,10240,512,0.013020444247457715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,65536,3072,0.531617800394694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,3584,0.07794577545589872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,3584,0.09823466671837701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,12288,0.027564444475703772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,512,0.24540445539686415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,3584,0.06075377596749199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,2560,0.024298666252030268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,3584,0.029449777470694646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,10240,0.01348355578051673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,10240,0.26785778999328613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,65536,0.07859733369615343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,16384,768,0.01958933307064904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,2560,0.4115866555107965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,16384,0.013839999834696451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,256,0.01276266657643848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,64,0.006326222171386083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,10240,0.05919910801781548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,10240,0.019137778215938143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,5120,0.016551110479566786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,12288,0.02884355518552992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,8192,0.06468266910976835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,1024,0.022414222359657288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,2048,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,4096,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,768,0.003244444520937072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,8192,0.010780444575680627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,65536,4.222096761067708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,512,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,6144,0.05457866523000929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,512,0.1162266731262207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,2560,0.009538667069541084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,3072,0.029071112473805744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,3584,0.016901334126790363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,4096,51200,0.8930400212605795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,12288,0.07549422317081027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,12288,2560,0.016230222251680162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,8192,6144,0.22397955258687338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,256,0.008286221987671321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,2560,0.017628444565667045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,65536,2048,0.08185066779454549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,51200,5120,0.10055289003584121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,5120,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,256,0.006439111298984951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,16384,0.020896888441509668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,65536,16.630291408962673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,65536,0.15286489327748617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,65536,0.19095910920037162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,3072,0.03494933247566223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,2048,0.011598221957683563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,2048,0.024381332927280005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,10240,0.02741777731312646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,8192,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,5120,0.014331556028789945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,16384,0.03151288959715102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,6144,0.22396622763739693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,16384,0.17329955101013184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,6144,768,0.004212444441186057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,16384,0.0332897769080268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,5120,51200,0.10933066738976373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,1024,0.00388355553150177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,16384,12288,7.8078418307834205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,51200,65536,11.906756083170572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,3072,0.030454221698972914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,16384,0.13567377461327448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,32,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2048,1536,0.008271999657154083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,1536,0.011560888753996955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1024,2048,0.034508443540996976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,1536,0.013880000346236758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,8192,4096,0.011297778122954898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,32,16384,0.045491556326548256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,51200,0.18789511256747773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,7168,0.04663822385999891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,256,16384,0.07495644357469347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,3072,0.045370668172836304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,128,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,10240,10240,0.2462115552690294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,2048,0.005396444350481033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,8192,0.011049778097205691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,3584,0.010050666828950247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,12288,1536,0.07921155293782552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,4096,0.23475466834174263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,256,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,6144,0.044658667511410184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,64,0.014030221435758801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,65536,0.1129102177090115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,2560,0.08172000116772123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,7168,8192,0.02329866588115692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,768,51200,0.0502071115705702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,256,0.006576889091067844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,128,7168,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,1536,0.006235555642180973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,5120,0.0366613335079617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,65536,6144,0.6229199833340114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,12288,256,0.08823555707931519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,3584,3584,0.10593244764539932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,128,0.004919111314747068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,3072,0.00795555528667238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,256,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,51200,0.12505333953433567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,65536,128,0.04963199959860908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,6144,64,0.12871289253234863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,6144,0.0320995549360911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,8192,0.02568888829814063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,2048,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,512,0.020226667324701946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,2048,0.024961777859263953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,3584,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,51200,1.7059742609659831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,3584,0.8089413113064237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,2048,0.023464888334274292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,12288,0.005256888767083486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,128,0.012631999949614206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,2048,0.005689777847793367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,768,0.01994755533006456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,51200,0.040411555104785495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,12288,0.052523556682798594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,2560,0.012658666405412885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,51200,0.25770311885409886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,2048,0.01147555559873581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,51200,0.17098755306667754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,256,0.004541333350870344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,8192,0.011336888704035016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,768,0.1585751109653049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,1024,0.015093333191341825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,768,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,5120,0.017330666383107502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1536,8192,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2560,4096,0.01662577854262458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,32,0.011016888750924004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,5120,16384,0.03614933292071024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,1536,1.842361874050564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,3072,0.0595395565032959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,8192,0.01999022232161628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2048,10240,0.11826222472720677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,10240,0.011342222491900126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,8192,0.0159599996275372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,12288,0.016695110334290397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,256,0.007267555428875818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,65536,0.06324711110856798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2048,1536,0.020322667227851022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,2560,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,12288,0.013397333522637686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,3072,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,16384,0.2596204545762804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,12288,0.08095288938946195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,3072,0.013004444539546967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,7168,51200,0.060101330280303955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,2048,0.1266231139500936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2560,4096,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,51200,2560,0.06608977582719591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,4096,0.008277333445019191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,8192,5120,0.0728364454375373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,256,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,16384,0.29392798741658527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,5120,0.0306951105594635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,12288,0.02390044430891673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,3072,0.06400533517201741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,64,7168,0.02421244482199351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,16384,0.6767324341668023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,256,0.11119911405775283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,2560,0.02052177819940779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,3584,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,3072,0.023974221613672044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3584,8192,0.02522666586769952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,65536,2560,0.41663021511501735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,256,0.04855644371774462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,512,0.0031600000543726813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,768,4096,0.027706666125191584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,256,0.030031111505296495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,7168,0.20092356204986572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,32,0.006311111152172089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,65536,0.13872710863749185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,10240,0.04217155443297493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,12288,0.07160977522532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,3072,0.013735999663670858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,65536,0.924774169921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,768,0.0031439999325407874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,512,0.24005156093173555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,65536,1.2920488781399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,256,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,65536,0.0879324475924174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,128,0.002436444370283021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,1024,0.013058666553762225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,6144,0.08192533254623413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,5120,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,768,0.01588088936275906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,6144,1.1514204872979057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,512,10240,0.01239466667175293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,256,51200,0.42506933212280273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,4096,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,512,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,8192,0.010760889285140567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,65536,0.03594577643606398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,65536,3584,0.3268124527401394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,51200,2.244000964694553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,65536,0.266628450817532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,8192,0.21203111277686226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,10240,0.035551110903422035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,768,0.008973333570692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,512,0.0052782222628593445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,12288,0.0677200025982327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,10240,1536,0.04014844364590115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,51200,1536,0.024312888582547505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,10240,0.06904888815350003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,1536,0.05977244509590996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,10240,0.14026578267415366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,256,0.028536889288160536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,12288,0.007668444679843054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,12288,0.14465510845184326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,3584,0.014008000493049622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,128,0.010319111247857412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,1536,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,1024,0.01754044493039449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,8192,0.02304711110062069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,65536,2.7926355997721353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,128,3072,0.022130666507614985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,5120,0.6442586580912272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,4096,0.2386524412367079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,256,0.015565334094895257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,32,0.0037679999238914917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,16384,0.032808889945348106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,2048,0.12531556023491755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,4096,0.011015111373530494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,5120,0.006304889089531369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,128,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,512,1536,0.020759110649426777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,7168,0.039749334255854286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,3584,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,1536,0.01889155473974016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,4096,0.015544889701737298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,512,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,8192,0.01074577785200543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,65536,0.06810222069422404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,51200,0.2429706785413954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,1024,0.013415111435784234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,6144,0.05152177810668945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,128,0.013223999904261695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,1024,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,8192,7168,0.028816888729731243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,7168,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,1536,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,7168,0.017782222893502977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,2560,0.0096124443742964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,51200,0.1720506615108914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,256,0.01349244432316886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,6144,2.4716044531928167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,3584,0.12814933723873562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,7168,0.13920088609059653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,768,0.014426666829321118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,2048,0.007002666592597961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,7168,0.05710133579042223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,6144,0.03154844376775954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,5120,0.016545777519543965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,10240,0.18110844824049208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,51200,0.10445066955354478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,65536,0.0990826686223348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,10240,0.04769955409897698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3072,3072,0.007667555577225155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,1024,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,32,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,6144,0.533957322438558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,6144,0.018441778090265062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,5120,0.03167111012670729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,8192,3584,0.12806755966610378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,1024,0.0178195552693473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3072,1536,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,7168,0.032680001523759626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,16384,12288,0.25583378473917645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,512,0.0031022222505675424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,65536,0.6260515318976508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,12288,0.04093155595991347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,1536,0.027303111222055223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,1536,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,6144,0.020246222615242004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,51200,0.05686044692993164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,12288,1.0813609229193795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,7168,12288,0.04920444554752774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,6144,128,0.013008889224794177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,51200,0.2789075639512804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,10240,0.028757333755493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,1024,0.01924888955222236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,1024,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,512,0.002546666603949335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,4096,0.015357333752844067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,64,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,4096,256,0.003504000190231535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,8192,0.011734222372372946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,512,0.03200711144341363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,1024,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,7168,0.06342933575312297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,3584,0.03086755673090617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,7168,1024,0.030621333254708186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,5120,0.009167999857001835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,4096,5120,0.022312889496485393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,16384,0.24544978141784668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,51200,0.9269235399034289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,3072,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,32,768,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,8192,0.020962667134073045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,12288,0.018774222996499803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,3584,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,256,64,0.00564177789621883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,768,0.12567467159695095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,3584,0.022224000758594934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,3584,512,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,5120,1536,0.019256888164414298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,16384,0.08271200127071805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,64,0.046726223495271474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,7168,0.012885332935386233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,768,0.00740533322095871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,6144,2560,0.008656000097592672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,65536,1.373694207933214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,4096,0.6293439865112305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,10240,0.04907822277810839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,3584,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,64,0.0023448889454205832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,1536,0.0046311111913787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,32,0.006752000086837345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,3072,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,5120,0.023074666659037273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,8192,0.13034755653805202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,64,0.002016000035736296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,12288,0.030789332257376775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,7168,0.21577866872151694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,3072,0.024488000406159297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,10240,0.3482053279876709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,2560,0.04089066717359755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,3072,0.7296889093187121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,3072,0.02072533302836948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,256,0.00602311102880372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,65536,1.1318160163031685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,65536,256,0.012741333080662621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,8192,0.11563555399576823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,16384,0.04590400060017904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,4096,0.070905778143141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,512,0.005630222045713001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,1536,0.020925333102544148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,1536,0.011576000187132092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,3072,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,51200,3584,0.07588355408774482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,3072,0.017512889371977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,768,0.01924000018172794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,2560,0.004581333448489507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,10240,0.05726666582955254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,8192,0.033207999335394964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,512,0.006167999986145232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,512,0.006214222146405115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,768,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,512,3584,0.01499111122555203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,256,0.0069982219073507525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,6144,0.18450755543178984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,32,0.034134222401512995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,256,6144,0.02943555514017741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,3584,0.00566133318675889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,6144,0.024642666180928547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,32,65536,0.2835306591457791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,3584,0.04895288745562235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,5120,0.03964799973699782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,4096,0.29324976603190106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,768,7168,0.006592000110281839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,6144,0.05397955576578776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,3584,0.023373333944214716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,10240,0.1325964397854275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,51200,0.08451910813649495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1024,2048,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,768,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,4096,0.028592000404993694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,64,0.007240888973077138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,8192,0.9761813481648763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,10240,0.200981338818868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,16384,0.05778400103251139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,768,16384,0.020940444535679285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,128,7168,0.006447111152940326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,5120,0.057681779066721596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,2560,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,768,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,256,256,0.0021768889079491296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,1024,0.004528000122971005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,256,0.036722666687435575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,6144,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,512,0.01036088913679123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,7168,0.013108444710572561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,512,2560,0.010977778169843884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,768,512,0.004577777865860197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,8192,0.14752533700731066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,8192,0.021758221917682227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,1024,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,5120,0.021207110749350652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,65536,0.7019848823547363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,6144,0.041040887435277305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,12288,0.13032621807522243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,1024,0.17705955770280626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,5120,2560,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,128,0.005436444448100195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,16384,0.04420533445146349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,768,0.087135997083452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,128,0.015126221709781222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,32,0.008064000142945183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,2048,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,2560,0.032059556908077665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,6144,4096,0.14549067285325792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,2560,0.06993777884377374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,8192,0.012121777567598553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,3584,0.0440684457619985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,2048,0.009026666482289633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,3072,0.031193776263131037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,4096,0.00424799985355801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,4096,3.2126941680908203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,5120,16384,0.03279644582006667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,7168,0.011133333047231039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,7168,0.013370667066839008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1536,768,0.017549332645204332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1024,6144,0.026322666141721938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,4096,0.020237333244747586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,51200,0.7273173332214355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,768,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,16384,0.033253331979115806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,7168,0.011721777419249216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,8192,0.012216000093354119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,65536,2.6047901577419705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,512,0.03144977821244134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,4096,0.09827733039855957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,512,0.012368000215954252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,3584,0.085917333761851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,7168,1024,0.004528000122971005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,3072,0.004232888834344016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,10240,0.02812088860405816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,1024,0.011975111232863532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,10240,0.033124443557527326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,512,65536,0.08451466427909003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,16384,0.03714222378200955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,128,65536,0.011618666350841522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,6144,0.07915822002622816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,1536,0.016192000773217943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,6144,4096,0.026022222306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,12288,0.06467377477222018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,51200,8192,0.078402664926317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,16384,0.025027554896142747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,2048,0.07049689027998182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,6144,0.15691999594370523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,10240,2048,0.17253600226508248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,5120,0.08897155523300171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,1024,0.020273778173658583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,51200,1536,0.32265154520670575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,4096,0.013165333204799227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,4096,0.09106133381525676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,3584,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,6144,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,12288,0.04554133282767402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,1024,1536,0.004885333279768626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,768,7168,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,3584,0.029045333464940388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,51200,0.06891377766927083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,256,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,256,0.016050666570663452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,4096,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,256,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,5120,0.14857244491577148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,4096,0.05892622470855713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,3584,0.022590221630202398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,5120,0.033017777734332614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,8192,0.04219466778967115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,2048,0.15733067194620767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,51200,32,0.1083022223578559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,6144,0.04394933250215319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,4096,0.30139999919467503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,256,0.006528888725572162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,12288,0.013649777405791812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,2048,0.007106666763623555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,8192,0.08839555581410725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,16384,0.07883911000357734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,1536,0.04348266786999173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,5120,2560,0.007691555553012424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1024,8192,0.01425333321094513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3584,2560,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,3584,0.035550223456488714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,1024,0.022244445151752893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,32,10240,0.012522666818565793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,8192,0.0980444418059455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,12288,0.025849777791235182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,3584,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,3584,0.01926577753490872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,64,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,6144,0.038613332642449275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,16384,0.07766666677263048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,2048,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,10240,51200,0.7217591073777941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,8192,0.013966222604115805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,7168,0.010403555300500657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1024,2048,0.00885511106914944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,768,5120,0.008715555899673039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,3072,0.017272000511487324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,12288,0.694710201687283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,7168,0.0973022249009874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,4096,0.021242666575643752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,32,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,256,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,3072,0.008368000388145447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,6144,3584,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,2560,0.033238222201665245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,2048,0.038859556118647255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,1024,51200,0.04994844396909078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,5120,0.0284826656182607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,16384,0.03126577867401971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,1536,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,4096,0.015964445140626695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,16384,0.1280399958292643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,7168,0.6682915687561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,512,0.007383999725182851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,128,0.0019804444164037704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,1536,0.11418933338589138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,2560,0.011566222541862063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,10240,4096,0.1691324445936415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,512,0.23295021057128906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,65536,0.15178310871124268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,12288,2.5253643459743924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,768,0.009663110805882348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,5120,0.11882400512695312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,12288,0.01684888866212633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,51200,0.2592826684315999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,6144,0.6034408675299751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,10240,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,768,0.023558222585254248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,8192,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,1536,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,1536,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,10240,4.092811584472656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,768,0.01777688827779558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1024,512,0.008999110923873054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,5120,51200,0.06293244494332208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,128,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,128,0.018264888061417472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,512,0.017306667235162523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,51200,0.3295075628492567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,4096,0.12239466773139106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,51200,7168,0.11495110723707412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,7168,1.0441128412882488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,6144,0.038336889611350164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,51200,0.1256355577045017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,512,0.010098666780524785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,2560,0.040223111708958946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,3072,12288,0.015769778026474845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,512,0.035262223747041486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,512,8192,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,3072,0.04607111215591431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,2048,0.012788444757461548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,5120,7168,0.27356089486016166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,3072,0.025784000754356384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,512,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,3584,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,3072,0.008323555191357931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,3584,0.11453244421217178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,6144,0.015899555550681222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,65536,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,7168,0.014750222365061441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,16384,0.019896888070636325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,512,51200,0.026734222968419392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,1536,0.07507377862930298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,3584,0.010323555933104621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,64,0.04535822073618571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,32,0.02553244431813558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,768,0.034847110509872437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,65536,0.31524621115790474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,8192,1.0975253846910265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1024,16384,0.049603554937574595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,5120,0.009384888741705153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,5120,0.06463111109203763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,6144,0.052357332573996655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2560,1024,0.023579556081030104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,3584,0.07797244522306654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,128,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,12288,0.10906222131517197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,2560,0.14862844679090711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,1024,0.008003555238246918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,65536,1536,0.24418134159511992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,65536,7168,0.14665510919358996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,1536,0.009906666974226633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,6144,0.27958043416341144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,12288,0.019542222221692402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,7168,0.01697422232892778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,3072,0.01475911173555586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,2560,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,4096,768,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2048,10240,0.1368808878792657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,5120,7168,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,2048,0.023818666736284893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,1536,0.0041884444653987885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,1536,1536,0.016159110599093966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,512,6144,0.01584533353646596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1024,2048,0.01687111126052009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1536,3072,0.016170667277442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,2048,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,2560,12288,0.050399998823801674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,3072,0.018619555566045973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1024,6144,0.029691557089487713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,6144,0.02716000046994951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,7168,0.05576533079147339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,65536,1536,0.0682835578918457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,256,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,1536,0.018378666705555387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,8192,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,128,1536,0.0038568890757030914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,2560,0.0048942222363419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,65536,0.16181955072614881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,16384,0.015783111254374187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,8192,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2048,1024,0.012438221938080259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,6144,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,5120,0.00868444475862715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,7168,0.01404088901148902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,51200,1536,0.028776887390348647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,5120,32,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,768,2560,0.007248000138335758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,65536,1.904327180650499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,16384,0.04684355523851183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,3584,0.02199466692076789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,1536,0.02093688812520769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,3072,0.07751199934217665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,5120,768,0.0058311112225055695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3072,12288,0.4100968837738037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,16384,0.0528728895717197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,768,65536,1.4895795186360676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,256,1024,0.015814221567577787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,6144,7168,0.0649777783287896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,3072,0.10986755291620891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,256,0.013620444469981723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,1536,0.0035662220584021676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1024,12288,0.04522933231459724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,16384,1.5419546763102214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,3584,0.1500142282909817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,65536,0.06253244479497273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,3072,0.004927999857399198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,7168,5120,0.05185955431726244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,5120,12.714126586914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,6144,0.11823377344343398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,32,0.002871111035346985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,768,768,0.005615111026499007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,1536,0.0063475556671619415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1536,3072,0.016559999850061204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,256,0.024978665841950312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,768,0.026503110925356548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,3072,0.010249777800507015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,256,2560,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,3072,0.23187645276387533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,10240,0.17362756199306914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,10240,0.005573333137565189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1024,1536,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,128,4096,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,6144,0.01036000003417333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,64,1536,0.0038773334688610504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,128,0.014479110638300577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,3072,0.174782223171658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,4096,4096,0.011402666568756104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,3072,0.01788444485929277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3072,512,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,51200,5120,0.3785635630289714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1024,1024,0.005913777897755305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,51200,0.06332711378733318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,5120,0.2044959995481703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,64,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,768,0.009924444059530893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,256,0.07441955804824829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,1024,0.018577777677112155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,5120,0.09297688802083333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,5120,0.009360888765917884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,64,0.0018595556418100994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,2048,0.0075031113293435835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,3584,0.008042666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,5120,0.022655111220147874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,1024,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,1536,0.01718311177359687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,12288,0.1264008945888943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,10240,64,0.03276888860596551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,51200,0.08536622259351943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,3072,0.0075191110372543335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,2560,0.004980444494220945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,4096,0.016300444801648457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,2048,0.02216888964176178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,16384,0.10566933287514581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,64,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,8192,12288,0.03978577918476529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,1536,0.032030221488740705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,3072,0.08412800232569377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,5120,0.05124533176422119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,65536,0.03462844424777561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,1024,0.007703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,65536,0.07661244604322645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,512,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,768,0.014834667245546976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,1024,0.006011555592219035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,6144,0.019983111156357657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,51200,0.06934666633605957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,2560,0.005014222115278244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,1536,0.029746668206320867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,6144,2048,0.03683822353680929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,32,0.002179555594921112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,128,16384,0.016197333733240765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,51200,7168,0.050154666105906166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,10240,0.03358577688535055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,7168,0.08141155375374688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,12288,0.04634311132960849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,8192,0.04389511214362251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,51200,0.06524178054597643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,4096,7168,0.011752000285519494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,3072,0.02827199962404039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,256,512,0.00245066670080026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,256,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,2560,0.02292977770169576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,2048,0.00815644446346495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,512,0.18843911753760445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,128,65536,0.5096426539950901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,3584,0.016905777984195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,128,0.005589333259397083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,2048,0.019542222221692402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,3072,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,51200,0.29904177453782826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,65536,0.3216035630967882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,128,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,256,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,51200,128,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1536,6144,0.05654578076468574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,128,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,51200,256,0.02086755633354187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,5120,0.014408000641398959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,2048,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,51200,0.7047626707288953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1536,256,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,2560,0.06419289112091064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,8192,0.13979821734958225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,51200,12.23908911810981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,5120,0.06410311328040229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,1024,0.07793600029415555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,512,128,0.013119999733236102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,1536,0.07183021969265409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,51200,4.573970794677734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,65536,0.4991031222873264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,51200,16384,0.2517022291819255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,64,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,8192,65536,0.08966844611697727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,2560,0.02760266595416599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,7168,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,12288,0.08323466777801514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,6144,0.011039999624093374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,1536,0.02755733331044515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,2560,0.04996000064743889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,1024,0.3028346697489421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,12288,32,0.011016888750924004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,16384,0.005623999983072281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,256,0.019884443945354886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,128,16384,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,256,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2048,10240,0.043453332450654775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3584,65536,0.3566088941362169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,64,32,0.002521777732504739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,3072,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,256,8192,0.020632889535692003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,512,4096,0.0188755558596717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,51200,32,0.09387022256851196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,2048,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,4096,3584,0.04495555493566725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,256,4096,0.015448888142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,128,0.0021066665649414062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,256,0.01648266613483429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,2048,0.024850666522979736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,2560,0.022653333014912073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,65536,0.17310933272043863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,2048,0.02529866662290361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,32,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3072,64,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,128,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,16384,768,0.06760266754362318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,4096,0.07539821995629205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,2560,0.05463822351561653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,768,0.027311111489931744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,512,65536,0.07469244135750665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,2048,0.004202666795916027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,16384,0.1048239999347263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,51200,0.11273511250813802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,7168,0.058042665322621666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,12288,0.009535999761687385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,2560,0.010312000082598792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,256,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,6144,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,32,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3072,256,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,6144,0.03499733408292135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,2048,0.007153777612580194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,32,0.002014222244421641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,4096,0.07878488964504667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,5120,0.1191822157965766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,2048,0.9916275872124566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,4096,0.022304000125990972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,2560,0.00903555585278405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,2560,0.009699555734793345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,12288,0.2555573251512316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2560,3584,0.020271999968422782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,5120,0.03979911075698005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,768,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,2560,0.025379555092917547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,10240,1024,0.025981333520677354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,8192,32,0.010682666467295753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,7168,0.0384880006313324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,2048,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,6144,0.04670133193333944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,512,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,51200,0.05189866820971171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,32,0.014819555812411837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,10240,0.015142222245534262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,768,256,0.015072888798183866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,512,2048,0.007487110793590546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,512,512,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1024,3584,0.009425777528021071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,32,0.0021066665649414062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,16384,0.10704177618026733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,5120,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,256,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,3584,0.0749724441104465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,7168,5120,0.03379999929004245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,5120,0.11204000314076741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,512,0.014049778381983439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,4096,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,256,0.008330666356616551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,3584,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,1024,0.004976888911591636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,6144,0.023172444767422144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,16384,0.01648977730009291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,4096,3584,0.02676711148685879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,1536,0.043655110730065234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,10240,0.028480887413024902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2560,51200,0.6833670934041342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,2048,0.027405333187845018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,1536,0.0038328886859946777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,16384,0.09984622399012248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,1536,0.00720000018676122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,2048,0.010673777924643623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,10240,13.605643378363716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,5120,0.02645066711637709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,768,51200,0.2523546748691135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,64,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,768,8192,0.03394488824738397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,4096,0.02910755409134759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,3584,0.029473778274324205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,5120,0.014578666951921252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,6144,12288,0.05548977851867676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,10240,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,3584,0.03615377677811517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,3584,0.03741777936617533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,5120,0.009701333112186855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,12288,0.0875706672668457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,6144,128,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,16384,0.02846933404604594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,2560,0.026551999979548987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,1024,0.014244443840450711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,16384,0.08904266357421875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,8192,0.34036978085835773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,6144,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,2048,0.05334844523006015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,16384,0.07970399989022149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,1536,0.021545777718226116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,8192,128,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,8192,0.01854311095343696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,8192,0.013426666458447775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,3072,0.010471999645233154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,128,5120,0.01516533394654592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,6144,0.10637866788440281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,512,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,4096,0.025196444657113817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,5120,0.009280888570679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3072,2048,0.2717217869228787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,64,0.0028488888508743713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,768,7168,0.05578577518463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,8192,2048,0.019716444942686293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,4096,0.03529955612288581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,6144,0.030014223522610132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,1536,0.012107555237081317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,8192,0.013778666655222574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,512,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,128,0.006222222414281633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,51200,0.08566844463348389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,2560,0.05134044422043694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,3072,0.010835555692513784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2048,7168,0.021759111020300124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,128,12288,0.013760888742076026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,7168,0.0379617777135637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,51200,0.04987022280693054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,512,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,2048,0.015025777949227227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,10240,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,12288,0.017490666773584154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,128,0.006554666492674086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,2048,0.013089777694808112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,10240,0.022965333527988855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,5120,6144,0.011108444796668159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,12288,0.5874648623996311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,51200,128,0.11823733647664388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,12288,0.5264337857564291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,3072,0.025729777084456548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,3584,0.03459111187193129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,1024,512,0.01274577776590983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,8192,0.17251733938852945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,4096,3584,0.014478221535682678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,12288,768,0.006467555546098285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,3072,0.5182808770073785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,32,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,16384,0.022890667120615642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,2560,0.0069102222720781965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,2560,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,512,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,12288,0.10700533125135635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,768,0.03384266628159417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,1024,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,256,12288,0.7793128755357531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,6144,0.04898666673236423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,64,0.0018524444765514797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,10240,0.028051555156707764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2560,12288,0.05827822287877401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,1024,0.018917333748605516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,32,4096,0.01998844411638048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,65536,0.20846311251322427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,2048,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,10240,768,0.005872888697518243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1536,2048,0.03501066565513611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,7168,128,0.019545776976479422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3584,2048,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,5120,1536,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,4096,0.05004622207747566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,16384,0.11284266577826606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,6144,0.04706755611631605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,1536,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,4096,0.021928000781271193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,2048,0.01607911123169793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,128,0.001965333397189776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,8192,0.08577955431408352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,768,0.0058542220956749385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,6144,51200,1.0837892956203885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,1536,0.09120355712042914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,512,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,768,0.009525333013799455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3584,16384,0.2204666667514377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,10240,0.3819119930267334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,1536,2560,0.019913777709007263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,512,0.012539555629094442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,768,0.005668444352017508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,3072,0.010952889091438718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,128,0.01514844430817498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,128,0.010374222364690568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,3584,0.008314666648705801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,16384,0.016176000237464905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,1536,0.05145155390103658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,12288,0.07099111212624444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,4096,0.06337333387798734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,51200,0.11446666717529297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,2560,0.023375110493765935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,7168,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,51200,0.7359813584221734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,1024,0.07689688603083293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,3584,0.012109333442317115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,10240,0.02088800072669983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,1024,0.1766995588938395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,16384,0.024177778098318312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,8192,0.014194667339324951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,128,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,51200,0.06501511070463392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,10240,0.05061689019203186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,8192,0.023541332946883306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,8192,0.04987022280693054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,7168,0.014352889524565803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,32,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,5120,0.009286222358544668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,65536,3072,0.0876764456431071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,512,0.022316444251272414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,1024,0.026055110825432673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,10240,12288,0.044600887431038745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,51200,0.425165335337321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,64,0.0105253333846728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,6144,0.06245244211620755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,768,0.016976000534163583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2048,4096,0.008084444536103142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,8192,0.028879112667507593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,2560,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,32,51200,0.12420088715023464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,3072,0.022388445006476507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,65536,0.13438933425479466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,65536,0.07729955514272054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,256,0.00906488878859414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,32,0.0069466663731469055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,12288,0.6873324182298449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,6144,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,8192,0.024792000651359558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,6144,11.918387518988716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,16384,0.0798435542318556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,2048,0.023992000354660883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,8192,2048,0.007480888730949826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,3584,0.007144889069928064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,512,0.03390044305059645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,128,0.005391110976537068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,2560,0.0810719993379381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,7168,0.04753511150677999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,32,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,1536,0.008030222521887885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,3584,0.014815110299322339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,12288,0.20908444457583955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,128,0.014482667048772177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,7168,0.09769155581792195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,6144,0.004915555732117759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,2048,0.00703288863102595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,128,1024,0.011708444191349877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,512,0.005395555661784277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,7168,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,1536,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,1024,0.28260355525546604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,6144,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,51200,1.346592903137207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,12288,8192,0.0904959970050388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,1536,0.015834665960735746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,3072,0.007948444121413762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,256,0.007452444069915348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,16384,4096,0.015073777900801765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,10240,0.04282933473587036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,5120,0.07854044437408447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,6144,0.02624977793958452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,2048,0.05526399943563673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,6144,9.778446621365017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,12288,0.04072177741262648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,1536,0.006226666685607698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,1536,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,65536,0.0991706649462382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,4096,0.11130932966868083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,1024,0.008793777889675563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,2560,0.02106933295726776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,256,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,256,0.008676444490750631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,64,0.006572444405820634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,3072,3072,0.033470223347345986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,16384,0.0936008890469869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,12288,0.015279110934999255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,1024,0.022672888305452134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,2560,0.031276444594065346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,16384,0.07441955804824829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,6144,6144,0.03839377893341912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,256,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,4096,65536,0.5332871013217503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,1024,0.019685332973798115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,128,12288,0.013992889059914483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,51200,0.05938755803638034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,4096,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,1536,0.025801777839660645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,16384,0.08167733086480035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,2560,65536,0.06545155578189425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,51200,0.6236569086710612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,1024,0.0032115555885765287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,51200,0.0764568911658393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,512,12288,0.037026666932635836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,7168,0.39396177397833926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,128,0.008288000192907121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,12288,0.030185778935750324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,3584,0.007999999655617608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,4096,0.012599999705950419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,8192,6144,0.03942844602796767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,6144,0.025066667132907446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,64,0.0015502222296264435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,3072,0.004559111264016894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,8192,1024,0.056399113602108426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,4096,0.026111111044883728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,51200,0.057743999693128795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,10240,3072,0.06367466847101848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,32,0.0022275555464956495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,256,0.00628977765639623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,2560,0.04702488912476433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,5120,0.042191998826132886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,51200,0.5276444223192003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,7168,512,0.006020444548792309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1024,2048,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,512,0.013448889056841532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,10240,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1536,768,0.019925332731670804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,65536,0.04184000028504265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,64,64,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,2048,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,768,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,3072,0.01928444372283088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,128,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,4096,0.08848000235027736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,512,0.007128000259399414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,5120,10240,0.1880835559633043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,65536,0.17235644658406576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2048,256,0.015792888071801927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,8192,0.4758453369140625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,1024,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,1024,0.009106666677527958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,1024,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,128,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,5120,0.008757332960764566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,512,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,3584,0.03872177667087979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,32,0.002149333349532551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,768,64,0.0018026667336622875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,6144,0.009639110830095079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2048,12288,0.13691733943091497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,512,0.023206222388479445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,4096,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,1024,0.02611200014750163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,51200,128,0.07196799914042155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,1536,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,7168,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,7168,0.060416890515221484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,8192,0.07184711429807875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,2560,0.08259466621610853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,3072,0.010030222435792288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,128,12288,0.013106666505336761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,64,16384,0.016116445263226826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,12288,0.900751961602105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,8192,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,2560,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,51200,0.04658222198486328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,128,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,512,0.045422222879197865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,1024,0.003630222131808599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,128,16384,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,6144,16384,0.033624887466430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2048,65536,0.04860533277193705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,5120,0.01260088880856832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,1536,0.010008888940016428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,1536,0.015248888068728976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,65536,0.2955697907341851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,65536,0.2640044424268934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,512,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,12288,7168,0.03808622227774726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,2560,1536,0.007270221908887227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,2560,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,10240,0.01259022206068039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,256,0.014848889576064216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,4096,0.010662222074137794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,256,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,768,0.023002665903833177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,1536,0.02947910957866245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,10240,0.037232001622517906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,3072,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,2560,0.06472000148561266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,5120,64,0.005704888867007361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,1024,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,2048,0.007286222444640265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,1024,0.03427911135885451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,128,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,8192,0.009024889104896123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,64,0.011095110740926532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,10240,768,0.0041288890772395665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,2560,0.013060444758998023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,10240,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,2048,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,16384,2560,0.07398666938145955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,2560,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,2048,0.016526222229003906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,2048,0.0042035554846127825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,7168,0.07650488615036011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,6144,0.056779556804233126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,768,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,51200,0.11975110901726617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,16384,3072,0.03527377711402045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,16384,0.009088888764381409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,2560,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,5120,5120,0.054027554061677724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,7168,0.5373999807569716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,65536,0.5168515311347114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,2560,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,10240,0.10874488618638779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,12288,3072,0.015944888194402058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,64,0.005965333431959152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,7168,0.040948443942599826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,128,0.03331289026472304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1024,7168,0.03534400131967332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,10240,0.43077778816223145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2048,512,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,4096,0.010609777437316047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,2560,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,16384,0.03948444459173415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,128,0.015449777245521545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,32,0.036878221564822726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,65536,0.3874071174197727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,8192,0.012111999922328524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,65536,1536,0.14989066123962402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,2560,0.01237155579858356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,5120,0.10956088701883952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,768,0.017903111047214933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,51200,2.06551382276747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,10240,12288,0.0410444438457489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,65536,768,0.032840887705485024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,51200,0.07391377952363756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,51200,7168,0.0682471129629347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,2048,0.0347635547320048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,8192,0.013226666384273104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,768,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,2560,0.032009777095582746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,768,0.012451555165979596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,3584,0.09203288952509563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,8192,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,2048,0.033775111039479576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,768,6144,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,2560,0.007983110845088959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,1024,0.011711999773979187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,65536,0.8903822369045682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,10240,0.16833155685000947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,256,32,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,1024,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,3584,0.12029066350724961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,65536,0.1391786601808336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,16384,0.043086220820744835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,16384,0.44742488861083984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,6144,2560,0.008767999708652496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,8192,0.007813333637184566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2560,16384,0.020276443825827706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,5120,0.014143110977278816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,2048,0.0595466693242391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,768,256,0.005626666463083691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,2560,0.013091555900043912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,6144,0.012011555333932241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,16384,0.5498604244656032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,3072,0.011360888679822287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,128,0.010112889111042023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,12288,0.0166640000210868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,768,3584,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,12288,6144,0.024096889628304377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,4096,0.07890133062998454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,51200,16384,0.27917154630025226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,65536,0.05307288964589437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,4096,65536,0.6163004239400228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,768,0.018740443719757926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,4096,0.845272011227078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,512,0.11635643906063503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,16384,0.04929600159327189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,512,0.0047608887155850725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,3584,0.039234668016433716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,512,0.01757333344883389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,512,2048,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,10240,0.21693244245317247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,6144,0.2508764531877306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,32,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,16384,0.04883289006021288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,4096,0.02530133393075731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,51200,0.3395084540049235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,1024,0.027620444695154827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,7168,6144,0.15574310885535345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,12288,1024,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,2048,0.024579554796218872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,5120,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,12288,0.016640888320075143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,512,0.026317333181699116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,2048,0.004550222307443619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,16384,0.032700445916917585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,128,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,4096,0.006924444602595435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,2560,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,65536,1.7931075625949437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,4096,0.03577866819169786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,16384,0.031883554326163396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,12288,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,6144,512,0.0028293333533737394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,64,0.0057653333577844836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,2560,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,32,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,3072,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,10240,0.011712888876597086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,1024,0.014906666345066495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,7168,0.01644355555375417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,5120,0.0515288876162635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,256,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,64,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,12288,0.2572266790601942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,6144,0.02780710988574558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,4096,768,0.011880888707107969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,5120,1.383671972486708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,65536,5120,0.14394756158192953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,8192,0.0622311102019416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,7168,0.011981333295504252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,12288,7168,0.020668443706300523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2560,256,0.016557332542207506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,7168,1.476837370130751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2048,256,0.018518222702874076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,128,2048,0.004631999880075455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,2560,0.04429422153366936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,3584,0.07493955559200711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,512,0.011058666639857821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,1536,0.059839997026655406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,1536,0.014078223043017916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,5120,0.38870400852627224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,2048,0.01531644496652815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,2560,0.02337155573897892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,3072,0.026714666022194758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,128,51200,0.12984089056650797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,1024,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,12288,0.055204444461398654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1536,6144,0.031298667192459106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,768,0.01738755570517646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,128,4096,0.03631555702951219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,128,0.002136888810329967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,65536,12288,0.19070933924780953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,32,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,51200,3584,0.03564000129699707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,768,0.00703111125363244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,1024,0.013759111364682516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,2048,0.006962666908899943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,3072,0.01000711073478063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,65536,2.032080968221029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2560,3584,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,4096,128,0.0019697778754764134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,1536,0.14927821689181858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,8192,0.17787467108832467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,6144,0.031792001591788396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,1536,0.020232889387342665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,12288,0.031567109955681696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,768,0.019741333193249173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,6144,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,2048,0.020968889196713764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,2048,0.023917333947287664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,128,1536,0.012267555627557965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,65536,0.08910399675369263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,2048,0.01276711126168569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,51200,512,0.046985778543684215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,10240,0.012736000120639801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,65536,0.9690435197618273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,6144,0.07193155421151055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,7168,0.037123554282718234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,8192,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,16384,0.016121778223249648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,7168,0.2843466599782308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,16384,0.021513778302404616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,256,0.009373333719041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,4096,0.1869893338945177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,768,0.01865333318710327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,64,0.0070168889231152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3584,4096,0.0090844440791342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,16384,0.1444000005722046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,1536,0.023714666565259297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,128,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,768,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,2048,0.01313333296113544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,7168,16384,0.028019554085201685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,32,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,10240,768,0.010991999672518836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2560,1024,0.014146667387750415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,3584,0.04474044508404202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,6144,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2560,65536,0.06443377998140123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,256,0.10258933570649888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,32,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,2048,3072,0.08380177948209976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,8192,51200,0.12955021858215332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,65536,5.66103278266059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,4096,0.01960266629854838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,1024,0.002883555574549569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,16384,0.12045599354637994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,64,8192,0.0107360002067354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,2048,0.009741333623727163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,768,0.008274666965007782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,4096,0.01316177762216992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,6144,0.008632000121805403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,6144,0.14930489328172472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,128,0.005059555586841372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,768,0.0047022220161226065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,2048,0.022986667023764715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,3584,0.3230240080091688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,6144,0.013213333156373767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,4096,0.02383022175894843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,5120,0.01105955574247572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,2048,0.011898666620254517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,768,16384,0.07714844412273832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,768,0.0025519999778932994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,4096,65536,0.3319359885321723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,8192,0.022656889425383672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2048,1536,0.016190222567982145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,2560,0.04210577739609612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,7168,0.04569955666859945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,3584,0.43378488222757977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,51200,0.08465511269039577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,128,65536,0.1757368908988105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,3584,0.008377778033415476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,32,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,4096,2560,0.1841964456770155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,512,1536,0.10968177848392063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,16384,2.2572248246934676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,768,6144,0.02472177810139126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,768,0.0026444445053736367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2560,2048,0.006747555401590135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,3584,0.00831822223133511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,6144,768,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,10240,0.02982577681541443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,256,0.009995555712117089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,512,65536,0.13151910569932726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,12288,0.012843555874294706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,1536,0.019888000355826486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,768,65536,0.06242577897177803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,65536,1.6471279991997614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,768,0.004192000048028098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,6144,0.07073422273000081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,8192,0.03116977877087063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,16384,6144,0.01943111088540819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,4096,0.01260355528857973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,65536,0.10149777597851223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,3584,0.01622222198380364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,768,51200,0.5348079999287924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,768,4096,0.06067022350099352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,16384,1.338823954264323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,7168,0.045279999574025474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,5120,3072,0.02994222111172146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,64,0.002536888958679305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,1536,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,768,0.05052711235152351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,5120,51200,1.1451519860161674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,3072,0.011057777537239922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2048,256,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,128,32,0.0017893332988023758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,12288,0.10795644256803726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,32,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,2560,0.007073777417341868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,64,0.005888888819350137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,2560,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,4096,0.04194133480389913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,10240,0.10747644636366104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,2048,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,3584,0.010694222317801582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,128,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,2560,10240,0.0384897788365682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,256,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,32,512,0.002130666747689247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,10240,0.1594933271408081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,5120,0.016695110334290397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,6144,12288,0.14879466427697077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,51200,0.08057333363427056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,128,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,8192,1536,0.006991110742092133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,128,1024,0.002946666752298673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,256,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,128,0.014164444473054675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,1024,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,7168,5120,0.017273777061038546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,3584,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,1536,0.009591110878520543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,2560,0.008738666772842407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,2560,0.14492089218563503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,7168,512,0.033033778270085655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,512,0.013405332962671915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,12288,512,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,8192,0.11573778258429633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,3584,0.05141244331995646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,6144,0.005210666606823604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1024,1536,0.017092444830470614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,768,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,12288,4096,0.02345955537425147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,51200,0.05504444572660658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,4096,1024,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,128,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,51200,0.05855911307864719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,8192,0.09877155224482219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,128,0.006747555401590135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,6144,0.005976888868543837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,32,0.0032088889016045462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,32,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,3584,0.7949075698852539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,128,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,3584,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,512,0.009366222553782994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,64,0.007319110963079665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,6144,0.1201288832558526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,3072,0.007670222057236566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,12288,0.1138497723473443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,3584,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,4096,1.066688855489095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,2048,0.7638417879740397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,6144,0.0446995562977261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,2048,64,0.02568622264597151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,51200,0.17229066954718697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,2560,0.012649777862760754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,4096,0.07606310976876153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1536,7168,0.01957155598534478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,256,0.03523733218510946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,8192,0.010406222608354356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,3584,3072,0.008922666311264038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,51200,1.2733209398057725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,65536,0.07382488913006253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,10240,0.039579557047949895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,6144,0.021935110290845234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,8192,0.020045333438449435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,2560,0.019451555278566148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,10240,0.04422488808631897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,2560,0.022642667094866436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,768,0.04121511181195577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,3072,0.007867555651399825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,768,0.011929777761300405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,32,8192,0.021914665897687275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,64,0.07112444771660699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,3072,0.03789600067668491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,51200,0.061358220047420926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,5120,10240,0.0490497781170739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,65536,51200,0.5058639844258627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,256,0.01577422188387977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,5120,0.006623111251327727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,7168,0.011002666420406766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,1536,0.021965333157115515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,64,12288,0.016131555040677387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,512,0.013932444983058505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,1536,0.011767999993430244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,768,256,0.008344888687133789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,2048,0.08685866991678874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,6144,0.07644710938135783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,5120,0.011063111325105032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,10240,5120,0.015885333220163982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,64,0.15330666965908474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,4096,0.019747555255889893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,128,16384,0.016247999336984422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,8192,12288,0.140209780799018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,1024,0.0057839999596277876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,5120,0.88537687725491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,6144,0.01185066666867998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,64,0.0021555556191338436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,2048,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,16384,0.01620533400111728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,16384,0.28317512406243217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,65536,0.40716534190707737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,2048,6144,0.03643200132581923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,65536,0.07152177890141805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,4096,0.01811911165714264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,1536,12288,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,3072,0.006272000157170826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,128,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1024,51200,0.12373956044514973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,3584,7168,1.2085226906670463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,1024,0.005962666538026597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,256,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,16384,4096,0.16241689523061117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,3072,0.007335999773608313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3584,65536,0.1785191165076362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,5120,0.008982222113344405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,12288,1024,0.007138667007287343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,256,0.11391022470262314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,128,0.001826666709449556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,16384,0.04638933473163181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,10240,0.03783911135461595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1536,65536,0.06696977880265978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,512,0.014531556102964612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,65536,0.061159114042917885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,16384,0.22193866305881074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,32,0.0069191112286514705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,256,6144,0.00518844442235099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,3584,0.025378665990299646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,16384,0.8864755630493164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,256,0.020853334003024634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,6144,0.026008889079093933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,16384,0.01371911085314221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,3584,0.04616711205906338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,51200,0.523089779747857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,51200,2560,0.331949339972602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,3072,0.010042666561073726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,4096,65536,1.0838168462117512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,5120,0.02859555681546529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,2048,0.01742844449149238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,64,0.00794933322403166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,256,1024,0.009646221995353699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,3072,0.37384621302286786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,512,16384,0.057985776000552706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,16384,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,12288,10240,0.03142133355140686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,65536,768,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,1536,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,1024,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,6144,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,6144,0.028920888900756836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,12288,0.2944942315419515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,8192,0.3016008800930447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,16384,12288,0.07519200113084581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,4096,0.11204355292850071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1024,2560,0.006599111275540457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,768,3584,0.008049777812427944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,1536,1536,0.018410666121376883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,51200,512,0.011041777829329172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,256,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,128,0.01369511087735494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,768,0.006436444405052397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,16384,128,0.020240889655219186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,3072,0.01739377776781718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3584,16384,0.02808088726467556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,1024,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,5120,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,4096,0.120798216925727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,8192,0.14768355422549778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,12288,0.27818044026692706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,128,51200,0.13095288806491429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,5120,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,12288,0.41452977392408585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,64,1024,0.01036977767944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,6144,0.024615110622511968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,4096,0.025620443953408137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,65536,0.07052799728181627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,6144,0.06467644373575847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,12288,512,0.03808800048298306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,1536,0.016201777590645682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,10240,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,12288,0.08423733048968846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,8192,0.022491556074884202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,10240,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,32,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,51200,0.12016977204216851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,5120,5120,0.02420444455411699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,1024,0.13345066706339517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,256,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,1536,0.02647466626432207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,1536,768,0.02572088936964671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,12288,0.01622222198380364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,3072,0.007495111061467065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,512,0.0032213332338465583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,51200,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,16384,0.014601776997248331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,12288,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,64,0.002472888885272874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,12288,0.37109687593248153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,128,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,2560,0.01754577789041731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,768,0.09763644138971965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,6144,0.02296977738539378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,10240,5120,0.07838310798009236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,65536,13.423631456163195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,128,0.024705777565638225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,8192,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,3072,0.00996800015370051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,2048,0.008954666554927826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,12288,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,3072,0.0702586637602912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,2048,0.05783200263977051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,8192,0.03338133295377096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,2048,0.0075902218619982404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,12288,0.08078044652938843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3072,256,0.016163556112183463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,3072,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,32,0.005516444643338521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,5120,0.016971555021074083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,10240,0.2131751113467746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,16384,0.04559288753403557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,1024,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,8192,2048,0.008513777620262569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,12288,1.8154036204020183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,4096,0.014627556006113688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,768,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,4096,0.017942221628295053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,1536,0.027985778119828966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,512,0.01751733322938283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,128,0.0038986665507157645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,8192,0.05794755617777506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3584,4096,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,8192,0.012803555362754397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2048,8192,0.0410426656405131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,51200,0.05985866652594673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,7168,4096,0.012288889123333825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,128,0.0031617778456873367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,5120,0.03108888864517212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,65536,0.06265333626005384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,64,0.011726222104496427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,2048,0.006392000036107169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,32,0.00518488883972168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,5120,0.017094223035706412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,3072,0.00830844458606508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,7168,0.26573067241244847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,4096,8192,0.016399110356966656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,10240,0.024872000018755596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,6144,0.049795554743872754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,2048,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,12288,0.09931555721494888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,5120,0.010363555616802638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,1024,0.020957332518365648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,51200,12288,3.941448211669922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,1536,0.008915555146005418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,4096,3584,0.007865777446164025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,32,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,1024,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,2048,0.1120871173010932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,4096,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,16384,0.0878408882353041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,6144,0.03259911139806112
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,1024,0.00444533344772127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,51200,2560,0.027338667048348322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,51200,65536,16.573577880859375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,256,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,4096,0.045010666052500405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,3584,0.009009777671760982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,3584,0.020888888173633154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,512,2048,0.006664888726340399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,12288,2560,0.02946044339074029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,128,0.005233777893914117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,7168,51200,0.41950755649142796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,10240,0.015981333123313058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,512,0.005734222216738596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,8192,0.02462844385041131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,7168,0.12019199795193142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,4096,0.008607111043400234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,2048,0.012819555898507437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,256,0.016884444488419425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,10240,0.2223893271552192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,8192,0.024005333582560223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,768,0.05592799848980374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,10240,0.042167110575569995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,65536,0.04965510964393616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,32,128,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,16384,0.23103645112779406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,6144,0.00923644420173433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,3072,0.01034044474363327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,512,256,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,16384,0.034225778447257146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,3584,0.011056000159846412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,7168,0.017893332574102614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,128,8192,0.07483555873235066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,6144,0.33500533633761936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,10240,8192,0.23223998811509874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,65536,1.2160275777180989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,2560,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,3072,0.6079324616326226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,7168,0.07691911194059584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,3072,0.015832889411184523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,7168,4096,0.010103111465771994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,16384,1.514767116970486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,6144,0.019316444794336956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,2560,0.02106933295726776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,8192,512,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,16384,0.06613866488138835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,2560,0.04795822170045641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,5120,0.018954666124449838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,2048,0.0750711096657647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,4096,0.03606933355331421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,256,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,2048,0.006096000058783426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,1536,0.006691555596060223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,16384,0.1099511120054457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,1024,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,2048,0.022271111607551575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,51200,0.009208888643317753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,3584,0.006599999964237213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,2048,0.013354666531085968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,6144,0.005625777774386936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,256,0.016522667474216886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,1024,0.003601777884695265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,5120,1024,0.005667555663320754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,16384,0.02132266594303979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,512,0.0221395558781094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,51200,0.0797742207845052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,2048,0.008760888543393876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,1024,0.019689778486887615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,2560,0.01033333357837465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1024,1024,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,8192,6144,0.14569600423177084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,2560,0.01016088906261656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,6144,0.17568533950381807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,4096,0.4674764209323459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,65536,16384,1.1437075932820637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,64,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,4096,0.02696888811058468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,12288,2560,0.052744001150131226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,64,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3584,3072,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,128,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,1024,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,768,7168,0.03011022342575921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,2048,0.014807111687130399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,64,2560,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,7168,0.028315554062525432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,1024,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,1024,0.018194667167133756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,16384,0.0162808895111084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,768,0.01754666699303521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,3072,0.018246221873495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,10240,0.17268978224860299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,3584,0.025037333369255066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,7168,0.008072000410821702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3072,5120,0.05236177643140157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,12288,0.4024853441450331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,16384,0.02682133350107405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,5120,0.03218133250872294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,10240,0.014832889040311178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,6144,8192,0.05643377701441447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,512,64,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,256,5120,0.008373333348168267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,512,16384,0.05046400096681383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,7168,0.026775999201668635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,65536,0.10673155387242635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,2048,0.02975377771589491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,1536,6144,0.043959998422198825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,768,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,512,0.0074471111098925276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,4096,2048,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2560,1536,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,1024,0.008525333470768398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,65536,256,0.13623466756608751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,4096,0.02942044536272685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,2560,0.007012444237867991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,2048,0.013395555317401886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,64,512,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,65536,0.027622222900390625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,3072,0.0140657772620519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,3584,128,0.0019146667586432563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,3072,4096,0.025361778007613286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,512,0.005377777748637729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,1536,0.006769777586062749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,7168,0.012719111310111152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,5120,0.01682400041156345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,768,8192,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,7168,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,5120,0.02104533380932278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,3072,0.08401511112848918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,65536,2560,0.028169777658250596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,7168,0.11026310920715332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,1024,0.2861057652367486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,65536,0.32733334435356987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,5120,0.010659555594126383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,768,0.018537777993414137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,12288,1536,0.028563555743959215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,51200,0.1136560042699178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3584,12288,0.03358577688535055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,1536,0.047170665529039174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,2048,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,128,0.0031999999450312722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,4096,256,0.020724443925751578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,4096,3584,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,8192,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,2560,0.0192284451590644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,3584,0.011064888702498542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,3584,0.0815253323978848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,32,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,5120,0.1457200050354004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,16384,0.09640177753236558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,256,128,0.002163555473089218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,5120,0.016207999653286405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3584,16384,0.03621600071589152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,768,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,1536,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,32,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,8192,0.40527645746866864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1536,3072,0.007787555456161499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,4096,0.0429075558980306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,64,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,8192,0.09582666556040446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,512,16384,0.011619555453459421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,16384,0.025432000557581585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,32,0.0019964445382356644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,5120,6144,0.055439998706181846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,768,0.04335022303793165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,2560,0.020267556111017864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,8192,0.017127111554145813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,64,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,65536,0.44421688715616864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3072,3072,0.007743111087216272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,5120,0.021518222159809534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,12288,0.03433777888615926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,7168,16384,0.04003644320699904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,12288,16384,0.4017973211076524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,10240,0.15597066614362928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,2560,0.026687999566396076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,128,0.015122666954994202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,3584,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,65536,0.39168087641398114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,8192,10240,0.04414577616585625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,768,0.015606222881211175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,768,0.015522667103343539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,2560,0.03615377677811517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,12288,0.30199021763271755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,512,0.0058702222175068325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,3584,0.02399377855989668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,12288,0.02701777716477712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,1024,0.1111840009689331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,7168,2560,0.025967111190160114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,65536,0.45425955454508465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,2048,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,3072,0.023376888699001733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,3072,0.009944000177913243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,3072,0.006064888917737537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,2048,512,0.008061333662933773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,12288,0.03479111194610596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,2560,0.01993333299954732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,128,768,0.012783110969596438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,2560,0.13936266634199354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,128,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,8192,0.03805511196454366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,65536,1536,0.07902310954199897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,6144,0.03547199898295932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3072,5120,0.09099555677837795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,3584,3584,0.021339555581410725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,10240,65536,6.293988545735677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,16384,0.020506666766272653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,3072,0.008410666551854875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,128,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,5120,0.008266666697131263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,7168,0.212973329755995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,12288,4096,0.03595022360483805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,256,0.012200888660218982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,4096,0.004201777693298128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,3072,0.007686222592989604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,1024,0.00628355559375551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,12288,0.03363822235001458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,3584,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,4096,0.04499466551674736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,3072,0.007110222346252865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1024,4096,0.014415999253590902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,512,0.06785688797632854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,51200,0.23538488811916777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,65536,0.1679848829905192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,6144,0.01368622233470281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,3072,0.030930668115615845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,2048,0.024303999212053087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,3072,0.011074666347768573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,4096,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,5120,0.18881421618991426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,128,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,3584,0.07479733228683472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,128,64,0.007313778003056844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,1024,0.0140684445699056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,1024,0.02625688910484314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,128,256,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,6144,0.23100177447001138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,7168,0.0064693333374129395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,5120,1.8855056762695312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,2048,0.008656889200210571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2048,3584,0.0229759994480345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,10240,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,51200,51200,0.4935715463426378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,512,0.021290666527218286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,256,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,8192,0.03324266605907016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,512,0.021244444780879553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,5120,0.01573955516020457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,10240,0.3063253296746148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,512,0.01754844519827101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,6144,0.009583999713261923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,6144,1024,0.11514400111304389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,768,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,12288,51200,0.19813777340783012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,128,6144,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,2048,0.014584889014561972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,2560,0.03752711084153917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,3072,0.009683555199040307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,256,1024,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,65536,0.1363404459423489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,4096,0.03003644280963474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,256,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,3072,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,1536,0.01514400045077006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,51200,4.244619581434462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,2048,0.026713778575261433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,65536,0.09133688608805339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,4096,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,4096,8192,0.013056889176368713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,7168,1024,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1536,3072,0.1133120059967041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,10240,0.017791999710930716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,256,0.008597333398130205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,4096,7168,0.028917332490285236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3072,12288,0.06137510803010729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,1536,0.0673617786831326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,1536,12288,0.07713333103391859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,2048,0.005327111317051782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3584,1536,0.0137155552705129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,65536,0.05793333053588867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,128,768,0.013932444983058505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,32,0.0018231111268202465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,2048,0.004798222333192825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,64,0.0020826665891541377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,256,16384,0.02642044425010681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,51200,1.756305800543891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2048,51200,5.363638983832465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,51200,0.14030933380126953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,12288,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,5120,0.11315911346011692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,51200,7168,0.6442329088846842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,256,0.02784622377819485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,768,0.01955111159218682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,7168,5120,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,8192,0.06933955351511638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,6144,0.01440711153878106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,16384,0.06736266613006592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,2048,0.007697777615653143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,4096,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,128,0.013032889200581444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,12288,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,768,1536,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1024,6144,0.029155555698606703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,2560,0.032524443335003324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,4096,0.013665777941544851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,64,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,1536,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,3584,2.8377172682020397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,2048,1024,0.015499555402331881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,2560,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,4096,0.007731555236710443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,128,0.019448000523779128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,12288,0.13765777481926814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,12288,2.1267358991834855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2048,1536,0.011843555503421359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,1024,0.0031386667655573953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,1024,0.02640177806218465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,2048,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,768,512,0.05071555574735006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,51200,1.3614151212904189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,3584,0.01316088851955202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,3072,0.008948444492287105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,128,0.011332444846630096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1024,5120,0.016881777180565726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,10240,0.01717422240310245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,6144,2048,0.07121421893437703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,512,0.00564444437623024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,512,0.012267555627557965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,3584,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,51200,0.062422222561306424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,10240,0.024383111132515803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,16384,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,3584,0.017981333865059745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,6144,0.0529457794295417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,512,0.03176711002985636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,65536,1.1264186435275607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,12288,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,64,0.00627377794848548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,51200,65536,0.5590151151021322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,768,10240,0.012795555922720166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,1536,1536,0.0038702223036024305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,3072,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,1536,0.17879111236996123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,65536,0.13901244269476995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,2048,2048,0.0064897777305708984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,1024,12288,0.06922666894065009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,768,0.01685244507259793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,65536,0.31307734383477104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,12288,1.4042995240953233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,4096,0.009015999734401703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,16384,0.018389332625601027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,4096,768,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,8192,0.12928267319997153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,3584,0.009656888743241629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,8192,0.07127910852432251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,7168,0.11202755239274766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,65536,0.5282017919752333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,1024,0.017551110850440133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,128,0.009639999932712978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,3584,1.138811535305447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,256,0.007708444363541073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,256,0.009775111244784461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2048,32,0.01812533371978336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,65536,0.10761600070529515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,6144,0.010671111444632212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,12288,0.011779555843936073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,7168,0.03665333324008518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,32,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,64,0.0019057778020699818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,3584,0.04885955651601156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,6144,0.06520089175966051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,10240,0.07444710863961114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,51200,0.056214219994015165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,8192,7168,0.042591998974482216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,1536,0.006688000013430913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,768,0.023759111762046814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,1536,0.007599111232492659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,768,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1024,2048,0.006903999795516332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,2560,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,65536,0.6988142331441244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,2560,0.007606222397751278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3584,2560,0.03642400105794271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,256,0.004996444616052839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,3072,0.011911110745535957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,512,0.00249777775671747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,7168,32,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,5120,2560,0.05156088868776957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,12288,0.10877155595355564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,4096,0.3443297810024685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,3072,0.02360444433159298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,12288,512,0.008596444295512306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,65536,0.7287022272745768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,3072,0.011343999869293638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,2048,0.04042933384577433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,32,65536,0.020575111111005146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,64,32,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,16384,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,12288,0.06186311112509834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,8192,0.01071466671095954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,256,0.01482399966981676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,3072,0.014953777194023132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3584,2048,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,51200,7168,2.3987714979383683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,512,0.03669333457946777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,51200,0.30551287862989635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,12288,0.0396382212638855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,51200,0.4753635724385579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,8192,0.035812444157070585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,8192,0.039311110973358154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,32,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,10240,0.04149688945876227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,1024,0.012455999851226807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,1536,0.0063324446479479475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,256,0.016385777129067313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,512,512,0.003516444315512975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,7168,0.01532977819442749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,3584,0.8618942366706001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,7168,0.5248471366034614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,64,1536,0.0035146665241983202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,3584,0.025947555899620056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,5120,0.041377776198916964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,4096,0.010636444720957013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,6144,0.03758666581577725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,6144,0.015107555521859063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,768,0.018089777893490262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,10240,0.011364444262451597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,6144,0.0690586699379815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,512,0.007071110937330458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,1024,0.016027554869651794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,6144,0.03407288922203912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,768,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,16384,0.029143999020258587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,12288,0.014475555883513557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,256,0.015235554840829639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,4096,0.011376000112957425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,51200,0.07854666974809435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,512,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,32,0.001818666648533609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,8192,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,1536,0.048671109808815845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,16384,0.01017599966790941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2560,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,768,0.0028737777223189673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,6144,0.06154488854938083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,8192,128,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,7168,0.04412533177269829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,256,0.006959111326270633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,10240,0.025272889269722834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,4096,0.02418400016095903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,2048,0.04340799980693393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,2048,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,65536,0.1725484397676256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,64,4096,0.0035706667436493766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,128,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,4096,0.028312000963422988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,3584,0.03147911032040914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,512,0.016512889001104567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,3072,0.03559199968973795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,128,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2560,2048,0.026328000757429335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,65536,8192,0.11203111542595758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,5120,0.00941422250535753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,10240,5120,0.04336533281538221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,128,0.010856889188289642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,65536,0.4383991029527452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,4096,0.05243644449445936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,32,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,512,0.0057839999596277876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,512,0.015471999843915304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,2560,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,16384,0.03677511215209961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,64,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,10240,0.026855111122131348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,6144,6144,0.016146666473812528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,8192,0.019309333629078336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,5120,0.023754666248957317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,128,0.002159111201763153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,2560,0.04615555538071526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2560,128,0.03516799873775906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2560,7168,0.033287998702791
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,3584,0.01868533260292477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,7168,0.105076445473565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,1536,0.0325528879960378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,7168,65536,2.540321773952908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,10240,0.06542222367392646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,7168,0.04580444428655836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,512,0.19335733519660103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,8192,0.05017066664165921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,4096,1024,0.015196444259749519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,5120,0.03347555465168423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,65536,0.5697191026475694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,7168,0.8715759913126627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,64,1024,0.007279111279381647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,3584,0.47443379296196836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,2048,0.006712889091836081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,8192,7168,0.040786667002571955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,64,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,65536,12288,2.749856948852539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,768,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,3584,0.018368888232443068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,16384,0.18210577964782715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,2048,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,512,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,512,12288,0.01752000053723653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,7168,0.03535200158754984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,512,0.00684799998998642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,512,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,2048,0.012474666866991254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,768,0.015152888165579902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,5120,0.019367110398080613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,2048,0.010447999669445885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,1536,0.012441777520709567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,5120,12288,0.024749333659807842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,512,0.002526222210791376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,2048,0.007979555262459649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,256,0.0025395556456512874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,64,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,4096,0.007910221815109253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,6144,0.021194666624069214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,51200,0.19017333454555937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,2048,0.08526311318079631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,5120,0.01753422286775377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,8192,0.07696000072691175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,32,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,7168,0.01888533267709944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,12288,0.030782222747802734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,256,0.01787644459141625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,65536,2048,0.07937244574228923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,51200,0.06756177875730726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,1536,0.13494222693973118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,2048,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,10240,0.0070364442136552595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,3072,0.06639911068810357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,2560,0.004861333303981357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,6144,0.10753599802652995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,5120,768,0.013359111216333179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,2560,0.005456888841258154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,3584,0.04812888966666328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1024,4096,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,7168,0.03367999858326382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,4096,51200,0.05784533421198527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,512,0.02718755602836609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,3072,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,2560,0.01860000027550591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,3072,0.007552000383536021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,1024,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,1536,0.007242666350470648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,512,1536,0.006594666590293248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,128,0.021928888228204515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,64,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1536,51200,0.04740266667471992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,256,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,5120,0.01235466698805491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,10240,2048,0.01071111112833023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,6144,3584,0.013906665974193148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,10240,0.12898133860694036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,10240,0.24094133906894258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,65536,0.05979377693600125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,16384,0.10662400060229832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,5120,0.09603822231292725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2048,7168,0.03274222215016683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3584,2048,0.007378666765160031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,51200,1536,0.20489954948425293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,2560,0.004898666507667965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,32,0.014741332994567024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,8192,0.015832000308566622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,2560,0.14204800128936768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,2048,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,3584,0.018797333041826885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,8192,0.06406044297748141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,128,0.16023733880784777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,256,0.026022222306993272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,512,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,16384,0.32035112380981445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,65536,3584,0.8503377702501086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,512,0.01997422178586324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,512,1536,0.008376888930797577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,8192,0.06092444393369886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,7168,3584,0.010435555544164447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,3584,0.014314666390419006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,3584,0.057017776701185435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,1536,0.023051554958025616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,4096,0.008001777860853408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,1024,0.05303288830651177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,12288,4096,0.01647822227742937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,51200,0.0957546697722541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,12288,0.035318222310807966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,768,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,768,0.03239999877081977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,10240,0.023864888482623633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,65536,2.563607109917535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,65536,0.09962222311231826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,5120,0.01312444441848331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,128,512,0.015512888630231222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,5120,0.01293955577744378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,6144,0.013631111217869652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,512,0.004845333182149463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,8192,0.016543999314308167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,128,0.005281777845488654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,10240,0.01676977839734819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,128,0.018986667195955913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,2048,0.09825066725413005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,12288,0.0306693348619673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,3584,1024,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,4096,0.0582266648610433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,1536,0.02942755487230089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,2048,0.02684088879161411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,5120,0.6616506576538086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,256,16384,0.021293333835071985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,6144,0.026280888252788123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,1536,0.007544889218277401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,512,512,0.015805333852767944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,3584,0.011340444286664328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,51200,3072,0.09151466687520345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,5120,0.019660444723235238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,32,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,2048,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,16384,5120,0.017450667089886136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,1536,0.08192000124189588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,5120,0.009373333719041612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,65536,0.07142666975657146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,5120,0.07451911105049981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,12288,0.05939022037718031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,128,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,16384,0.1863395505481296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,128,65536,0.04872533347871569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,768,0.031981332434548274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,3584,0.011922666596041786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,6144,0.983459578620063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,8192,0.07849333021375868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,64,12288,0.06251821915308635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,6144,0.005232000102599462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,4096,0.016308445069524977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,6144,19.072474161783855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,6144,0.32061510615878636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,3072,0.033210668298933245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,16384,0.04598755637804667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,4096,3584,0.030568887790044148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,8192,0.3499751091003418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,5120,0.23914400736490884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,65536,0.7914960119459363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,2048,0.02335022224320306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,512,0.12201333045959473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,256,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,65536,0.08358399735556708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,64,512,0.002467555511328909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,10240,0.04333422250217862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,768,16384,0.02176355487770504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,1536,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,512,0.03240622083346049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,51200,0.05083644390106201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,7168,0.29163821538289386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,256,0.011194667054547204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,10240,0.01203733351495531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,4096,0.01776355504989624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,2048,0.01665244499842326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,768,0.015616888801256815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,128,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,512,0.011389333340856763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,65536,0.0599457754029168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,1024,0.07497688796785143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,3072,0.06169778108596802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,51200,0.05180799961090088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,768,3072,0.02158044444190131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,512,0.002536000063021978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,3584,0.09055199888017441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,2560,3072,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,51200,12288,0.6989066865709094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,768,0.003168888803985384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,1024,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,2560,0.006241777704821692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,8192,0.8992915683322482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,8192,256,0.014820444915029736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,2560,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,12288,0.0439644455909729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,256,2048,0.0069439998931354946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,512,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,8192,0.03955555624432034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,51200,0.88409423828125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,512,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,8192,0.06199200285805596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,7168,0.026622222529517278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,512,0.005361777626805835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,64,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3584,10240,0.017887999614079792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,65536,0.8548453119066027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,2560,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,128,0.014057776994175382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,4096,128,0.009384888741705153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,128,0.0024497778051429326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,256,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,3072,0.012534221841229333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,7168,0.030840890275107488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,8192,0.030461334519916113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,7168,0.04483911063936022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,3584,512,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,768,0.004190222256713443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,1536,0.006660444455014334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,128,0.002850666642189026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,128,0.013463999662134381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,6144,0.02570755614174737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,1024,0.018583110637134977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3584,2048,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,2048,0.017826666434605915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,12288,0.5604533089531792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,2048,0.022234666678640578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,7168,0.16317689418792725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,7168,0.009106666677527958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,10240,0.04400355617205302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,16384,2.177398257785373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,5120,0.016900445024172466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,128,0.006155555446942647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,3072,0.01909244391653273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,51200,3584,0.20557867156134713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,2560,6144,0.3289155430263943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,6144,12288,0.21355288558536104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3584,16384,0.03935555617014567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,2560,0.008986666798591614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,64,0.005215999980767568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,768,32,0.009283555878533257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,768,0.004924444274769889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,3584,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,128,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,7168,0.028155555327733357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,7168,16384,0.43102489577399355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2560,2560,0.00960888879166709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,8192,8192,0.024482667446136475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,32,10240,0.017280000779363845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,6144,5120,0.13006488482157388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,64,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,3584,0.11047377851274277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,6144,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,512,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,768,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,512,65536,0.06389244397481282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,10240,0.13710221979353163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,7168,0.1240968836678399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,3584,0.008285333712895712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,8192,0.014816888504558139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,1024,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,768,0.002520888836847411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,768,1536,0.028564443190892536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,16384,0.09133955505159165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,12288,16384,0.460693359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,16384,0.08204711145824857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,1536,0.03380089004834493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,768,0.003526222374704149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,128,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,12288,0.1373573276731703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,6144,0.017217778497272067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,10240,0.03339111142688327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,7168,2048,0.014487110906177096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,256,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,4096,1024,0.018265777164035372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,1024,0.009847111172146266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,8192,0.03692444496684604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,32,0.025382222400771245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,2560,0.17620266808403862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2048,5120,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,32,3072,0.003850666599141227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,1024,0.009200000100665623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,5120,0.011744000017642975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,6144,0.05513422356711494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,256,0.020755555894639757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,65536,2560,0.03401422169473436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,256,0.014156444205178155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,10240,0.14796355035569933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,8192,3584,0.02544622288809882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,16384,5120,0.13114489449395075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,6144,0.6348969141642252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,2048,0.0750711096657647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,8192,0.093922668033176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,256,0.02717422280046675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,1536,0.008515555825498369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,3072,6144,0.010919999745157031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,8192,0.21546844641367593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,4096,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,128,128,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,7168,0.024969778127140466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,3072,0.7108968628777398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,4096,5120,0.022263111339675054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,256,0.016148444679048326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,16384,0.06844000021616618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1536,256,0.007634667058785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,512,2560,0.0096133334769143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1024,128,0.006629333313968446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,6144,16384,0.0432586669921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,256,0.042895999219682485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1024,768,0.003535110917356279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,768,3072,0.01627200014061398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,2048,0.5229404237535265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,3072,0.030758221944173176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,7168,0.1301457749472724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2048,6144,0.011238222320874533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,512,0.011583999627166323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,3072,0.012149333126015134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,10240,0.023229334089491103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3584,8192,0.01536622146765391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,10240,0.016470222009552848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3584,7168,0.10562666919496323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,32,8192,0.04052177733845181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,4096,0.016984888248973422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,512,0.00833155545923445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,16384,0.06480622291564941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,6144,0.18546044826507568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,12288,1.0040648778279622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,51200,0.9927946726481119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,65536,0.020228443874253165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,4096,0.013166222307417126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,6144,1024,0.044103112485673696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,10240,0.019668444991111755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,10240,0.03523555397987366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,2048,0.028441776831944782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,3584,0.0247013337082333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,1024,0.015607110328144498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,4096,0.02645155621899499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,768,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,3584,0.031139555904600356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1024,8192,0.01073333372672399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3072,12288,0.018075555562973022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,8192,0.06645422511630587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,1536,0.005583999885453119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,51200,0.05990133020612928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,128,2048,0.007318221860461765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,7168,4096,0.1288355588912964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,3584,0.09850311279296875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2048,6144,0.010034666293197209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,6144,1.7230933507283528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,768,256,0.009002666506502364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,12288,3584,0.2981111208597819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,10240,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,1536,7168,0.011800888511869641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,4096,0.03357777661747403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,5120,0.03751644492149353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,3072,6144,0.05320355627271864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,64,0.007206222249401941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,65536,0.43744977315266925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,3072,0.02093422247303857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,2560,0.023344889283180237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,6144,0.01961599952644772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2560,256,0.005240889059172736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,3072,0.015527110960748462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,51200,0.0607208874490526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,768,0.01588355501492818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,10240,0.20891200171576607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,8192,128,0.01851733360025618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,10240,0.01679733395576477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,1024,0.03400977783732944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,3584,0.007923555870850882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,768,0.013835555149449242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,7168,0.12511555353800455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,3072,0.021319111188252766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,512,0.05083022183842129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2560,12288,0.04944711261325412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,12288,0.36612534523010254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,7168,0.016677333248986136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,12288,65536,0.15591022703382704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,64,3584,0.01715555621518029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,2048,0.03015466531117757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,6144,0.046401778856913246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,2560,0.12929599814944798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,512,12288,0.03455200129085117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,768,0.00573333352804184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,4096,0.09392799933751424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,256,0.005205333232879639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,512,6144,0.019084443648656208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,7168,0.4482862154642741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,1024,0.007589333587222629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,768,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,16384,128,0.026538666751649644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,1536,0.05061599943372938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2048,3072,0.00943022221326828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,128,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,64,0.0025333333760499954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,512,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,16384,0.06713066498438518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,128,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,10240,8192,0.2586817741394043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3072,1536,0.03673955467012193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,51200,256,0.11754043896993001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,51200,51200,0.4948168860541449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,768,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,768,0.005736888696750005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,32,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,128,256,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,12288,23.024792141384548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,51200,0.0588408907254537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,256,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,7168,0.046169777711232506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,8192,0.28920356432596844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,6144,0.005950222412745158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,512,0.005979555762476391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,8192,0.30635200606452095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,12288,0.12851999865637884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,128,0.002191999927163124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,8192,0.7187280125088162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,7168,0.03481866584883796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,12288,0.034134222401512995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,65536,0.037794666157828435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,768,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,10240,0.18052889241112602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,2048,0.014863999353514777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,5120,0.04996177885267469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,512,0.007500444849332173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,8192,0.15769688288370767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,4096,0.09097688727908665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,256,0.00556355549229516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,7168,0.02200177808602651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,1536,0.006009777800904379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,10240,0.011900444825490316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,6144,256,0.009104889300134447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,16384,2048,0.031173331869973078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,32,0.0058871110280354815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,512,0.05124444431728787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,16384,0.06425511174731784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3072,3072,0.007532444265153673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,65536,0.48887909783257383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,32,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,51200,0.1160702175564236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,3584,0.015127110812399121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,4096,0.008464888566070134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,1024,0.014137778017255994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,7168,0.027787556250890095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,12288,512,0.02988977895842658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,2560,0.009689778089523315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,768,16384,0.0432204438580407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,12288,0.022425777382320825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,6144,0.006824000014199152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,16384,0.02652799917591943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,7168,0.7103688981797961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,3072,0.05025777882999844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,256,0.0026764445420768526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,4096,0.018458666072951425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,2048,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,7168,1.0154817369249132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,7168,0.5047564506530762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,65536,0.12369867165883382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,12288,0.03341955608791775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,3584,0.06791200240453084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,2048,0.019692444139056735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,1024,0.09023377630445693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,512,0.005389333185222413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,12288,0.12804799609714085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,6144,0.010452444354693094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,2048,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,8192,0.014389332797792224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,3584,0.09574577543470596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,3072,1024,0.011184888581434885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,65536,0.6819146474202474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,7168,4096,0.10352444648742676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,8192,0.05322755707634819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,10240,0.028260442945692275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,4096,0.030807110998365615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,3584,0.010056888891590966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,768,0.015837333268589444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,1536,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,256,4096,0.02032177812523312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,65536,5.950159284803603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,768,2560,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,65536,20.179228040907116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,768,0.023356444305843774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,7168,0.02181333303451538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,2048,64,0.0035582222044467926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,64,0.0035626664757728577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,2560,0.03066133459409078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,6144,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,512,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,768,0.01700088878472646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,12288,64,0.003571555432346132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,51200,0.04935377836227417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,2048,0.004553777890072929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,10240,0.031423999203575984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,1024,0.011723555624485016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,5120,3584,0.010703110860453712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,4096,0.03511733478969998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,64,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,12288,7168,0.3804017702738444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,2560,0.008549333446555668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,65536,256,0.12506400214301214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,128,0.0052382221652401825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,32,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,16384,0.38942045635647243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,12288,0.37419644991556805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,256,0.006529777828190062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,32,0.0021359999146726397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,768,6144,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,2560,0.01255555533700519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,3584,0.0964026649792989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,65536,0.0724124444855584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,3072,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,65536,0.215810669793023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,4096,0.02609688871436649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,65536,10240,0.11386133564843072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,16384,2048,0.21890221701727974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,8192,0.0651528901524014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,4096,0.012103999654452005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,64,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,256,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,256,0.02181422213713328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,16384,0.042262223031785756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,3584,512,0.00870577742656072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,7168,0.19508800241682264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,64,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,51200,3.311767154269748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,5120,51200,0.6762248675028483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,128,0.1656257841322157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,2560,0.007702222300900354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,3584,0.02141333288616604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,64,1536,0.00958933350112703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,1536,0.007972444097201029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,1024,0.01790488925245073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,512,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,2560,0.09812266959084405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,6144,0.09406133492787679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,3584,0.010022222167915767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3584,2560,0.041731556256612144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1024,10240,0.07051110929912992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,1024,0.0032017777363459268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,128,2048,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,32,0.006264888991912206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,3072,0.017350221673647564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,32,0.00701155596309238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,1024,512,0.01310222182008955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,6144,0.09826311137941147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,3072,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,2560,0.017616000440385606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,3072,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,768,0.012049777640236748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3584,2560,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,3584,0.08869955274793838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,16384,12288,0.6710640059577094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2048,2048,0.03538844320509169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,128,0.001875555556681421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,5120,0.09182933304044937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,768,0.004001777619123459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,1536,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,16384,0.02342222299840715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,1024,0.0038906666967603895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,2048,0.023531556129455566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,12288,0.17466044425964355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,6144,0.032219555642869734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,4096,0.06852977805667453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,32,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,7168,12288,0.422376897599962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,16384,0.22238755226135254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,1536,0.019223110543357003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,128,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,128,0.029327111111746892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,256,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,8192,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,5120,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,6144,0.03226666649182638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,51200,0.05096888873312208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,6144,0.009375111096435124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1024,51200,0.6771964497036405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,128,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,16384,0.12751555442810059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,7168,0.27580444018046063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,128,0.005526222288608551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,2560,0.005201777650250329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,7168,0.011048888994587792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3584,32,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,1536,0.026709333062171936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,3584,0.02924088968171014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,7168,0.013493333425786761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,10240,0.1587724420759413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,512,0.18585422303941515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,3584,0.00905600024594201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,1024,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,10240,0.03965244359440274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,8192,0.3587226602766249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,51200,2048,0.1034053365389506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3584,65536,0.09631733099619548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3584,3584,0.018935999936527677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,51200,0.04744088980886671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,65536,10240,16.23258802625868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,8192,0.06162666612201267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,32,256,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,5120,0.09631022479799058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,5120,0.009973333113723332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,2560,0.007544889218277401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3584,256,0.0162426663769616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,3072,0.05753778086768257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,51200,3584,0.26873776647779674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,128,1024,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,4096,1536,0.018938667244381376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,2560,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,2048,0.01071911139620675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,512,5120,0.01758577757411533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,10240,0.05317777726385328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,4096,0.04690222276581658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,7168,0.008359111017651027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,6144,0.08621155553393894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,768,51200,0.22014578183492026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,1024,0.021935999393463135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,3072,0.01516177753607432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,3584,0.09054311116536458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,7168,0.13666844367980957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,12288,0.09903289212120904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,10240,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,768,0.008991999758614434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,128,0.0053048887186580235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,128,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,12288,0.310843547185262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,2048,0.017328000730938382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,64,0.0018079999006456798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,256,12288,0.0075937774446275495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,3584,0.013398222625255585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,128,0.0018053334206342697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,4096,0.009972444838947719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,65536,10240,0.17485688792334664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,256,0.06736799743440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,16384,0.11573689513736302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2560,3072,0.017668444249365065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,4096,0.008572444319725037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,4096,32,0.049590223365359835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,51200,0.06291288799709745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,32,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,65536,0.1932284434636434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,8192,51200,0.07676888836754693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,2048,5120,0.011388444238238864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,8192,0.025990222891171772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,512,7168,0.03923644291030036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,7168,5120,0.10722844468222724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,10240,2.140424940321181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,256,8192,0.043486220969094165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2048,12288,0.017116443978415597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,32,0.010900444454616971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,768,256,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,51200,32,0.003540444291300244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,2048,0.01995733380317688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,128,0.01484444406297472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,5120,128,0.005566222386227713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,1024,0.02618933386272854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,16384,128,0.026025777061780293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,3072,0.010790222220950656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,6144,3584,0.09673955705430771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,65536,10240,0.49610312779744464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,3584,0.0655662218729655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,7168,256,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,2048,0.004200889004601372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,10240,0.024715556038750544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,1536,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,12288,0.15307999981774226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,10240,0.12878666983710393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,6144,0.025758221745491028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,1536,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,5120,0.00960177762640847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,8192,2.581296920776367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,8192,0.014378666877746582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,256,16384,0.029127111037572224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,2048,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,256,0.005578666511509154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,1536,0.012599999705950419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,768,0.006643555644485686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,7168,768,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,16384,2048,0.11505777306026882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,4096,0.008852444589138031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,10240,0.01993777851263682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,12288,0.034993777672449745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,51200,0.10301599899927776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,8192,0.5821448961893717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,5120,0.011809777882364062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,2048,0.029666665527555678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3072,32,0.005723555468850666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,5120,0.4313173294067383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,10240,0.015789333317014907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,4096,0.007691555553012424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,8192,2560,0.011620444556077322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,32,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,1024,0.0038399998512532976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,2048,0.012703110774358114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,768,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,256,0.02016622159216139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,3584,3584,0.007349333001507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,32,6144,0.00998755544424057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,64,64,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,768,2048,0.006683555742104848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,1536,0.028186665640936956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,6144,0.03553244471549988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,1024,0.02834133307139079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,6144,0.02141688929663764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,2560,0.012890666723251343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,16384,0.036900444163216486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,16384,0.02184355590078566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,6144,0.0203475554784139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,32,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,2048,0.019506666395399306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,1024,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,32,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,51200,0.1505262189441257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,3072,0.027023111780484516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,2048,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,7168,0.0993751088778178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,1536,65536,0.4446889029608832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,5120,0.01737955543729994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,5120,0.03516711129082574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,1536,0.006635555376609166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,65536,3072,0.1114506721496582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,3072,0.013740444348918067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,3072,10240,0.10818311240937974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,768,65536,0.029512001408471003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,2048,0.023365333676338196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,51200,0.3393431239657932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,12288,0.013409777647919126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,10240,65536,0.8664835294087728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,16384,4096,0.030357334348890517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3072,10240,0.021732444564501446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,8192,0.0203804439968533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,8192,0.17095732688903809
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,12288,2048,0.011934222446547614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,16384,0.18794133928087023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,256,0.01089422239197625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,128,0.019295111298561096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,3584,0.054754667811923556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,7168,8192,0.068631112575531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,10240,3584,0.026368000441127356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,2560,0.027095999982621934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,128,0.009314667019579146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,3584,2048,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,256,2560,0.013443555268976422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,256,0.014077333940400017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,256,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,32,3072,0.01886044442653656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,128,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,2560,0.007195555501514011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,5120,0.06578310992982653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,1536,0.024452444579866197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,128,0.007626666790909237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,16384,0.05034133460786608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2048,12288,0.45204798380533856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,256,0.03592977921168009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,1024,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,6144,0.027676444914605882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,5120,0.015607110328144498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,10240,0.020908445119857788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,5120,0.1133448945151435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,512,256,0.014469333820872836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,2560,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,12288,0.20065866576300728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,2560,0.031657778554492526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,256,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,12288,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,6144,0.01939733326435089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,2560,0.005264889034960005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,6144,0.057709336280822754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,256,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,1024,0.020273778173658583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,12288,512,0.01644177734851837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,3072,1.1752231385972765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,1536,3584,0.008391111261314815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,12288,12288,0.050127112203174166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,7168,0.010381333529949188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3072,2048,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,2560,7168,0.027641778190930683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,1536,0.010688888529936472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,16384,3072,0.032335999939176775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,10240,0.07838400204976399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,2560,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,32,0.05628444751103719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,10240,64,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1536,4096,0.02574488851759169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,8192,0.09567999839782715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,3072,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,5120,0.6042479938930935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,768,0.007280888656775157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,8192,0.042688889635933765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,8192,0.03460266523891025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,10240,2560,0.02472977836926778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,512,0.005607999861240387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,256,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,512,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,2048,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,3584,0.01755555636352963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,3584,0.01614488826857673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,4096,0.04422310988108317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,2560,0.004895110925038655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,16384,0.12456977367401123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,4096,2560,0.026914666096369427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,256,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,2048,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,1536,0.03018666638268365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1024,12288,0.011140444212489657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,256,0.014467555615637036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,4096,0.009497777455382878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,2560,0.0069306666652361555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,65536,1536,0.26536888546413845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,2560,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,10240,0.014122666584120857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,12288,0.040178666512171425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,51200,0.1299626694785224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,768,0.05798044469621447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,256,0.005577777822812398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,64,0.0021742222209771476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,51200,1536,0.07140444384680854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,4096,6144,0.023323555787404377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,32,768,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,8192,0.08594133456548055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,4096,0.8064124319288465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,8192,0.01275111072593265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,1536,0.025820444027582806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,3584,0.027639110883076985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,2560,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,8192,3072,0.0769875579410129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,3584,0.007681777907742395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,768,0.015443555182880826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,2048,0.008651555412345463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,1536,0.019428445233239066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,6144,0.02407466702991062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,1536,0.018043556147151522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,4096,0.007871111234029135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,8192,0.07639466391669379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,3072,0.026260443859630164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,12288,0.061055110560523145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,51200,0.5466186735365126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,64,0.012078222301271228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,3072,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,2560,0.048230220874150596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,10240,0.4958506690131293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,5120,0.034719109535217285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,6144,32,0.006897777732875612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,6144,0.02164977788925171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,2560,0.010096888575288985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,7168,0.04996711015701294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,2560,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,512,0.11822756131490071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,12288,0.0606000026067098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,7168,0.02944355540805393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,6144,128,0.009657777845859528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,2560,2048,0.006677333265542984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,5120,0.038872000243928694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,16384,5120,0.021314667330847845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,65536,1.269504017300076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,12288,0.10195022159152561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,768,256,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,16384,0.03552177879545424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,1536,0.01421866648726993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,6144,0.24259821573893228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,1024,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,16384,0.026424888107511733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,1024,0.46543110741509336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,4096,0.027385777897304956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,128,10240,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,2560,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,4096,0.9386986626519097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,32,0.00793688909875022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,2048,0.28259823057386607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,8192,0.010755555497275459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,10240,3584,0.017120000388887193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,1536,0.007921777665615082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,768,0.006978666616810693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,256,128,0.008291555775536431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2560,8192,0.01330311099688212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,51200,0.7310115496317545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2560,6144,0.0621582203441196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,16384,0.03682311044798957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,51200,0.04471910993258158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,768,0.0028026666906144884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,256,0.01756355497572157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,16384,0.08918133046891953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,16384,64,0.008674666285514832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,12288,12288,0.3489377763536241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,2048,0.037238223685158625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,3072,0.061216890811920166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,256,10240,0.05520977907710605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,128,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,6144,32,0.0214008887608846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3584,1024,0.015061333775520325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,12288,0.05927466683917575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2560,64,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,4096,0.2577226691775852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,128,0.007997333175606197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,5120,12288,0.019831111033757527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,5120,4096,0.03281599945492215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1024,16384,0.02125511070092519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,65536,5120,0.30907734235127765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,512,0.012771555946932899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,2048,0.009019555317031013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,8192,7168,0.03698933455679152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,4096,2560,0.04969244533114963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,12288,0.03315111001332601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,51200,10240,0.1705626646677653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,7168,2.09153069390191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,512,8192,0.06957155466079712
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,10240,0.02200266718864441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,2048,0.01591999994383918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,32,2560,0.006625777731339137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,3072,3584,0.029574222034878198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,3584,0.016167110866970487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,64,10240,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,64,0.052072889275021024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,512,2048,0.025014221668243408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,12288,2.2199334038628473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,1024,0.025910221868091162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,3584,0.03318399853176541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,7168,0.016454221473799813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,128,0.005574222240183089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,3072,0.007316444483068254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,7168,0.016364443633291457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,256,2560,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,3072,0.009085333181752099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,256,0.015141333142916361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,2048,0.03319199879964193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1024,64,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,51200,0.9275947146945529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,7168,0.030802667140960693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,2048,0.005888000130653381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,64,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,3072,0.053433779213163585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,1024,0.02534577747186025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,16384,0.02077244387732612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,6144,0.024352888266245525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3584,2560,0.02216177847650316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,7168,0.11371200614505345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,768,0.016139555308553908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,2560,3584,0.036419557200537786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,32,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,3072,0.11509066157870823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,256,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,128,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,1536,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,6144,0.019204444355434842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,3072,3584,0.008271111382378472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,3584,0.15059822135501438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,128,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,7168,0.016160888804329764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,8192,7168,0.06144444147745768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,1024,4096,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,65536,10240,0.08696533573998345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,5120,0.037979556454552546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,10240,8192,0.030637333790461223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,3072,0.007811555431948767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,65536,0.06295644574695163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,768,1024,0.03871733281347487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,4096,0.007538666327794393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3584,3072,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,12288,0.030260443687438965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,16384,0.11110488573710124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,128,128,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,12288,0.018242667118708294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,3072,0.8440773222181531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,5120,0.01125511113140318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,5120,7168,0.013402666482660504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,65536,0.03719644414054023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,128,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,51200,0.04418666826354133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,10240,0.02306488818592495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,51200,0.8294489118787977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,12288,8192,0.02687377731005351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3072,1024,0.012431999875439538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,64,0.00701333334048589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,6144,0.017127111554145813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,128,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,1536,0.016195555528004963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,8192,1536,0.03283199999067519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,3072,0.021249777740902368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,3072,0.014874666929244995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,2560,0.053114665879143604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,6144,5120,0.03446755475468106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,51200,0.29236088858710396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,1024,0.0057591112951437635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,768,65536,0.22001243962181938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,3584,0.07191199726528592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,51200,1024,0.015303999185562134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,2048,0.01054133309258355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,12288,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,12288,0.04563822348912557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,8192,8192,0.027270222703615826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,2048,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,3584,11.272940741644966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,10240,0.041280001401901245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3584,128,0.01511111193233066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,8192,0.0609875586297777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,10240,0.012092444631788464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,768,1024,0.020149333609475028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,12288,0.005274666680230035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,12288,0.0188044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,1536,0.0039013334446483185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,2048,0.005922666854328579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,51200,0.10348622004191081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,768,8192,0.009728000395827824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,128,1536,0.0042239998777707415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,5120,7168,1.7380879720052083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,256,0.013813333378897773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,3584,0.004892444445027245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,12288,32,0.006263111200597551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,65536,15.425701565212675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,12288,128,0.0188400000333786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,128,0.0021679999513758556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,256,0.01752622259987725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,2048,0.006606222026877933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,4096,3584,0.034083555142084755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3072,12288,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,3584,0.7482426431443957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,8192,3584,0.09401955869462754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1536,12288,0.0340266658200158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,5120,256,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,51200,2560,0.43569066789415145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,65536,1.4191946453518336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,5120,0.008253333469231924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,3072,0.014294221997261047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,6144,0.17713867293463814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,256,0.005567111074924469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,256,65536,0.16333688629998103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,6144,0.025083555115593806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,4096,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3072,51200,0.052368889252344765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,128,0.011720000041855706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,256,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,1024,0.022830223043759663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,51200,0.10291644599702622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,51200,0.10433155298233032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,3072,0.0791813333829244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,512,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,16384,256,0.041947556866539844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,16384,0.14578844441307917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,51200,0.07505955298741658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,8192,0.026604443788528442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,4096,512,0.014479110638300577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,2560,0.05905510981877645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,7168,0.005527110977305307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,64,0.00794577764140235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,2560,0.015104000767072042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3072,10240,0.03329244587156508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,65536,0.5311644342210557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,768,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,3584,4096,0.009951999617947472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2048,256,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,512,512,0.01517777807182736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,768,0.005594666633341048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,7168,0.01830844415558709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,12288,128,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,4096,64,0.011688888900809817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3072,16384,0.020635555187861126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,10240,0.03524177935388353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,65536,0.06239022148980034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,2560,0.07431466711892022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,5120,0.018230222993426852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,65536,0.34651554955376523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,512,7168,0.019270223047998216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,128,0.016883555385801528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,256,0.0052133335007561584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,5120,0.009686222506894005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,51200,0.05603733327653673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,5120,1024,0.0061928890645504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,2048,512,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,2560,0.004871110949251386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,2560,0.382950226465861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3584,3072,0.027273777458402846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,10240,0.02337155573897892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,128,128,0.0017955555684036678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,1536,0.01608622239695655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,32,0.14608177873823378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,12288,1.471572452121311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,4096,2560,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,10240,512,0.04080977704789903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,8192,5120,0.014494222071435718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,3584,0.005534222142563925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,51200,0.13531022601657444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,16384,0.019880000087949965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,5120,4096,0.0103502223889033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,10240,0.06060444646411472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,2048,0.006239999913507038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,64,0.007663999994595845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,4096,3072,0.009673777553770278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,65536,0.02565777798493703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,12288,0.014715555641386243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,51200,0.039260443713929914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,6144,0.020564445190959506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,5120,0.037250667810440063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,4096,2048,0.006950221955776215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,3072,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,12288,7168,0.09518755806816949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,32,0.07971822553210788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,16384,0.03928177886539035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,128,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,768,12288,0.014070222775141398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,512,0.016511999898486666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,6144,0.020048000746303134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,51200,0.063009778658549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,5120,0.022796443767017786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,6144,768,0.017242666747834947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,512,2048,0.004571555389298333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,512,0.038817776574028864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,3584,0.057108441988627114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3584,32,0.00423911131090588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,65536,16384,0.23264887597825792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,2048,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,51200,6144,0.34456978903876406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,8192,16384,0.050591111183166504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,256,0.017328888177871704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,51200,5.118826548258464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,2048,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3584,7168,0.03422933485772874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,16384,0.020224889119466145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,512,1536,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,3072,0.06841422451866998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,5120,0.04104177819357978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,512,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,512,3072,0.016149333781666227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,65536,3584,0.0958462225066291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,4096,0.02338933282428318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,4096,0.02274399995803833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,51200,0.058564444382985435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,5120,4096,0.024833778540293377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,5120,0.02351288828584883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,128,0.8354506492614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,1536,0.02852977646721734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,256,64,0.0032248888164758682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,32,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,7168,0.023390221926901076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,1024,0.006586666736337874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,2560,0.011165333290894827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,64,0.0018871112002266778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,12288,7168,0.29151466157701283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,32,1536,0.006996444529957241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3584,8192,0.014488889111412896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,1024,0.017502221796247695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,4096,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,512,128,0.0021297778520319196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3072,64,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,256,6144,0.01717066599263085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,512,1536,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,65536,0.1728364494111803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,256,0.11821866035461426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,768,0.17834666040208605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,6144,0.023034666975339253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,12288,0.020566221740510728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,65536,0.07218666871388753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,2048,3584,0.015253333581818474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,51200,4096,0.6327333450317383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,10240,0.16791378127204049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,2560,0.008647999829716152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,2048,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,2048,0.2573946581946479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,8192,0.017356443736288283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,512,0.003206222214632564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,51200,0.08606221940782334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,256,0.015454222758611044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,1536,0.0034808889031410217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,768,0.01829866733815935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,4096,1024,0.015415110521846347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,3584,0.011360888679822287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,65536,6144,0.1283608939912584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,51200,0.23795734511481392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,64,3072,0.004263999975389904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,2048,0.0782524413532681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,10240,0.040800889333089195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,1024,0.014641778336630927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2048,2560,0.005497777627574072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,65536,0.2511715624067518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1536,1024,0.0031386667655573953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,512,0.011592888997660743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,7168,12288,0.03197066651450263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,6144,0.11151999897427028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,2048,0.004498666773239772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,1536,10240,0.09801066584057277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,2048,0.06963733169767591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,64,0.008958222137557136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,4096,0.04081511166360643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,16384,5120,0.12576622433132595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,4096,6144,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,6144,0.5688213242424859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,2560,0.05004444387223986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,64,0.010344888601038191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,2048,0.027497778336207073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,512,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,16384,4096,0.12173688411712646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,32,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,3584,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,8192,0.015245333313941956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,16384,1.341183132595486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,5120,10240,0.0358240008354187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,64,0.022633777724372015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,5120,0.013064888616402945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,6144,512,0.017204445269372728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1536,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,768,128,0.005607111172543631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,128,10240,0.017196445001496207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3072,65536,0.1252924468782213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,512,0.005666666560702854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,10240,65536,0.14446311526828343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,51200,256,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,64,0.014463111758232117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,16384,65536,4.606692420111762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,6144,0.009367999931176504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,128,0.009930666950013902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,128,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,2560,0.0580622222688463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,65536,1.6227110756768122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,6144,0.12155021561516656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,8192,0.3480551242828369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,10240,0.060773332913716636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,5120,0.010999110837777456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,65536,8192,1.5945119857788086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,32,8192,0.01668711172209846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,16384,0.8239502376980252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,256,128,0.01686044368478987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,512,0.007446222007274628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,65536,0.3475511074066162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,16384,1536,0.019699555304315355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,768,5120,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1536,51200,1.7155413097805445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,12288,0.018778666853904724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,128,0.0021546667234765156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,3584,0.01683111157682207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,768,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,6144,0.12144800027211507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,512,65536,0.17576177914937338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,2048,0.02939200070169237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,1536,0.008646222452322641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,7168,0.016155555844306946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1024,4096,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,7168,0.03215288784768846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,128,8192,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,2560,0.020936000678274367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,5120,256,0.01721333298418257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,768,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,12288,3584,0.14492800500657824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,512,16384,0.01035377797153261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,128,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,6144,0.07314311133490668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,51200,128,0.057723555299970836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,51200,0.15730399555630153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,4096,0.014461333552996317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,7168,4096,0.0890551143222385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,256,0.023947555157873366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,6144,0.010866666833559671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,16384,0.05871822436650594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,512,0.00555288874440723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,128,6144,0.00885333369175593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,1536,0.011739555332395764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,7168,0.011055111057228513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,256,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,12288,1024,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,6144,0.7346382141113281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,4096,0.036700444089041814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,10240,0.034268445438808866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,65536,0.0757146676381429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,768,0.004475555486149258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,8192,0.028288000159793433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,7168,0.07864889171388414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,1536,0.007301333049933116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,3072,0.005551110953092575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,12288,2048,0.010461333725187512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,3584,0.026782222919993933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,51200,1536,0.01958311100800832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,1024,2048,0.016121778223249648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,3584,0.00868977771864997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,2560,0.01572177807490031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,51200,0.0766542222764757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,1536,0.05007644494374593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,512,0.01051288925939136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,3584,0.013782222237851886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,7168,0.03138488862249587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,1024,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,4096,0.011744889120260874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,512,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,128,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,8192,0.03262844350602891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,768,0.030814223819308813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,51200,0.8893982569376627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,16384,4096,0.0828897754351298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,2048,0.024263999528355066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,2560,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,2560,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,12288,0.03894933395915561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,16384,0.13238578372531468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,10240,1536,0.019225777851210702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,32,0.007277333074145847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,2048,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,51200,1024,0.11188088523017035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,3584,2560,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,3584,0.06637866629494561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,128,0.002181333386235767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2560,7168,0.010065777434243096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,1024,0.003882666842805015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,8192,256,0.005616000129116907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,4096,0.009325332939624786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,65536,0.05749155415429009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,768,4096,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,1536,0.03379644287957086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,12288,0.038047999143600464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,7168,0.020253333780500624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,12288,512,0.05167555477884081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,3584,0.02643022272321913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2048,3072,0.022057778305477563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,10240,0.1898355616463555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,64,0.003148444410827425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,32,0.0021075554605987337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,768,0.01686311099264357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,3584,0.10980088843239678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,128,0.01575644479857551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,128,3584,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1536,16384,0.07317599985334608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,5120,8192,0.018020444446139865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,6144,0.009655999640623728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,12288,0.05727555354436239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,128,0.00184888889392217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,65536,0.07585244708591037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,128,0.01719911065366533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,768,0.017472000585661996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,2560,0.07448977894253202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,3584,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,7168,0.019401777121755812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,1536,0.009328000247478485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,10240,7168,0.039356443617079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,512,0.018753778603341844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,2560,0.006966222491529252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,12288,0.01793866687350803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,3072,0.007621333003044128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,2048,0.034183998902638756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,768,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,51200,0.7606542375352648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,51200,0.0794737802611457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,10240,0.11293778154585098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,2048,0.023056889573733013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,512,0.10116089052624172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,6144,0.03786133395300971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,128,3584,0.014881778094503613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,10240,32,0.007644444704055786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,128,0.010285333626800114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,12288,0.03420711225933499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,64,0.0021671110557185277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,3072,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3584,12288,0.01921688848071628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,16384,0.09879644711812337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,8192,2560,0.02574488851759169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,51200,256,0.01072533345884747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,51200,0.07057778040568034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,3584,0.038945777548684016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,12288,3584,0.0774773359298706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,256,0.02460266649723053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,2560,2048,0.006931555353932911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,512,0.012392000191741519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,5120,0.009294222626421187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,7168,0.032065778970718384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,65536,0.09694755739635891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,12288,0.07260888814926147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,12288,0.042118221521377563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,4096,0.004560889055331548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,5120,51200,0.12527644634246826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,4096,65536,0.09108533461888631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,8192,0.16200444433424208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,768,2560,0.21934489409128824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,65536,0.24670222070482042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,16384,10240,0.24578934245639375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,6144,0.13827644454108343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,65536,0.1865262190500895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,1024,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,51200,0.0646995570924547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,2048,0.014901333385043673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,12288,7168,0.04540888799561394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,1024,0.007387555307812161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,4096,0.019260444574885897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,10240,0.015801777442296345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,1536,0.026102221674389307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,51200,0.07119377454121907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1536,6144,0.05769955449634128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,16384,0.2803742090861003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,256,51200,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,65536,16384,7.419216579861111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,768,0.022035555707083807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2560,3072,0.04414844512939453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,3584,0.10182844267951117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,1536,0.019163555569118924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,1024,0.06963555680380927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,512,0.026728888352711994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,3584,0.06467111243142022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,64,0.00757155567407608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,512,128,0.01408622165520986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,64,0.07662400272157457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,3584,2048,0.006859555426571105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,1536,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,3072,512,0.013742222554153867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,5120,16384,2.004127926296658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,32,1536,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,256,0.09417600101894802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,128,0.02314577831162347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,32,0.002599999929467837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,512,0.024168888727823894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,51200,64,0.041498667663998075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,3072,0.007258666886223688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,64,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,256,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,5120,0.02840177880393134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,768,0.017514665921529133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,4096,0.008940444224410588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,12288,0.05657866928312513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,32,0.0028479999552170434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,128,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,768,0.024711999628278945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,5120,2560,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2560,32,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,256,0.005571555760171678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,6144,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,7168,0.9633902443779839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,32,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,12288,0.23919910854763457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,1536,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,768,0.026549332671695288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,8192,0.04513511061668396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,2560,0.6550346480475532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,1536,0.01704088846842448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,2048,0.022758222288555566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,6144,0.017161778277821012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,51200,768,0.09891733196046616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,6144,0.07183111376232572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,3072,0.09675377607345581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,7168,512,0.026848889059490625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,2560,0.01960355540116628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3072,16384,0.034621334738201566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1024,65536,0.20763466093275282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,768,768,0.0028640000770489373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,10240,0.1548115544848972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,5120,3072,0.01385600037044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,5120,0.061774220731523305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,128,0.016564443707466125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,32,0.005141333159473208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,16384,12288,0.05809422334035238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,256,2560,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,12288,64,0.040955556763543025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,1024,0.028583112690183852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,64,0.023266666465335425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,512,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,3072,0.014110222458839417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,12288,0.2583955658806695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,2560,0.011657777759763928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,16384,0.26350755161709255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,7168,0.0458186666170756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2560,1536,0.017886221408843994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,8192,0.0189057770702574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,10240,0.024675556355052527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,65536,0.8793768882751465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,10240,0.07582133346133761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,51200,0.16599022017584905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1024,10240,0.010315555665228102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2560,2560,0.03455822335349189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,3072,7168,0.054551111327277295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,2048,256,0.01517511076397366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,256,32,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,10240,8192,0.0372648901409573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,8192,768,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,4096,0.00848355558183458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,768,64,0.001820444439848264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,3072,0.01036800030204985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,128,0.13974489106072321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,256,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,8192,0.05070755547947354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,5120,512,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,4096,12288,0.10403644376330906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,65536,0.05816266934076945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,8192,0.10642844438552856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,16384,3072,0.01370044466522005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,4096,8192,0.09716355800628662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,2560,0.08346133099661933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,5120,0.021592888567182753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,2048,0.013845333622561561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,65536,0.8078479766845703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,12288,0.11809955702887641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,5120,0.014186667071448432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3584,256,0.014818666709793938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,51200,512,0.01755822201569875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,512,0.007967111137178209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,7168,12288,0.02979733215437995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,256,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,2560,7168,0.040876444843080305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,2560,0.018373333745532565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2048,128,0.01203644441233741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,128,768,0.005946666830115848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,128,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,64,0.003257777748836411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,1536,0.02720977697107527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,2560,512,0.02139733400609758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,256,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,3584,0.037801778978771634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,5120,0.009128000173303816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,10240,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,8192,64,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,3072,0.028759999407662287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,2048,0.008984000318580205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,51200,0.10699644353654648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,12288,6144,0.2500426769256592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3584,2048,0.03194933467441135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,8192,12288,0.03135733471976386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,12288,4.971726311577691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,64,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,1536,0.006639110959238476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,8192,0.025429333249727886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,4096,0.11109422312842475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,1536,0.023807111713621352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,10240,51200,0.11587822437286377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,51200,0.12644444571601018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,16384,0.0251226673523585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,128,51200,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,3584,0.004930666751331753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,256,0.0026613333159022858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,2560,0.069541335105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1536,256,0.01957244508796268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,16384,5120,0.029827555020650227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,2560,0.0064586665895250105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,7168,0.02127911150455475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,256,0.09103200170728896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,3072,0.017824888229370117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,7168,0.022919111781650122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,16384,0.3313608964284261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,8192,0.015136889285511441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,51200,12288,0.5739244355095757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,4096,0.009307555854320526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,2560,64,0.011363555159833698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,2048,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,16384,0.24641156196594238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,65536,0.1914853387408786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,5120,0.01186222251918581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,51200,256,0.006592888798978593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,6144,2560,0.03741866681310866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,32,0.00795288880666097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,10240,0.08581778075959946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,12288,0.01444355481200748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,256,65536,0.05151999990145365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,2560,0.023187556200557288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,16384,10240,0.29038222630818683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,128,0.0052435555391841466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,3072,0.007721777591440413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,5120,8192,0.07793688774108887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3072,7168,0.0694133308198717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,1024,0.00870311094654931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,2048,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,64,7168,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,16384,0.4256871011522081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2560,10240,0.0639537771542867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,2048,0.04105955693456862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,65536,1.123843510945638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,7168,0.1496853298611111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,1536,0.01928622192806668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,128,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,512,0.016837333639462788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,2048,0.01330044451687071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,1024,0.030922667847739324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,128,0.015842666228612263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,768,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,128,0.020921778347757127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1024,2560,0.008623110751310984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,768,2048,0.026178666286998328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,4096,256,0.005227555417352253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,2048,0.03229422370592753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,128,512,0.013402666482660504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,16384,512,0.03300977746645609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,1024,0.015221332510312399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,2048,0.007703111403518253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,5120,0.9358417722913953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,4096,65536,0.08658488591512044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,512,5120,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,3072,0.021333333518770006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,4096,8192,1.5991289350721571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,7168,16384,0.17437244786156547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,8192,0.01350488927629259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,3072,0.02913244565327962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,10240,0.017343110508388944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,10240,64,0.03132889005872939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,65536,768,0.022720000810093347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,16384,0.13661600483788386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,64,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,10240,768,0.04037244452370538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,1024,0.011591110792424945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,6144,0.13181600305769178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,2048,0.07102311319775052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3584,2560,0.027085334062576294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,512,12288,0.07852266894446479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,2048,0.016553777787420485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,2048,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,16384,0.29229866133795845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,5120,64,0.008586666650242275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,65536,0.05424000157250298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,3584,0.024015110399987962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,6144,2560,0.09090844127866958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,512,5120,0.008357333640257517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,1536,3584,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,128,0.011160000330872007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,12288,0.032999111546410456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,16384,1536,0.01072355525361167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,256,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,7168,51200,0.09883111053042942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,2560,0.010384000009960597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,10240,0.014067555467287699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,64,7168,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,512,0.02068000038464864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,10240,0.0616168909602695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,65536,0.22564355532328287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,16384,2560,0.017993777990341187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,3072,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,64,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,1536,0.0070222218831380205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,1536,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,16384,0.02168088820245531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,2048,0.006446222050322427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,65536,0.05938488907284207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,10240,0.023436443673239812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,7168,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3072,3584,0.3936791155073378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,128,0.011317333413494958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,51200,0.09188710980945164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,64,0.006591111007663939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,4096,1024,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,8192,0.12626577748192683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,16384,768,0.07235466771655612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,7168,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,12288,128,0.0036799998746977914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,3072,0.26557421684265137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,4096,0.09649599923027886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,1536,0.006552888701359431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,1536,0.006883555402358373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,65536,2560,0.2963155640496148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,12288,2048,0.013416888813177744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,768,256,0.015476443701320224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,256,0.014444443914625378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,7168,0.09151911073260838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,768,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,3584,0.028696000576019287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,16384,0.03971822063128153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,7168,0.01975555552376641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,5120,3072,0.01276266657643848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,51200,0.04700977934731377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,512,2560,0.04829333225886027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,6144,0.715454207526313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,1024,0.017841777867741056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,8192,4096,0.07228355275260077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,12288,65536,0.2961217827267117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,12288,0.05015733175807529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,5120,64,0.016750221451123554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,4096,0.04975110954708523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,4096,0.04320444332228767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,3072,0.012787555654843649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,8192,0.13753067122565374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,512,512,0.003604444364706675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,512,65536,0.016693333784739178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,3072,0.010685332947307162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,10240,12288,0.03258400162061056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,8192,0.041659557157092623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,16384,0.3428213331434462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,51200,0.3548240131802029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,10240,0.014442667365074158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,3584,0.13037599457634821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,5120,0.018237334158685472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,16384,0.07138755586412218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,4096,0.0463137792216407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,51200,0.1802417834599813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,2048,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,768,0.0791048871146308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,10240,0.031256887647840716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1536,128,0.007983999947706858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1536,512,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,7168,0.021319111188252766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,2048,0.006647111227114995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,512,0.003530666646030214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,128,768,0.005802666561471091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,3584,0.014131555954615274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,51200,0.7072390980190701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,5120,0.011428444749779172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,128,0.027124444643656414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,2048,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,128,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,128,0.016198222835858662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,64,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,65536,0.158860445022583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,3584,0.023408888114823237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,16384,0.04721066686842176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,512,16384,0.04933511217435201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,32,65536,0.028016000986099243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,10240,0.12615111139085558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,4096,0.014918223023414612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,3584,0.016182222300105624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,1024,0.004670222186379963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,8192,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,51200,0.13437422116597494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,65536,7168,0.06292177571190728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,1536,5120,0.019621334142155118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2048,64,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,4096,0.7310560014512805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,64,3584,0.006596444381607904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,65536,256,0.03240177697605557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,7168,0.014249778456158109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,65536,0.06603288650512695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,4096,0.03458133339881897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,768,0.04000711109903123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,3584,0.008278222547637092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,2048,0.006062222023804982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,128,0.01517777807182736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,1536,2560,0.007278222176763747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,4096,3584,0.009104000197516548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,4096,12288,0.020208888583713107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2048,1536,0.018179555733998615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1024,2048,0.004550222307443619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,256,0.014768888552983602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,8192,128,0.002157333410448498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,8192,0.013844444519943662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,5120,0.027041777968406677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,5120,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,128,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,6144,8192,0.1251360045539008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,256,0.017850667238235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,768,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,1024,0.01455200049612257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3584,2560,0.03628977802064683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2048,65536,0.29210488001505536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,4096,0.016011555989583332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,1536,0.00832177781396442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,256,0.0052800000541739995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,51200,1.078094270494249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,64,0.006917333437336816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,512,0.009135999613338048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,12288,0.02275288932853275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,64,7168,0.0317466656366984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,51200,0.043416887521743774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,6144,0.041768001185523145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,3072,0.012898666991127862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,512,1536,0.0041253334946102565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,3072,0.01404088901148902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1536,65536,0.6409058041042751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,2048,0.15487999386257595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,12288,512,0.006616888774765863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,512,0.008338666624493068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,8192,0.02752799954679277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,512,0.08416444725460476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,51200,0.0667182207107544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,768,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,8192,0.012131555212868584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,8192,0.15464622444576687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,65536,4096,0.9915182325575086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,32,0.011310222248236338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,12288,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,1536,16384,0.017972444494565327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,128,10240,0.031261334816614784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,512,3072,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,7168,0.010240000155236986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,64,0.0059573331640826324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,51200,0.14271999730004206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,12288,0.027445332871543035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,7168,0.03655822078386942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,256,0.02197066611713833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,6144,0.01901511185699039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1024,10240,0.008779555559158325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,2560,0.020008888509538438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,32,0.011111111276679568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,7168,0.010320000350475311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1024,32,0.00796622203456031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,128,5120,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,768,0.017847999930381775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,1536,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,2560,0.05023911264207628
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,768,0.0028391112056043413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,6144,0.024503111839294434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,16384,0.018021333548757765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,10240,0.0373982224199507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,8192,4096,0.012034667034943899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,64,3072,0.08187822500864665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,2560,0.007311111523045435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,512,0.01146044416560067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,2560,0.08864621983634101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,10240,128,0.024280000064108107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,128,16384,0.016541333662139047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,768,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,768,0.003626666549179289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,128,0.0069164443347189165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,4096,0.03460177779197693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,6144,0.1262106630537245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,3584,0.014814222852389017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,12288,256,0.02461155586772495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,128,0.014504889647165934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,256,0.008631111019187504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,8192,0.011358222199810876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,10240,5120,0.017453332742055256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,8192,256,0.016506666938463848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,3072,0.007932444413503012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,768,0.005809777726729711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2048,768,0.012387555506494312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,256,0.014454222387737699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,51200,5120,0.26223821110195583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,12288,3584,0.017567111386193168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,1024,0.013071999781661563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,51200,0.3846008777618408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,3072,0.028804444604449805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,128,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,10240,16384,0.04411822226312426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,7168,0.023873777853118047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,5120,0.009348444640636444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,64,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,8192,0.020940444535679285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,768,0.00591111100382275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,1024,0.007128888534175024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,8192,0.0417502224445343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,512,0.006063111126422882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,3584,0.030043555630577937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,4096,0.15997778044806585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,256,0.012065778175989786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,6144,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,2048,0.007985778152942657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,512,0.015222221612930298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,4096,4096,0.00962222201956643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,12288,6144,0.18917244010501436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,3072,0.03394666645261977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,5120,0.00777333312564426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,3584,3584,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,65536,0.5178382131788466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,6144,0.011337777806652917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,6144,768,0.0031893334041039148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,2048,0.017493334081437852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,64,0.006614222294754452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,128,768,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,256,0.005418666534953647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,512,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,768,512,0.017895110779338412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,6144,0.07900889052285089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,3584,0.1149839957555135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,12288,0.03730310996373495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,1536,0.006634666687912411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,64,0.019128888845443726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,7168,8192,0.06203733550177681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,7168,0.06565066840913561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,10240,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,128,0.005554666535721884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,10240,8192,0.04965244399176704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,3584,128,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,5120,512,0.01587199999226464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,10240,0.23635644382900664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,65536,0.08905688921610515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,16384,16384,0.053561776876449585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1536,16384,0.021027555068333943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,16384,0.011026666396194033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,4096,0.015275556180212231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,2560,51200,0.05152977837456597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,4096,0.015435554915004306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,5120,0.009380444056457942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,1536,0.045181334018707275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3072,12288,0.02940800123744541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,2560,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,5120,16384,0.04891733328501383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,6144,0.11734844578637017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,12288,0.03587111168437534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,7168,512,0.01648355523745219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,16384,0.17612622843848336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,3072,0.011724444727102915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,2560,0.004932444542646408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,4096,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,768,3072,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,4096,0.025062221619817946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,10240,3584,0.10744533273908828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,5120,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,512,0.009654222263230218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,51200,0.10274400313695271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,6144,0.040949334700902305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,4096,0.030138668086793687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,32,0.002423111142383681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,8192,0.04299644298023648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,3584,0.019189332922299702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,4096,0.017634666628307767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,65536,16384,0.373489777247111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,12288,0.1721857786178589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,65536,0.055847108364105225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,768,0.01594577729701996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,512,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,256,0.02371911042266422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,65536,0.19218756092919242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,768,0.003564444267087512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,512,0.006142222219043308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,51200,128,0.0670622189839681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,2048,3072,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,512,0.007620444728268518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,256,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,10240,0.011437333292431302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,6144,0.11951200167338054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,10240,0.013355555633703867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,51200,0.11360889010959202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,768,0.012921777864297232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,4096,0.04472800095876058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,512,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,2048,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,256,51200,0.05364177624384562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,12288,0.02269422180122799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,2048,0.03543466660711501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,256,0.0038844446341196695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,64,0.007941332956155142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,2048,0.0042408886882993905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,4096,2048,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,5120,0.08083466688791911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,2560,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,10240,0.09227644734912449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,8192,0.036764442920684814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,6144,0.010885333021481832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,2048,0.007649777664078607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,65536,0.25475645065307617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,8192,0.04280089007483589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,256,0.01441511180665758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2560,5120,0.027259555127885606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,16384,0.02141155633661482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,3072,0.013790222505728403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,65536,0.05819111400180393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,2048,0.018887110882335238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,16384,0.02029333346419864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,3072,6144,0.04242755638228523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,1024,0.013799111048380533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,32,2560,0.004903111192915174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,3072,0.008864888714419471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,65536,0.4018408987257216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,3072,0.10919466945860122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,12288,0.010149333212110732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,512,0.012494222157531314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,512,0.021957332889238994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,5120,6144,0.013396444420019785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,51200,0.09921599758995904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,7168,128,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2048,8192,0.01088355564408832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,5120,0.008664000365469191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,12288,5120,0.06832533412509494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,16384,7168,0.0528400010532803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,3584,0.023360000716315374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,512,0.015579554769727917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1536,512,0.0025066667132907445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,768,0.01960533360640208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,512,0.14785333474477133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,5120,0.07365777757432726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,2560,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,3584,0.004943999979231092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,4096,0.00830666638082928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,256,0.002487999904486868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,3072,0.028833776712417603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,256,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,7168,0.27672354380289715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,512,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,512,0.005972444597217772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,512,0.020245333512624104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,2560,0.007673777639865875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,128,0.0052791109515561005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,768,0.006569777925809224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,768,0.003506666670242945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,2560,0.050386667251586914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,3584,0.011163555913501315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,12288,0.15549777613745794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,3584,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,6144,0.22128800551096597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,16384,0.06394577688641019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,51200,6144,0.058668441242641874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,5120,0.013463111387358772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,5120,0.0458808872434828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,4096,0.11553866333431667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,768,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,3072,0.01719466679626041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,1536,0.004249777644872665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,16384,0.2487724357181125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,4096,1.117736922370063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,10240,0.0259226659933726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,256,512,0.010335110955768162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,128,0.009669333696365356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2560,64,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,2048,0.012720888687504662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3072,1024,0.013088000317414602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,8192,2048,0.0075039996041191956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,4096,10240,0.15928888320922852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2560,5120,0.012454222473833295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,5120,2560,0.044349332650502525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,10240,0.28536889288160533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,3584,0.09998310936821832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,10240,3.8453581068250866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,128,0.009681777821646797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,4096,0.01609866652223799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,2560,0.015966221690177917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,768,3072,0.006099555641412735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,2048,0.011005333728260465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,2048,0.016536000702116225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,10240,0.023018666439586218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,51200,5120,0.03881688912709554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,32,0.0028328889360030494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,6144,1536,0.03866399990187751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,5120,10240,0.08521778053707546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,4096,0.09382044606738621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,5120,0.004936888813972473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,5120,256,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,2560,0.008054222497675154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2048,2560,0.00554577757914861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2048,12288,0.027291556199391682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,12288,0.04442044430308872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,768,0.008640888664457533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,3584,0.015108444624476962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,7168,0.021984888447655573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3584,51200,0.5564231342739528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,7168,0.04030133287111918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,10240,0.031903998719321355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,3584,0.06170666880077786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,16384,0.03618755605485704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,12288,0.026146666871176824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,3584,0.00793511089351442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,256,0.010998222563001843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,16384,0.014959999256663851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,51200,0.2438595559861925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,16384,0.19570044676462808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,32,64,0.0019964445382356644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,128,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,65536,1024,0.40149421162075466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,256,128,0.011713777979214987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,256,768,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,4096,0.007965332931942409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,512,0.028868443436092798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,768,3584,0.008431110945012834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,8192,0.022797332869635686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,64,0.0021280000607172647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,65536,128,0.01016088906261656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,6144,0.05214222272237142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,32,32,0.0018328889790508484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,256,512,0.005894222193294101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,3072,0.016138666205936007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1024,256,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,768,0.003264888914095031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,51200,0.918823136223687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,5120,0.025806221697065566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,64,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,2048,4096,0.015507555670208402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,32,51200,0.102528002527025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,768,512,0.009072889056470659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,10240,0.10894755522410075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,3072,0.008343111309740279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,7168,768,0.14937867058648005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,128,0.0028719999310043124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,512,8192,0.04464088877042135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,3584,10240,0.10257599751154582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,2048,0.02642044425010681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,65536,512,0.021525333325068157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,2048,2048,0.008672888908121321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,512,0.005893333504597346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,7168,8192,0.05048177639643351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,3072,5120,0.11117066277398004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,10240,0.0314684444003635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,8192,0.053596443600124784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,5120,0.042120890484915845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,12288,4096,0.696183098687066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,4096,1024,0.018477333916558158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,51200,0.3952017890082465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,2048,0.004358222087224324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,16384,256,0.027112889620992873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,16384,0.013800000150998434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,1536,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,256,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,3072,0.032046222024493747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,65536,12288,0.4584346877204047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,2048,0.013345777988433838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,51200,0.08478311035368179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,128,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,6144,0.022500443789694045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,10240,0.018732445107565988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1024,2560,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,256,0.02368800010946062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,512,3072,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,1536,0.01723288827472263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,1024,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,7168,7168,0.04088266690572103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,256,7168,0.016879111528396606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,6144,0.025292444560262892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,128,0.002395555583967103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,2048,128,0.019885333047972787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,65536,0.0888035562303331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,128,65536,0.06026044156816271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,32,0.005610666755172942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,8192,0.042704890171686806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,256,10240,0.31527911292182076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,7168,65536,0.12690933545430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,3072,5120,0.009323555562231276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,5120,2560,0.007892444729804993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,8192,0.019941333267423842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,16384,0.16677510738372803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,2560,0.017765333255132038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,256,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3584,8192,0.014211555322011312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,7168,1536,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,3072,0.019289778338538278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,512,65536,0.06322489182154338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,128,0.006447111152940326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,1024,0.007890666524569193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,32,1024,0.006275555739800136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,2560,0.026382222771644592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,4096,0.03928177886539035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,1024,0.0028808888875775864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,10240,0.08817511134677464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,64,0.0069271110826068455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,256,7168,0.020640000700950623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,7168,0.014422222971916199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,128,0.012130666938092975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,256,3584,0.014488000008794995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,8192,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,2048,3072,0.026757333013746474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,7168,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,5120,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2560,65536,0.0700942211680942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,5120,0.02696711156103346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,3072,0.038022223446104265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,1536,0.03668444355328878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,32,256,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,256,0.0025635556214385563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,32,0.0018515555808941524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,12288,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,2560,12288,0.01754399968518151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,6144,0.04155022237035964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,51200,0.06975644164615206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,256,0.005590222362014983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,768,0.02544622288809882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,16384,1.342898686726888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,1536,0.004531555705600315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,64,51200,0.006481777876615524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,768,0.012764444781674279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,8192,0.14805332819620767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,10240,0.017826666434605915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,512,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,128,0.00461422238085005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,3072,0.00997688869635264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,256,0.008341333104504479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,7168,3584,0.029436445898479883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,5120,16384,0.040384001202053495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,51200,0.04740088846948412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,32,0.009688888986905416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,768,0.008307555483447181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,32,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,2560,0.018219555417696636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,256,512,0.005572444448868434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,2560,0.037054220835367836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,8192,0.07878400219811334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,4096,256,0.014658666319317289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,1024,0.009744889206356471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,2560,0.00832088871134652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,8192,512,0.017581333716710407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,256,0.005320888840489917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,10240,0.00924800005224016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,256,512,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,128,0.014856888188256158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,10240,0.011326221956147088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,256,0.005978666659858491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,768,0.020974222156736586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,3072,0.007663110891977946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,8192,1536,0.007623111208279927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,3072,0.07879555225372314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,128,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,16384,0.18480266465081108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,512,0.0158933334880405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,4096,0.07724266582065158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,65536,0.01182755579551061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,3072,0.02975911233160231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,1024,0.005976000179847081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,3584,0.014691554837756686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,16384,0.08920622534222072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,10240,0.02713777787155575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,7168,0.02924088968171014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,7168,0.12275377909342448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,3072,0.011038222246699862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,128,0.015662221444977652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,7168,17.336207071940105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,16384,51200,0.2709386613633898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,65536,10.24147203233507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,5120,0.010612444745169746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,10240,0.42455199029710555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,5120,0.01722044414944119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,1024,2560,0.010098666780524785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,512,0.009528000321653154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,2560,5120,0.05473866727617052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,1536,0.057842665248446994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,3584,0.02823466724819607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,128,0.0144533332851198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,51200,0.15838044219546848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,1536,0.010984000232484607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,10240,7168,0.026668444275856018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,51200,3072,0.026195555925369263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,51200,0.16355378097958034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,128,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,256,0.013803555733627744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,2048,0.06628088818656074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,128,0.00490488898422983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,5120,0.01571377780702379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,32,0.02299022177855174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,3072,1536,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,3584,0.011000889043013254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,51200,0.02906222144762675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,2048,0.049689776367611356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,12288,8192,0.04872711168395149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,7168,0.01718844473361969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,512,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,8192,768,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,4096,0.0052426668504873914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2048,768,0.017845332622528076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,32,0.019183999962276883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,256,0.013609777722093793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,512,0.005621333503060871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,51200,0.1195146640141805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,16384,0.04201777776082357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,7168,256,0.0024888888001441956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2560,4096,0.03537777728504605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,2560,0.02382044494152069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,51200,0.03795377744568719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,32,8192,0.01310222182008955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,128,0.014459555347760519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,6144,16384,0.02287111183007558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,3584,0.012221333053376941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,16384,0.024990222520298425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,10240,0.11632443798912896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,256,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,1024,0.006294222341643439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,12288,1024,0.041144887606302895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,4096,0.008301333420806462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,12288,0.04976711008283827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,2048,0.1948773331112332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,12288,65536,10.979425218370226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,768,0.02769244379467434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,6144,0.7635119756062826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,768,16384,0.0424906677669949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,16384,0.2551066610548231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,128,5120,0.009294222626421187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,12288,0.22054221895005968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,768,0.15473422739240858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,1536,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,2560,0.018583110637134977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,65536,0.06701866785685222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,10240,6144,0.2025617758433024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,128,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,128,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,128,51200,0.0655662218729655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,128,0.03152799937460158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,3584,0.07413778040144178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,5120,64,0.013807999591032663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,128,0.01721244388156467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1024,51200,0.033896889951494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,51200,0.06487199995252821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,128,0.01551555593808492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,16384,0.023898666103680927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2048,3072,0.007643555601437886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,4096,16384,0.1931635538736979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,3072,0.007911110917727152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1024,6144,0.008520888785521189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,7168,0.05694399939643013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,128,0.01204177737236023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,8192,0.010693333215183683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,1536,0.020278222031063504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2560,3072,0.019532443748580087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,32,5120,0.02404444416364034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,256,6144,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,3584,0.0076657773719893555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,10240,512,0.006923555499977536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,7168,2.2700587378607855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,768,64,0.0018240000224775737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,1024,0.007557333343558841
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,4096,5120,0.02160444524553087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,32,6144,0.010110222631030613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,256,256,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,7168,2560,0.02591466738118066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,5120,0.029640889830059473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,65536,0.2312471071879069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,128,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,256,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,2560,0.00815644446346495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,1024,0.01070844464831882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,65536,7168,0.12748977873060438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,768,8192,0.02792177928818597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,7168,0.014779556128713818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,1024,0.010016889207892949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,65536,12288,1.4697661929660375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,16384,2560,0.05767644113964505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,4096,0.010657777388890585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,12288,0.014321777555677625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,7168,0.6484844419691298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,128,0.0031964443624019623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,768,0.02530755599339803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,65536,51200,1.157406171162923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,51200,1.8488409254286025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,3584,0.02092444399992625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,768,0.038098666403028704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,8192,512,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,3584,2560,0.02677244444688161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,16384,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,512,2560,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,512,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,1536,0.04160355528195699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,3072,0.13025332821740046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,7168,0.05522400140762329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,768,5120,0.03119022316402859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,65536,0.05151999990145365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,128,3584,0.007958222594526079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,5120,0.06094577577379015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,65536,6144,0.3344568941328261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,2048,0.028580443726645574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,6144,256,0.012915555801656512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,32,0.00998133338159985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2560,1024,0.02921066681543986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,7168,0.01740533279048072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,12288,0.05794488721423679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,7168,0.08868088987138535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,256,0.01681422193845113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,64,0.007611555357774098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,256,0.016879111528396606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,768,0.009324444664849175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2560,256,0.025449777642885845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,1024,0.27086488405863446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1024,4096,0.017678222722477384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,3584,0.010324444207880232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,512,0.03868800070550706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,8192,0.620399104224311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,768,0.007532444265153673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,3584,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,10240,0.026955554882685345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,6144,0.06149599949518839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,7168,0.0042275554604000514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,512,2048,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,6144,0.013334222137928009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,51200,0.8008284568786621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3072,1024,0.021515554851955835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,12288,1.651860449049208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,10240,0.0326222214433882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,10240,0.03286133209864298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,51200,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,2560,0.016680000556839835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,768,0.006581333362393909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,768,16384,0.01650311052799225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,51200,0.27506934271918404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,128,0.00656355544924736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,256,0.07080444362428454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3584,4096,0.009344889058007134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,1536,0.009016888837019602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,3584,0.024033778243594702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,7168,0.05074844426578946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,12288,0.05608266592025757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,6144,7168,0.040892445378833346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1024,3584,0.0074240002367231585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,6144,51200,0.05793244308895535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,10240,0.1888320048650106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,32,51200,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,512,0.012899555265903473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,128,512,0.015118221441904703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,1536,0.0359280010064443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,2560,0.012079999678664737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,10240,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,10240,1024,0.015378667248619927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,128,256,0.0021102221475707158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,10240,0.2493226793077257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,7168,0.08833066622416179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,1024,0.016349333855840895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,768,0.007050666544172499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,2048,256,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,5120,128,0.014848000473446317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,1536,0.016194666425387066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,3072,0.012879999975363413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,512,128,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,6144,0.2163555489646064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,1536,0.009317333499590555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,2560,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2560,512,0.009308444129096137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,5120,4096,0.013400000002649097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3584,3072,0.03254133462905884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,2560,0.008087111016114553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,256,4096,0.010033778018421596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,6144,0.037779556380377874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,32,7168,0.010235555469989777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,10240,0.015088889333936902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,16384,0.04463466670778063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3584,7168,0.09264266490936279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3584,1024,0.0064071110553211635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,2048,0.02472977836926778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,7168,0.06599822309282091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,3584,0.7084213362799751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,65536,0.0940435528755188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,65536,1.2362480163574219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,3072,128,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,128,10240,0.02030666669209798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,8192,65536,0.17023555437723795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,512,0.002560888934466574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,128,0.0021644443687465456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,2048,0.010996444357766045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,2560,10240,0.059296886126200356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,10240,0.02760444415940179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,512,10240,0.028112000889248315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,12288,0.03822044531504313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,128,0.018566222654448617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,128,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,12288,0.0647982226477729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,64,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,3584,0.006495111104514863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,64,0.006612444503439798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,128,51200,0.0634853310055203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,3072,0.06971910927030775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,5120,0.08500888612535264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,6144,12288,0.02091644373204973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2560,64,0.011696889168686338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,128,3072,0.0551182230313619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,128,1024,0.005585777676767773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1024,3584,0.00776088900036282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,6144,7168,0.6101084285312229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,64,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,8192,0.0525911119249132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,768,128,0.006301333506902059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,8192,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,5120,0.03920977645450168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,51200,0.059824890560574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,16384,0.09113866753048366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,3072,768,0.02882044514020284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,6144,8192,0.014458666245142618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,768,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,10240,768,0.010932444698280759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,6144,65536,7.126128302680121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,5120,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,16384,0.06534400251176622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,6144,5120,1.4185351265801325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,16384,64,0.03208266695340475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,1536,0.009963555468453301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,51200,0.19988000392913818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,512,0.014059555199411182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,256,0.026713778575261433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,3584,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,12288,768,0.018623999423450895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,3072,0.04935466580920749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,1024,0.011686222420798408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,3072,0.0466195543607076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,32,0.001793777777089013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,1024,0.018350222044520907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,10240,0.01920977731545766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,65536,51200,15.442911783854166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,7168,51200,0.348523563808865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,4096,0.039930668142106794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,2560,0.006871999965773688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,5120,0.1540959941016303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,3072,0.03567466802067227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,4096,32,0.0028719999310043124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,7168,0.21263199382358125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,256,0.005591111050711737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,65536,12288,2.3548755645751953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1536,2048,0.01037599974208408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,6144,768,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,2560,0.025790222816997107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,2560,0.08176088995403714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,10240,0.026854222019513447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,4096,1024,0.015405333704418607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,32,0.0031679999083280563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,128,65536,0.4564844237433539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,65536,1536,0.019167999426523846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,65536,5120,1.2165386411878798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,3584,0.01681866745154063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,16384,0.0416515568892161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,16384,3584,0.01642311116059621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,3072,0.018360000517633226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,6144,2048,0.009770666559537252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,12288,0.3719608783721924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,7168,3072,0.025128000312381323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,3072,0.021936888496081035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,1536,0.02599999970859951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,3072,0.0052737775776121355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,4096,2560,0.024783111280865137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,512,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,256,0.0028284444577164124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3584,1024,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,2048,0.04994666576385498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,1536,0.009336888790130615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,6144,128,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,256,0.012435555458068848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,768,0.017543110582563613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,65536,0.24401243527730307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,10240,0.033710221449534096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2048,65536,0.28101245562235516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1024,512,0.012849777936935425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,7168,7168,0.020444444484180875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,3072,0.0432604451974233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,8192,0.019553777244355943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,5120,0.07169422176149157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,10240,0.01628444426589542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,128,768,0.015863110621770222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,512,0.011195555329322815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,12288,0.10652622249391343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,1024,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,5120,5120,0.02801688843303257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,6144,0.009712888962692684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,3584,0.008760000268618265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,6144,0.01254488858911726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,2560,1536,0.015164444843928019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,128,0.006631999793979857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,256,2560,0.006909333169460297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,2048,0.006654222392373615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,2048,0.05269866519504123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,8192,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,3584,0.013195555243227216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,512,4096,0.018204443984561496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,12288,0.20542666647169325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,6144,0.2709999879201253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,3584,16384,0.024109333753585815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,256,12288,0.013802666631009845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,4096,768,0.017885333961910672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,2560,0.007259555160999298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,512,10240,0.012192000117566852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3584,1536,0.028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,4096,1536,0.028704000843895808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,4096,0.03418933351834615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,5120,0.008672888908121321
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,512,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,3584,0.043472889396879405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,4096,0.03474222289191352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,768,0.006221333311663733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,51200,0.07427910963694255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,2560,0.02438933319515652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,1024,8192,0.011398221883508893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,256,0.005872888697518243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,6144,6144,0.017344888713624742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,4096,0.017842666970358956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,4096,0.004917333523432414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,3072,65536,0.7338915401034884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,5120,0.2440799872080485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,6144,0.03930755456288656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,2048,0.010640888578361936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,2560,128,0.020249777370029025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,768,0.011387555963463254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,1536,0.009191999832789103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,3584,0.020624000165197585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2048,12288,0.271871116426256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,512,256,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,12288,0.16591200563642713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,12288,0.019924443629052904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,2560,0.014001778430408902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,51200,1536,0.016249777542220224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1536,1024,0.013470222552617392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,1024,2048,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,16384,51200,0.1222391128540039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,768,0.03726844324005975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,6144,0.014974221587181091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,7168,0.022358222140206233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,1024,0.017456000049908955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,5120,0.09926399919721816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,1024,0.02368533280160692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,65536,2560,0.23595023155212402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,65536,0.14975289503733316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1536,768,0.005912000106440649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,512,0.0024844445288181305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,7168,0.011001778145631155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,10240,2048,0.12432177861531575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,65536,0.06390221913655598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,3584,0.014449778530332776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,10240,0.0188719994492001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,16384,32,0.04468355576197306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,256,0.0024684444069862366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,65536,0.35437599817911786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,5120,2560,0.008861333131790161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,7168,3584,0.01015733347998725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,8192,512,0.015613334046469795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,2048,0.013433777623706393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,3072,0.05375022358364529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,256,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,5120,0.029799110359615747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,5120,0.10133600234985352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,51200,0.20640799734327528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,2048,0.014877332581414117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,3584,0.015821332732836407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,512,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,768,1024,0.007630222373538547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,2048,0.01479288935661316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,8192,0.022088888618681166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,32,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,768,0.03237955437766181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,5120,0.017200888858901132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,3072,0.015412444869677225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,12288,0.06190311246448093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,4096,0.024284443921513025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,768,32,0.007925333248244392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,16384,0.02017244365480211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,256,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,2560,0.024911110599835713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,8192,0.07750844293170504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,65536,1024,0.3662231233384874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,10240,0.2964746687147352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2560,1536,0.024093333217832778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,4096,51200,0.07798666424221463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,6144,16384,0.041617777612474226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,7168,0.04605155520968967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1536,4096,0.00887022250228458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,1024,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,3584,0.03472177849875556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3072,3584,0.19567022058698866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,768,0.005580444302823808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,256,0.005601777798599667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,3072,0.11701688501569961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,12288,0.034308443466822304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,3584,1024,0.009309333231714036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1536,8192,0.015448888142903646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,64,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,3584,6144,0.026418666044871014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3072,6144,0.03420977791150411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1024,64,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,1536,128,0.005662222289376789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,5120,0.004592888885074192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,1536,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,3584,0.06247911188337538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,512,7168,0.019376888871192932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,16384,0.01984177695380317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,256,0.003863999827040566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,10240,7168,0.03457333313094245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,10240,0.08886577685674031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3584,1536,0.006313777632183499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,1536,51200,0.07234933641221789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,256,7168,0.017139555679427255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,10240,0.054901334974500865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,51200,0.023167999254332647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2048,8192,0.755177762773302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,1024,0.018911111685964797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,6144,0.011065777805116443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,10240,512,0.0196106665664249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,3072,0.005863111052248213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,128,0.06455377737681071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,12288,0.02035911050107744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,6144,0.045075555642445884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,8192,10240,0.07092089123196073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,16384,64,0.004520888957712385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,6144,0.36080177625020343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,768,0.017109332813156974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,768,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,4096,65536,0.26774666044447154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,512,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,6144,3584,0.013482666677898832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1024,7168,0.0321777794096205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,10240,0.059024002816942006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,4096,4096,0.06150844362046984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,128,256,0.0019519999623298645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,32,0.005986666513813867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,128,4096,0.01661422186427646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,4096,6144,0.0340320004357232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1536,4096,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,16384,8192,0.3252222273084852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3072,128,0.014594667487674289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,12288,256,0.03074666526582506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,1536,3584,0.023079999619060095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,128,0.0028693332440323303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,65536,256,0.2868826654222277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,6144,2048,0.05594044261508518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,3584,0.007265778051482306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,768,0.005720888988839255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,51200,0.17133422692616782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,8192,16384,0.046963555945290454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2560,10240,0.05695733096864489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,3072,0.013758222262064615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,1024,0.0038559999730851916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,12288,10240,0.32159821192423504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,2560,0.028140445550282795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,768,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,1536,7168,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,12288,0.36224977175394696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,32,0.003176888864901331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,1024,0.008006222546100616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,10240,0.028155555327733357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,128,0.001960888918903139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,1024,0.09733866982989842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,128,0.005976888868543837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,10240,6144,0.020941333638297185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,5120,0.008162666526105668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,3072,0.025719111164410908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,4096,0.035511109564039446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,10240,2048,0.031954665978749595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3584,5120,0.016720000240537856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,65536,0.14812800619337294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,6144,512,0.022622222701708477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,6144,3072,0.04411377840571933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,768,1024,0.005324444423119227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,10240,0.016394666499561734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,3072,0.008617777791288164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,51200,6144,0.1386515564388699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,10240,3072,0.030438221163219873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,6144,0.21418133046891955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,8192,0.013427555561065674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,4096,2048,0.07324888971116808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,768,6144,0.01924533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,64,0.029457777738571167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,6144,0.07022133138444689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,1024,0.0028799999919202593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,51200,0.11730044417911106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,7168,512,0.018012444178263348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,12288,0.08577422300974528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,16384,0.06849688953823514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,8192,5120,0.01759733259677887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1024,3584,0.007296889192528195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,128,0.009356444080670675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,8192,0.011434666812419891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,32,512,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,5120,0.03405244482888116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,64,64,0.0017893332988023758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,512,0.006206222292449739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,8192,0.02035999960369534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,8192,0.014072888427310519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,4096,0.04520977867974175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,512,16384,0.12667022811041936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,51200,0.04638311266899109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,51200,0.17108444372812906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,1024,0.03724799884690179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,6144,0.026878222823143005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,256,10240,0.011759111450778114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,16384,3584,0.05481333202785916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,32,0.0025128889828920364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,4096,10240,0.047206223011016846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,128,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,12288,0.030062221818500098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,65536,12288,0.1856764422522651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,8192,5120,0.05105955402056376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,1024,0.019783111082182992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,10240,0.17219466633266875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,256,0.00998133338159985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,7168,0.029226667351192896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,64,2560,0.015029332704014249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,2048,6144,0.03654488921165466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,64,1024,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,51200,0.04678666591644287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,64,1024,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,8192,0.03561688794030084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,512,1024,0.016582222448454965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,2048,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3584,512,0.002529777793420686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,128,0.002120888895458645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,8192,3072,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,128,0.022639999787012737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,32,4096,0.009863110880057016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,128,1024,0.00867555538813273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,3072,256,0.01620444489849938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,4096,0.7796097861395942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,6144,0.00997866690158844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,7168,7168,0.29988710085550946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,6144,0.015474667151769003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,10240,0.01240888900227017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,16384,0.8644275665283203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3072,12288,0.030238221089045208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,3584,0.011741333537631564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,3584,128,0.016158221496476066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,51200,0.18616178300645617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,256,0.007251555720965068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,10240,1024,0.0069253332912921906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,3584,0.019944889677895438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,8192,0.32687732908460826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,10240,0.05503999855783251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,8192,0.029703110456466675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,16384,8192,0.22392445140414766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,2048,1024,0.005602666487296422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,1536,0.007314666277832455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2048,64,0.005648889061477449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,2560,0.08702666891945733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,65536,1024,0.015513777732849121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,1024,0.01071466671095954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,7168,0.041703999042510986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,16384,16384,0.05024888780381945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,256,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,2048,0.017664000391960144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,256,0.006970666348934174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,3072,0.02288977801799774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,7168,3072,0.013167111410035027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,10240,768,0.004849777867396672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,1024,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3072,5120,0.008956444760163626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,3072,0.010886222124099731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,512,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,768,1024,0.010741333166758219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,7168,0.010010666317409938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1536,3584,0.021596444977654353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,8192,8192,0.026566222310066223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,2048,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,1024,0.022617777188618977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,768,0.015812445018026564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,1024,0.013111111190583972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,10240,0.08903555737601386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,10240,0.05259644322925144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,4096,768,0.01314222233162986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,6144,0.009658666948477427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,5120,0.012285333540704517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1024,2560,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,2048,0.0053022222386466126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,8192,4096,0.06985688871807523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,16384,0.030933333767784968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,3584,0.009215999808576372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,3584,2048,0.04150311152140299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,128,4096,0.014446222119861178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,256,0.0031635556370019913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,51200,0.07325244612163968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1024,7168,0.02609777781698439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,768,0.005949333310127258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,8192,0.005550222264395819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,5120,0.01310488912794325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,1536,4096,0.007803555164072249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,8192,0.037307557132509016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3072,2560,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,16384,12288,0.07254222366544935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,3072,0.25038666195339626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,2048,0.05782844622929891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,768,0.013883555928866068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,768,0.01512977812025282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,4096,0.006958222223652734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,16384,0.33232177628411186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,16384,0.020259555843141343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,65536,64,0.0039022221333450745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,6144,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,8192,0.02517866591612498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,65536,1024,0.018255111243989732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,1536,0.14279199971093073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,12288,0.04046133491728041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,12288,0.03664799862437778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,64,0.0021217777911159727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,6144,0.3444240093231201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,10240,256,0.006411555740568373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,128,0.018244443668259513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,4096,0.07014577918582492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3072,65536,0.49654579162597656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2048,1536,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,512,2048,0.018626666731304593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,51200,0.19234666559431288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,512,0.019932443896929424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,12288,0.062257779969109424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,3584,0.00813688917292489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,65536,2560,0.03339555528428819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,7168,1.2923155890570748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,8192,0.03825689024395413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,10240,51200,0.11254578166537815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,32,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,10240,3072,0.01405422223938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,256,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,3072,0.00812000036239624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,12288,4096,0.1631271176868015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,1024,0.005960888746711943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1024,32,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,7168,0.02404977712366316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,6144,12288,0.020000888241661918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,2048,0.15097778373294407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,2048,0.022980444961124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,768,0.035832888550228544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,3072,0.03136000037193298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,5120,0.01868799991077847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,256,0.01737955543729994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,1536,0.023028444912698533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,512,768,0.009329777624871995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,1024,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,3072,0.011703111231327057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,8192,0.04105866617626614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,1024,0.0035137778355015647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,8192,0.01647644407219357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,51200,128,0.020968889196713764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,128,0.018167111608717177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,8192,0.022506665852334764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,4096,768,0.0031431110368834603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,5120,3584,0.023345778385798138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,128,16384,0.025672889418072168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,51200,0.07181244426303439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,128,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,16384,4.442433675130208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2560,10240,0.11849510669708252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,3072,0.013160000244776407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,1024,0.053120000494851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,4096,6144,0.014075555735164218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,8192,0.07356266842948066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,3584,0.01069155583779017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,5120,10240,0.04122755593723721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,1024,0.017714665995703805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,768,0.018245332770877414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,51200,2048,0.04036888811323378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,4096,3584,0.007655999726719326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,64,0.00319377767542998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,5120,1024,0.0038417776425679526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,2560,16384,0.020403555697864957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,2048,0.007312888900438945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,8192,0.11319377687242295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,512,4096,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,51200,0.04619200030962626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,51200,0.21724977758195665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,768,51200,0.11559822824266221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,10240,0.19274844063652885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,256,0.0041146667467223275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,51200,0.28003644943237305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,8192,0.024051555328898962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,10240,5120,0.019256888164414298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,65536,0.1361066632800632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,128,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,10240,6144,0.06037511428197225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,3584,0.024449777272012498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,2560,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,768,0.030538668235143025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,51200,16384,0.2798231177859836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3584,65536,0.08954311079449123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,3072,0.008023111356629265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,512,0.011776888536082374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,10240,256,0.034840001000298396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,768,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,3584,0.015111999379263984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,64,10240,0.012823999755912356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,10240,0.03698488738801744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,512,51200,0.1880542172325982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,16384,0.022893332772784766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,10240,0.06668711370891996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2560,51200,0.47939199871487087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,256,0.012725333372751871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,1536,0.0665608909395006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,16384,2048,0.04382488793796963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,256,12288,0.034865776697794594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,4096,0.028220444917678833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,8192,0.07015911075803968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,1024,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,32,0.003535110917356279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1024,3072,0.03902311126391093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,768,1024,0.004182222402758068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,2560,0.007584888901975419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,1536,2560,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,8192,0.03976000017589993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,12288,128,0.02868266569243537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,3072,0.021225778592957392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,51200,0.05805777841144138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,5120,0.009630222287442949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,10240,0.026717333330048457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,32,1024,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3072,8192,0.053123556905322604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,10240,0.05332444442643059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2560,8192,0.05097600155406528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,2048,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,12288,32,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,1024,0.0059368887709246735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,5120,0.0461662213007609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,256,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1536,2560,0.017431111799346078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,64,0.016136000553766888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,10240,0.055416001213921436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,5120,0.1979964441723294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,2560,0.013074666261672974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,3072,0.006976889239417181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,16384,3584,0.014088888963063559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,2560,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,65536,128,0.02088177700837453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,10240,51200,0.12370488378736709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,256,2048,0.009558222360081142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,768,0.0177831103404363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,5120,0.010045333041085137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,8192,0.016394666499561734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,512,51200,0.12319021754794651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,10240,2048,0.009740444521109263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,16384,0.10736178027258979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,16384,0.03388266762097677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,3072,0.0086666668454806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,1024,0.003157333367400699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,1536,0.021533333592944678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,16384,0.17839999993642172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1536,3584,0.03431111243036058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,16384,2.792828453911675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,10240,0.016040000650617812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,16384,0.398144006729126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,1536,0.01292888902955585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,512,0.012064889073371887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,64,0.0035004446076022256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,128,7168,0.027853333287768896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,1536,0.026087111896938745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2560,6144,0.009533333281675974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,1536,0.006582222051090664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,1536,0.026833777626355488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,32,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,1536,0.013032889200581444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,3584,0.2344142331017388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,8192,768,0.0284222231970893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,1536,0.011743110915025076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,51200,0.06732622120115492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,32,12288,0.06245244211620755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,12288,0.052017778158187866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,51200,0.778897762298584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,5120,0.02687377731005351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,1536,0.05382133192486233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,1536,0.011687111523416309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,3072,64,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,8192,3584,0.08797955513000488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,51200,65536,83.71803114149306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,512,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,5120,0.3085964520772298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,6144,6144,0.027058665951093037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,7168,0.011855999628702799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,512,0.016184000505341422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,64,0.002838222309947014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2560,3072,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,8192,3072,0.017063111066818237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,7168,0.02507466740078396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1024,4096,0.008151999778217739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,32,7168,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2048,2048,0.007134222322040134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,10240,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,64,8192,0.014101333088344999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,128,0.012109333442317115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,32,51200,0.007272000114123027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,3584,0.04845244354671902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,5120,5120,0.055982225471072726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,512,0.013613333304723104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,4096,0.029225776592890423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,51200,0.6359813478257921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,5120,3584,0.0475777784983317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,6144,0.038755555947621666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,128,10240,0.037554668055640325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,51200,1.8892203436957464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,1024,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,16384,0.05738400088416206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,6144,0.17919199996524385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,256,0.08925155798594157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,256,0.012111110819710625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,3584,12288,0.17844622664981416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,65536,0.1977004475063748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,6144,10240,0.032912890116373696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,51200,1.0615164438883464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,51200,0.01834133267402649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3072,10240,0.01868888901339637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,768,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,7168,0.06373600165049235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,16384,6144,0.07725689146253797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,1024,0.015571556157535978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,7168,0.011208000282446543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,10240,0.016215110818545025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,256,1024,0.0035004446076022256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3072,7168,0.08951110972298516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,16384,32,0.00885333369175593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,51200,0.10209066338009304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,4096,0.008365333080291748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,768,51200,0.1033066643608941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,16384,0.3116755485534668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,65536,0.07778400182723999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1024,8192,0.010372444159454769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,768,0.005967111223273807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,256,0.02797244323624505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,12288,0.021330666210916307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1536,8192,0.02854222059249878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,12288,0.1510159969329834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,12288,0.017976888351970248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3584,128,0.0052871112194326185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,16384,0.033527112669414945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,128,0.08028533061345418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,7168,65536,15.818705240885416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,2048,0.37969599829779727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,3072,0.33110933833652073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,65536,512,0.22556887732611763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,256,0.004567999806669024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,512,0.016206221448050607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,64,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,128,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,256,3072,0.007607999775144789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,512,0.00831733312871721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,7168,0.024534222152498033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,3072,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,7168,0.00907199995385276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,10240,0.05112444361050924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2560,3584,0.05271555648909675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3072,5120,0.030780444542566936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,10240,65536,0.15377155939737955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,1024,0.00699733363257514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,2048,0.006908444480763541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,768,10240,0.02271466619438595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,5120,0.03195377853181627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,768,4096,0.02944355540805393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,5120,51200,0.08624444405237834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,2560,768,0.012518222133318583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,3584,0.05305155449443393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,8192,0.011356444822417365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,2048,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1024,3584,0.022312000393867493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,65536,1.1914372973971896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,3584,0.017856889300876193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,1024,0.020984888076782227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,4096,65536,0.07529511054356892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1024,2560,0.01201244443655014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,10240,0.016809778081046212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,2560,0.014612444572978549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,65536,0.08816444211535984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,16384,0.02125599980354309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,2048,0.02363377809524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,3072,0.016473778420024447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,8192,512,0.006308444258239534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,3072,0.012892444100644855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3072,12288,0.17412088976966011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,256,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,512,0.003198222153716617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,12288,4096,0.01883555617597368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,3584,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,16384,51200,0.13287466102176243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,7168,0.031090666850407917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,10240,6144,0.19068088796403673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,6144,0.020023110840055678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,2560,0.014543111125628153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,6144,0.02159822152720557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,3072,6144,0.012607999973826937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,768,768,0.00831911133395301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,51200,0.4604791005452474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,512,0.22793245315551758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,128,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,65536,2560,0.0281333327293396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,8192,0.056518223550584584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,4096,3072,0.017924444542990792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3584,512,0.006252444452709622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,5120,0.060066666867997914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,1536,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3584,51200,0.07210310962465075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,5120,0.0957840018802219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,16384,12288,0.06611022022035387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,8192,0.015081778168678284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,128,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,256,0.002491555487116178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,2560,16384,0.08755022287368774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,768,65536,0.06057955821355184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,65536,0.17659378051757812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,10240,8192,0.030511998467975195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3072,3584,0.007647111184067196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,3072,0.007256888680987888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,4096,1024,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,2048,0.014477334088749357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,768,0.006253333141406377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,8192,512,0.017656000124083627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,16384,64,0.022696000006463792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,64,256,0.0023306666149033443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,5120,0.00904266701804267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,4096,0.04411466585265266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,1536,0.029395557112163965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,3072,4096,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,6144,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3072,2048,0.0069120000633928515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,32,3072,0.007284444239404466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,5120,0.013424889081054263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1024,32,0.007956444389290279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,5120,10240,0.20179200172424316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,16384,5120,0.10785600211885239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,65536,0.04720888866318596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,8192,0.03973244296179877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,65536,10240,1.8326390584309895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,256,0.0021386666016446217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,6144,0.010703999963071613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,8192,0.027424888478385076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,2048,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,2560,5120,0.02188355558448368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,7168,0.03201955556869507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,32,65536,0.07433511151207818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,5120,32,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,64,0.08921600050396389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,51200,0.6106284459431967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,51200,3584,0.037556442949506975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,2048,8192,0.3777475621965196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1024,4096,0.010048000348938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,8192,1536,0.007585778004593319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,12288,64,0.020223110914230347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,1024,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,8192,65536,0.10075289011001587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,5120,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,64,512,0.0026791110220882627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1536,2048,0.016159999701711867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2048,16384,0.01964977714750502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,1536,0.011722666521867117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,7168,0.021301334102948506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3584,256,0.006583999842405319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,16384,3584,0.020629333125220407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,2048,7168,0.010739555789364709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,16384,0.5202746921115452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,1536,256,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,3584,12288,0.7613066567314996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,256,0.007298666569921706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,256,0.012892444100644855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,512,8192,0.059789333078596324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,4096,7168,0.01920088960064782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,7168,1024,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,2560,0.00490399988161193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,64,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,32,0.09727022382948135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,16384,0.027368000812000696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,1024,0.017847110827763874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,2560,0.028588443994522095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,3584,0.02369155486424764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,256,0.015153777268197803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,65536,1.2454924053615994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,1536,0.006292444550328785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,768,0.01872711049185859
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,3584,0.006261333409282897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,2560,0.007746666669845581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,4096,0.047296000851525195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,3584,0.03500533435079787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,1536,0.007361777954631382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,12288,2560,0.011347555451922946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,10240,0.012440000143316058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,10240,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2048,256,0.009339555270142024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,2560,0.006912888752089606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,768,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,768,0.007296000089910295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,2048,8192,0.04411288764741686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,65536,2560,0.03379022081693014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,12288,0.11725510491265191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,512,0.1585057841406928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,128,16384,0.11114133728875054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,128,0.014111111561457316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,3072,0.01018311083316803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,16384,768,0.020541333489947848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,128,65536,0.05061955584420098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,6144,1536,0.03257688879966736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,16384,0.03394844465785556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,256,128,0.0018373332503769135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,6144,0.008988444175985124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,65536,20.706285264756946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,5120,0.09375111262003581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,8192,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,7168,0.011189333266682096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,32,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,12288,7168,0.2943013244205051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,768,16384,0.016499555773205228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,768,1536,0.0048071112897661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2048,768,0.01758044461409251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,768,128,0.008363555702898238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,6144,0.04282844397756788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,512,65536,0.0842026670773824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2048,768,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,8192,0.017679111825095285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,7168,0.13843111197153726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,16384,0.0702400008837382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,8192,0.16686577267116973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,7168,0.029140445921156142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,768,0.01553600033124288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,768,0.0075839997993575195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,12288,0.041540443897247314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,2048,0.20400710900624594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,12288,7168,0.019928000039524503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,5120,0.02197333342499203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,32,12288,0.013441777891582914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,256,0.014075555735164218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,4096,0.04760444495413038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,768,0.08378666639328003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,65536,0.09886666801240708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,768,0.003630222131808599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,6144,4096,0.05499644411934746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,1024,0.04064711266093784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,2560,0.013815110756291283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,6144,2048,0.05798133214314779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,51200,0.03700444433424208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,256,0.0024666666156715816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,64,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,65536,1536,0.04042844308747186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,12288,0.013325333595275879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,3072,256,0.004909333255555895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,51200,10.182676527235243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,5120,0.1373573276731703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,51200,51200,0.9234480328030057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2048,3072,0.007114667031500075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,768,0.02084355552991231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,5120,0.018384888768196106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,7168,65536,0.09178133143319024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,5120,256,0.011735999749766456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,4096,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,16384,65536,2.3751369052463107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,5120,6144,0.026519111461109583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,51200,7168,0.06724444362852308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,256,8192,0.011054221954610614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3584,512,0.022664000590642292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,4096,3584,0.12621600098080105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,10240,32,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,4096,0.006541333264774746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,512,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,8192,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,12288,10240,0.1584284438027276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,3584,0.007616889145639207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,2560,1024,0.004575110971927643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,4096,1536,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,2560,0.015531554818153381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,3584,128,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3072,10240,0.16302933957841662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,7168,0.01036266651418474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,51200,12288,0.0839395523071289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,128,0.01643200053109063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2560,3584,0.008964444200197855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,7168,0.009986666341622671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,3584,0.05191733439763387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,3584,0.007661333514584436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,768,0.014185777968830533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,51200,0.17764888869391549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,65536,10240,0.16560978359646267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,16384,16384,0.04457155532307095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,64,0.15285332997639975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,2048,3072,0.021985777550273474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,16384,0.44406488206651473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,32,0.012793777717484368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,1024,0.006239111224810283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,16384,0.14489155345492893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,3584,6144,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,12288,0.0492524438434177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,51200,0.22657778528001574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,1536,0.013428444663683573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,6144,0.07247910896937053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,768,16384,0.018626666731304593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,8192,0.06595200300216675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,1536,768,0.0069671107663048645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,768,0.005744888964626525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,256,10240,0.02385066615210639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,6144,5120,0.05064533485306633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,4096,0.015490666031837463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,12288,128,0.011898666620254517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,3072,0.016529776983790927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,65536,0.06307999955283271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,12288,0.1420604387919108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,8192,3072,0.035848001639048256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,51200,0.2530311213599311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,6144,0.07523822122149997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,1536,0.10836978091133966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,12288,1536,0.028176890479193792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,5120,0.044545776314205594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,10240,0.034239110019471906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,12288,0.013502221968438892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1024,256,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,8192,7168,0.017651556266678702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,3072,0.015319110618697273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,10240,51200,3.8999752468532987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,256,0.023991111252042983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,6144,0.01924088928434584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,64,0.006895111252864202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,65536,0.931181377834744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,32,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1536,2560,0.01548266741964552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,512,0.011436444189813403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,2048,0.018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,51200,0.7819200091891818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,8192,0.040608889526791044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,2560,0.01404622197151184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,10240,128,0.027440889014138117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,1536,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,2560,0.007655111451943715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,128,65536,0.08340444167455037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,65536,0.033915556139416166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,3072,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,768,0.002522666628162066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,256,0.009300444689061906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,512,0.00657066661450598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,7168,0.04464266697565714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,1024,10240,0.04730044470893013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1024,65536,0.3452577855851915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,51200,2.5898089938693576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,12288,0.17834933598836264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,1024,0.006129777679840724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,2048,0.04094666573736403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,256,1536,0.038093331787321306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3584,256,0.012759110993809171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,5120,12288,0.20512710677252877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,6144,0.004600888739029567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,4096,0.02088533341884613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2560,8192,0.05319999986224704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,65536,0.21381866931915283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,3584,0.029437333345413208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,3072,128,0.019634667370054457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,4096,0.006004444426960415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,51200,0.7128924263848199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,6144,0.0210533340771993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,3072,0.013073777986897362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,2048,0.009860444400045607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,3584,3072,0.011338666909270816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1024,1536,0.006969778074158563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,3072,0.29722844229804146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,65536,7168,0.06638666656282213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,128,6144,0.015746666325463187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,1536,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,10240,6144,0.1924755573272705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,65536,1.0743991004096138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,4096,0.08079200320773654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,12288,8192,0.05369244350327385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,7168,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,51200,2.0510995652940536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,6144,0.08474489053090413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,512,1024,0.0181697772608863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,12288,0.08478577931722005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,4096,16384,0.06384266747368707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,6144,0.03757866554790073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,3072,0.04747733473777771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,2048,2048,0.0273964438173506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,768,1536,0.009089777866999308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,5120,0.021406221720907424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,128,0.003160888950030009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,4096,0.00999111102686988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,4096,0.015792000624868605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,3072,0.02570399973127577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,1024,0.013743999931547375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,51200,0.007672888537247975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,10240,0.18786844942304823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,32,0.004234666625658671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,65536,3584,1.047228495279948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,12288,0.028024001253975764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,8192,0.06184622314241198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,16384,0.05885777870814005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1536,3072,0.021640888518757288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,16384,0.06524355543984307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2560,512,0.020255110330051847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,2048,0.0069751110341813825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,51200,0.13775289058685303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,2048,16384,0.07406310902701484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,7168,0.02093066606256697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,64,256,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,65536,0.08302222357855903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,10240,6144,0.05850578016704983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,2048,8192,0.020503110355801053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,3072,0.01588088936275906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,10240,3072,0.09650311205122207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,512,128,0.015223110715548197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,3072,0.04410399993260702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,51200,0.6027973492940267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,65536,0.44078932868109805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,5120,0.012898666991127862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,768,0.019319110446506076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,51200,0.03805511196454366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,16384,4096,0.18642132812076143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,6144,0.03220977716975742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3072,2560,0.046889778640535146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,8192,0.05897688865661621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,5120,0.15927199522654215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,51200,16384,42.16026814778646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,6144,0.0866968896653917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,51200,6144,1.2592755423651802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,3584,0.010288889209429422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2048,3584,0.01587466730011834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,16384,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,16384,0.24624710612826875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,4096,0.07472178008821276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3584,2048,0.019744000501102872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,512,65536,0.0563653310139974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,16384,12288,0.04986133178075155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,10240,768,0.007930667036109501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,5120,2560,0.04574311110708448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,12288,0.055734223789638944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,5120,7168,0.12019733587900798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,2560,3584,0.034437331888410784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,3072,0.0317795541551378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,7168,768,0.011367110742463006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,10240,0.15995111730363634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,65536,0.5630053414238824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,4096,12288,0.018773333893881906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,512,0.026791110634803772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,10240,51200,0.20683111084832084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,5120,0.05883200301064385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,128,3584,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,6144,0.02199733257293701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,1536,0.006633777585294511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,8192,0.030885332160525854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,6144,6144,0.01664622293578254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,3072,0.0852924452887641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,51200,0.07116088602277967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,1536,0.008107555409272512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,3584,0.0075804442167282104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,8192,0.024009777439965144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,51200,2048,0.03943644298447503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,51200,0.05137955480151706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,6144,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,1536,12288,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,51200,1.3309289084540474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,4096,2560,0.011227555572986603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,3072,0.15154843860202366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,10240,0.027887112564510767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,768,1536,0.010870222416188983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,256,3584,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,10240,0.17106933063930937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,4096,3584,0.10563021898269653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,768,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,16384,0.13916977246602377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1536,1536,0.01091644416252772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,10240,10240,0.12000000476837158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,16384,12288,0.31240177154541016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,1024,8192,0.08951733509699504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,32,0.0018399999373488957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,1024,0.005955555372767978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,512,1024,0.0028782222006056043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,3072,0.32446932792663574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,512,0.04873155554135641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,6144,0.020430222153663635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,6144,256,0.018926221463415358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,1024,0.0063288890653186375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,8192,0.015925332903862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,64,0.0018311111877361934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,768,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,512,0.01960088809331258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,16384,0.02088622252146403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,5120,16384,0.04045866595374213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,8192,6144,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,2560,0.013459555804729462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,512,0.12149422698550755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,32,0.0053084443012873335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,51200,0.16344800260331896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,768,6144,0.11572710673014323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,256,0.01703555550840166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,4096,32,0.09420000182257758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,12288,32,0.033490667740503945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,10240,3072,0.1094355583190918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,8192,10240,0.10830489132139419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,1536,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,256,0.005995555470387141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,51200,0.03799466623200311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,51200,0.1976177824868096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,5120,64,0.00563466673096021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,1024,0.015846222639083862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,3584,32,0.008254222571849823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2048,256,0.002458666761716207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,2048,0.004557333472702238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,6144,1536,0.02199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,4096,128,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,128,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,2560,0.007306666837798224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,1024,0.017837334010336135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,8192,0.04986400074428982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1024,65536,3.3677378760443797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,7168,12288,0.022568888134426538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,256,768,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,12288,0.06706488794750638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,1536,0.027309333284695942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,1536,65536,0.19425066312154135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,1024,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,16384,0.015820443630218506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,3584,0.4835235807630751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,5120,512,0.02383377816942003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,51200,0.22062222162882486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,6144,0.049152000082863696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,16384,7168,1.7771404054429796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,768,7168,0.0099982221921285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,7168,1024,0.014463999205165438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,32,51200,0.039206223355399236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,6144,5120,0.03289422061708238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,3584,0.034809778134028115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,2048,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,768,0.002571555475393931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,2560,0.1586364507675171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,65536,0.0668462249967787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,16384,0.017998221847746108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2560,7168,0.08111021916071574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,5120,1536,0.02514222264289856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,51200,5120,0.06538044744067721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,512,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,65536,1536,0.02480977773666382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,10240,0.03380889031622145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,6144,0.03085244364208645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1536,768,0.01868088874551985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,4096,256,0.015497777197096082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,768,2560,0.038013332419925265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,128,0.002165333264403873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,12288,7168,0.1734773317972819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,768,8192,0.027476444840431213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3584,64,0.005712888720962737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,5120,0.01275288893116845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,128,3072,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,4096,0.058099554644690625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,12288,0.03048355711830987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,6144,0.02607733342382643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,16384,0.14395556184980604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,10240,0.023702222439977858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,32,768,0.0026942222482628296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,7168,256,0.017887999614079792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,12288,0.0226239992512597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1536,10240,0.010046222143703038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,51200,1024,0.025834666358100042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,51200,10240,0.13291911284128824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,51200,0.06221155325571696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,10240,65536,0.20302489068773058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,2048,10240,0.10270133283403184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,7168,0.17445066240098742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,5120,0.03418666786617703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,6144,128,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,2048,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,65536,32,0.07178933090633817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,768,0.02973777718014187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,2048,0.09032622310850356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,10240,0.31143999099731445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,16384,3072,0.017800889081425138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,768,128,0.005600000007285013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,768,0.05069422059588962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,10240,0.06582311126920912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,65536,0.09686755471759373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,1536,2560,0.02134311033619775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,64,256,0.006638222270541721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,12288,10240,0.1797644429736667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,256,0.015174221661355762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,128,0.010048000348938836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,65536,0.9423448774549695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,3072,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,4096,4096,0.01239911135700014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2560,256,0.007828444242477417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,8192,0.015092444088723926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1536,3584,0.04344711038801405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,768,0.02072177827358246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,4096,0.1193679968516032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,512,0.05187288920084635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,512,32,0.0018337778747081757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,2560,0.020626667473051283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,16384,3072,0.18908888763851592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1024,32,0.0028817777832349143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,64,16384,0.018239110708236694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,512,7168,0.012478222449620565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,256,10240,0.013456888496875763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,7168,0.0458497769302792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,16384,2560,0.02408711115519206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,512,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,128,3072,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,2048,0.03250933355755276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3072,6144,0.05976088841756185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,4096,0.008295111358165741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2048,10240,0.014153778553009033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,64,0.001842666624320878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,256,0.005305777821275923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,768,0.009034666750166152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,1024,1536,0.013078221844302284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,256,10240,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,768,0.018561777141359117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,8192,0.03615466753641764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,256,10240,0.020641777250501845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,3072,0.22275021341111925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,512,0.008601778083377415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,32,0.0018071112119489247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,4096,0.12718221876356336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1536,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,4096,16384,0.32095111740960014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,1536,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2048,7168,0.027227555712064106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,5120,2560,0.05748889181349012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,2560,0.07461511426501803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,1024,0.01683644453684489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,512,0.048729777336120605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,16384,0.03586133321126302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,256,16384,0.017768889665603638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,768,65536,0.13069155481126574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,16384,0.009466666314336989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,256,2048,0.004237333519591225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,1536,0.08912889162699382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,512,1024,0.03098755412631565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,5120,0.04678666591644287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,16384,2048,0.012347555822796293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,2048,0.004576000074545543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,128,512,0.0406275557147132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,256,5120,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,12288,0.35332711537679035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,4096,0.015119999647140503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,256,0.012093333734406365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,12288,768,0.006690666493442323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3584,5120,0.021776888105604384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,8192,4096,0.013542222479979197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,128,0.09488444195853339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,768,0.018144889010323417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,3584,4096,0.010267555713653564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,51200,0.11378578344980876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,16384,2.420166227552626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2048,65536,0.6055155330234104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,1536,512,0.017851556340853374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,12288,0.14041511217753092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,12288,3072,0.08923555745018853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,7168,5120,0.21313511000739205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,5120,0.03106400039460924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,256,0.012219555675983429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,512,1024,0.005584888988071018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,51200,0.1409333282046848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,1536,0.011772444678677453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,128,8192,0.018556444181336295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,4096,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,128,10240,0.018449778358141582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,2560,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,16384,6144,0.01977511081430647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,10240,0.15555555290646023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,2560,1024,0.005964444329341252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,64,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,768,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,16384,0.02306488818592495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,1024,0.1148408916261461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,128,0.010471999645233154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,65536,0.018007111218240526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3072,3584,0.05295377638604906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,65536,128,0.010371555884679159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,256,0.013574221895800697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,16384,1.1467129389444988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,32,1536,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,65536,0.06774755318959554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,12288,0.01481244464715322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,65536,2048,0.24586934513515898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,3584,0.2014231152004666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,7168,0.17394399642944336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,7168,0.011002666420406766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,128,32,0.006256888724035687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,2048,65536,0.45410754945543075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,3072,0.016491555505328707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,256,7168,0.02027466727627648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,1024,0.006276444428496891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,65536,6.82137934366862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,65536,256,0.010270222193664974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,8192,0.019666666785875957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,32,0.001496000008450614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,10240,256,0.003825777934657203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,65536,0.05386666787995232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,1536,256,0.014780445231331719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,8192,0.01441333360142178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,7168,0.016213332613309223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,51200,0.7279786533779569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,2048,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3584,6144,0.013080000049538083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,5120,0.1275564432144165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,3072,2048,0.022710222336981032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,1024,0.007893333004580604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,5120,0.016134222348531086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,3584,12288,0.1317680014504327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,65536,1536,0.40930133395724827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,65536,0.17832888497246635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,3072,0.007849777738253275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,16384,0.04176888863245646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,65536,1024,0.014862222803963555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,16384,0.04969066712591383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,6144,0.009304000271691216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,64,0.001867555495765474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,10240,0.02347822156217363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,2048,0.007594666547245449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,1536,0.0069066666894488865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,2560,0.01648000048266517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,128,0.018216000662909616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,8192,1024,0.011677333050303988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,10240,128,0.007652444144090016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,6144,12288,0.027162666122118633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,10240,0.08364266819424099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,3584,0.004307555655638377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,2048,0.009331555830107795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,32,0.008126222425036961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,256,0.01920355525281694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,64,1536,0.0032231110251612137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,3584,0.00793422261873881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,256,0.0075697774688402815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,65536,768,0.01646933290693495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,8192,51200,4.007862091064453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,3584,0.020524443851576913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,1536,0.0069520001610120135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,16384,2560,0.02084622283776601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,10240,0.029362665282355413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,6144,64,0.019211555520693462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1536,12288,0.03464711043569777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,8192,0.009371555513805812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,3584,0.01953155630164676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,1536,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,8192,0.004516444272465176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,8192,6144,0.014076444837782117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,1024,16384,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,10240,0.11401689052581787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,65536,7168,0.08057866493860881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,64,5120,0.02090666691462199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,65536,0.34079021877712673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,64,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,512,0.021167111065652635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,32,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,3584,2560,0.007240888973077138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,4096,128,0.0058959999846087555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,2048,0.06479733520083956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,1024,0.016215110818545025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,51200,1536,0.05980889002482096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,65536,5120,0.8234320216708713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,128,5120,0.009025777379671732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,768,0.007380444142553542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,10240,1.6632035573323567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,1024,0.0295448899269104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,768,0.007183111376232571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,7168,4096,0.015489778584904142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,256,3072,0.0069644442862934535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,2048,0.08282399839825101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,51200,2560,0.04765422145525614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,7168,0.0317831105656094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,2048,512,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,16384,3072,0.024655110306209985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,8192,0.016843555702103507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,8192,3584,0.13367556201087102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,256,256,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,10240,0.07243288887871636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,1536,0.030671109755833943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,768,0.007648000286685095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,4096,0.017418666018380057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,6144,768,0.019444444113307528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,12288,16384,0.04862044586075676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,128,51200,0.11752622657352024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,16384,2560,0.0664079984029134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,51200,0.5169359842936198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,65536,0.09005955855051677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,8192,0.030413332912656996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,128,12288,0.023021333747439917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,10240,0.08832177850935195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,512,0.028466665082507666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,12288,10240,0.38671289549933535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2048,51200,0.04939377639028761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,4096,7168,0.11387111081017388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,16384,7168,0.06526399983300103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,6144,0.024392000503010217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,768,6144,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,65536,6144,0.13283199734157985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,4096,0.0329715543323093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,8192,0.23024000061882866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2560,2048,0.004701333327425851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,16384,51200,0.11542044745551215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,4096,0.01667022208372752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,512,3584,0.011024889018800525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,8192,16384,0.027144889036814373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,3584,0.03718488746219211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,768,0.017834666702482436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1536,51200,0.051912890540228955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,1536,7168,0.011754666765530905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,2048,8192,0.5907582177056206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,65536,1.3705369101630316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,32,0.006247111078765657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,6144,0.012897777888509961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,64,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,512,0.010645333263609143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,2048,0.008110221889283922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,12288,0.041971554358800255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,768,0.009666666388511658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,3584,0.0075457774930530125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,768,0.24423199229770234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,4096,10240,0.017647110753589206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,7168,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,8192,0.059501330057779946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,65536,2.0731093088785806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,32,0.01959111127588484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,7168,2560,0.0369137790468004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,1024,0.017267554998397827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,8192,32,0.016657777958446078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,51200,0.38692532645331484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,16384,0.22951645321316191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,256,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,7168,7168,0.017532444662517972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,128,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,12288,0.006981333096822103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,2048,4096,0.06545422474543254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,5120,0.558309343126085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3584,4096,0.013096888860066732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,512,7168,0.08039644691679212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2560,65536,0.10760177506340875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,65536,3072,1.5090817345513239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,2048,0.006683555742104848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,4096,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,5120,65536,0.3947173224555121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,1024,51200,0.062493330902523465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,4096,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,12288,0.09307289123535156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,768,0.008223111430803934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,51200,0.054514666398366295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,2560,2048,0.21385599507225883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,768,32,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,2560,2560,0.016791999340057373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,16384,5120,0.021457778082953557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,256,16384,0.03831911087036133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,768,2048,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,256,51200,0.08817511134677464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,65536,3.4865404764811196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,128,3584,0.01036533299419615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,3072,0.05340977840953403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,65536,7168,0.9971546596950955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,6144,0.029715554581748113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,512,0.012796444197495779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,1536,3072,0.20600799719492593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,5120,5120,0.09381333324644302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,16384,3584,0.019848000672128465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,256,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,128,256,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1536,512,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,3584,0.023944000403086346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1536,1024,0.008847999903890822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,512,0.0075786668393347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,10240,2560,0.012834666503800286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,12288,7168,0.023781332704755995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,128,0.005932444499598608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,2048,8192,0.036533332533306546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,1024,1.9843084547254775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,51200,0.06515466504626803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,768,4096,0.01663377715481652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,4096,0.009007111191749573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,1536,0.012989333106411828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2560,1024,0.013595555391576556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,16384,256,0.002509333400262727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,6144,2048,0.0440924432542589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,4096,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,64,0.0017680000099870895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,16384,0.04832622077729967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,7168,0.03089155422316657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3072,3072,0.02243822150760227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,256,0.008671110702885522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,10240,0.015782222151756287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,51200,0.04308088951640659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,51200,0.40851200951470273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,7168,0.0767351124021742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,768,2560,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,65536,1024,0.2137697802649604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,128,5120,0.02029244436158074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,768,2048,0.018033777674039204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,3072,0.016944888565275405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,7168,0.010683555569913654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,51200,7168,0.0648239983452691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,10240,128,0.032551109790802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,4096,0.5980870988633897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,12288,0.07649599843555026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,12288,2048,0.03571733170085483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,10240,0.028512001037597656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3072,16384,0.026208889153268602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,16384,2048,0.0685359968079461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,1024,0.0328106681505839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,2560,2560,0.04483111037148369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,4096,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,12288,3584,0.01351999988158544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,4096,8192,0.015089778436554803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,5120,0.034133334954579674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,16384,0.032404445939593844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1024,51200,0.05032444331381056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,12288,0.03573688864707947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,12288,128,0.006299555715587404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,1536,512,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,512,0.006255110932721033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,5120,5120,0.05379022161165873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,6144,0.0879582232899136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,4096,0.07603733407126533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,51200,0.041723555988735624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,3072,2560,0.019927110936906602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,16384,0.0488026671939426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1024,32,0.005956444475385878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1024,2560,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,32,2048,0.005199999858935674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,7168,16384,0.028142223755518597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,4096,0.012727110750145383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,16384,51200,1.3008168538411458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,2048,0.0158906661801868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,64,0.002875555513633622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,256,0.02089333368672265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,3072,0.014671110444598727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,65536,0.29634756512112087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,51200,0.04927644464704725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,6144,0.014769777655601501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,2048,0.011716444459226398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,51200,5120,0.19944800270928276
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,7168,2560,0.02475022276242574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,5120,0.04923644330766466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,12288,128,0.021943999661339655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,128,0.0052586665583981415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3584,32,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3072,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,4096,6144,0.010608889162540436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,8192,16384,0.03862844573126899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,12288,128,0.15456177128685847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,1024,0.021993777818149988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,2560,0.025416889124446448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,3072,0.00657155571712388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,5120,128,0.011367999845080905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,12288,32,0.006242666807439592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,51200,3584,0.06260266568925646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,2048,51200,0.06810666455162896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,2560,0.012051555845472546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,2560,0.01237511138121287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,65536,0.05704889032575819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,128,2048,0.009678222239017487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,7168,0.035805334647496544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,51200,0.051570667160881885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,65536,128,0.00518844442235099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,12288,0.04150311152140299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,1024,0.016653334101041157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,256,0.013704000247849358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,256,4096,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,5120,0.010069333016872406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,1536,0.014431110686726041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,64,1536,0.01073688848151101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,32,2048,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,51200,0.07099466853671603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,256,0.013680000272062091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,768,0.0075795559419526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,12288,512,0.028703110085593328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2560,7168,0.014873777826627096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,1536,7168,0.010787555740939246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,3072,0.26855021052890354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,5120,6144,0.01093155559566286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3072,768,0.005655111124118169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,2048,0.004543999830881755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,256,51200,0.0448968874083625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,10240,0.014144889182514615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,256,8192,0.06615200307634142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,2048,0.024919110867712233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,6144,4096,0.011318222516112857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,3072,0.018208000395033095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,256,16384,0.12370933426751031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,3584,0.0064604443808396654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,5120,64,0.005261333452330695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2560,768,0.010061333576838175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,3072,0.06693333387374878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,1536,0.020243555307388306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,7168,0.03941866755485535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,7168,0.05173866616355049
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,1024,3584,0.019296889503796894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,10240,0.018403554956118267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,51200,2048,0.5606248643663194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,65536,0.5318960083855523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,65536,0.07905066675610013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,256,2048,0.005616888817813661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,10240,1024,0.009682666924264696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,1024,256,0.0028133332315418455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,12288,0.037296000454160906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,10240,0.0203484445810318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,1024,0.004862222406599256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,768,0.007033777733643849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,65536,0.06468355655670166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,7168,2048,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1024,1536,0.022477333744366963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,768,0.014460444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,65536,1024,0.031183110343085393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,65536,0.07719288931952582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,3584,0.023504889673656885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,256,2048,0.01314133322901196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,4096,0.2678399880727132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,6144,2560,0.021572444174024794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,5120,1536,0.04927644464704725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,512,0.0052897776994440295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,6144,64,0.008316444853941599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,4096,0.010687111152542962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,4096,0.07641688982645671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,3072,2560,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,6144,0.012834666503800286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,3584,0.016171556380059984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,7168,0.2201697826385498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,51200,3072,0.15491555796729192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,4096,51200,0.11386577288309734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,12288,0.041082666979895696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,7168,0.029345777299669053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,5120,0.059524443414476186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,32,12288,0.013417777915795645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,51200,0.04929866724544101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,10240,6144,0.02111111084620158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1536,4096,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,512,65536,0.07956444554858737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,7168,1536,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2560,3072,0.03156088789304098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,8192,51200,0.10417955451541477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,2560,768,0.025960889127519395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,1024,10240,0.06524266799290974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,6144,65536,0.709096007876926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2048,12288,0.03834755553139581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,4096,768,0.026925333672099646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,4096,0.011371555427710215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,12288,1.471917364332411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,256,64,0.0029760001020299066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,7168,0.011025778121418424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,2560,3072,0.007886221839321984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3072,16384,0.07636888821919759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,7168,0.009983110758993361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,6144,0.23616443739997017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,2560,0.11568266815609402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,512,4096,0.015303999185562134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,10240,0.01645511057641771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,7168,0.09981511036554973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,65536,0.06646844413545397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,51200,16384,22.99814012315538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,51200,7.068669213189019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,8192,0.010692444112565784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,4096,0.00798044436507755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,768,1024,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,2560,4096,0.022618666291236877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,5120,128,0.005649777750174205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,51200,64,0.1582204500834147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,128,0.002519999941190084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,512,0.011406222151385413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,7168,256,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,8192,512,0.0076142218377855085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,2048,0.012303111453851065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,12288,1536,0.008402666283978356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,8192,3584,0.06359822220272489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,6144,3584,0.013832888669437833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,3072,0.007992000215583378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3072,1536,0.006628444625271692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,8192,0.16056977377997503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,1536,0.037532445457246565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,10240,0.006440889090299606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,512,0.002493333278430833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,768,128,0.0021608887861172357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,12288,0.013458666702111563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,51200,0.23935111363728842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,12288,7168,0.023687111006842718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,8192,0.014845333165592618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,65536,1536,0.07427199681599934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,4096,128,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,65536,7168,0.13264444139268664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2560,3584,0.007618666523032718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,3584,0.05013777812321981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,128,0.0019004444281260173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,256,0.008005333443482717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,5120,128,0.006933333145247565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,51200,1.038126203748915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,1536,512,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,16384,128,0.0021253333737452826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3072,2560,0.008234666453467475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,12288,0.2573911084069146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,7168,10240,0.48870664172702366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3584,3584,0.028034667174021404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,2560,0.012367111113336353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,10240,0.008608000146018134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,2560,0.00778399987353219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,768,16384,0.017128000656763714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,2560,0.005251555393139522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,12288,1536,0.00829777783817715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,4096,0.01608444419172075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,5120,4096,0.026921777261628047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,768,5120,0.0192275560564465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,4096,65536,0.9334292941623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,1024,0.020248000820477802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,7168,0.015294222368134392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,4096,51200,2.59117062886556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,12288,0.07999822166230944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,8192,0.05413066678576999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3072,51200,0.05337599913279215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,12288,2048,0.009900444083743626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,65536,10240,0.181167999903361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,3584,3584,0.02770400047302246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1024,8192,0.010727110836240979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,2560,0.008621333373917473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,3584,0.011791110866599612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1536,8192,0.024146666129430134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,768,0.004870222260554631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,3072,256,0.014817777607176038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,128,0.020976889464590285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,32,0.006624889042642381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,51200,2560,0.21169511477152506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,64,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,4096,1024,0.018385777870814007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,16384,0.09472977452807957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,12288,65536,1.5585341983371312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,10240,128,0.005207111024194294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,256,6144,0.041111111640930176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,12288,7168,0.0547617773214976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,128,256,0.006573333508438534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,7168,7168,0.04651200109057956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,51200,0.06142489115397135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,3072,0.007658667034573025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,10240,1536,0.02299822204642826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,128,0.005586666779385672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,12288,0.0679360032081604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,768,0.00756444450881746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1536,64,0.0018168888572189545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,768,256,0.010114666488435533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,7168,1536,0.007285333342022366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,12288,0.3070124520195855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,5120,5120,0.04753866791725159
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,64,8192,0.006670222100284364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,10240,0.042176001601748996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,51200,1.0235164430406358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,3584,2048,0.04549511273701986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,5120,3072,0.010633777413103314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,51200,0.056187556849585645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,2560,1536,0.03355200091997782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,8192,12288,0.035451554589801364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,7168,0.04052177733845181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,64,0.004171555654870139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,1024,0.02386933399571313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,128,768,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,7168,0.17909867233700225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,128,0.005986666513813867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,2048,256,0.0021626665774318907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,8192,3072,0.00997155573632982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1024,65536,5.02913581000434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,12288,0.33090488115946454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3584,7168,0.012541333006487953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,3072,0.0701671110259162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,4096,768,0.006121777825885349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,64,16384,0.02337155573897892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,8192,0.024136000209384497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,256,6144,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,6144,0.00868622213602066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,51200,16384,0.19341156217787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,512,128,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,7168,0.1130853361553616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,2048,0.00664533343580034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,4096,0.007989333735571967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,5120,128,0.002508444504605399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,256,4096,0.012074666718641916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,3072,0.010646222366227044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,32,0.0059262220230367445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,7168,0.012959111067983838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,51200,0.6316755612691244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,6144,512,0.004559999952713649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,3584,0.029677334758970473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,2560,0.007300444775157505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,3584,5120,0.030971556901931763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,256,0.0244186669588089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,32,64,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,2048,3072,0.027256000373098586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,8192,0.07095199823379517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,512,0.014539556370841132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,256,768,0.0038488888078265717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,256,768,0.005717333406209946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,768,1536,0.014480888843536377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,5120,2560,0.022800000177489385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3072,512,0.025389333566029865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,10240,0.01205599970287747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,2048,1536,0.03464622298876444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,12288,3072,0.028112000889248315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,51200,64,0.009697777529557547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,12288,16384,0.04304800099796719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,256,64,0.0018035556293196147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,6144,0.024858666790856257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,7168,12288,0.25216354264153373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,64,5120,0.023694222172101338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,10240,1536,0.021174222230911255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,12288,0.020098666350046795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,256,768,0.0032133333798911837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,7168,0.01902311046918233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,16384,51200,0.12698045041826037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,3072,0.01719199948840671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,6144,1536,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,65536,16384,0.3798844549391005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,65536,0.04942399925655789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,768,2048,0.006888000087605582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2048,7168,0.023896889554129705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2048,128,0.011040888726711273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,7168,0.05893333090676201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,512,1536,0.007271111011505127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,16384,6144,0.07318133115768433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,2560,0.22959823078579375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,512,256,0.006955555743641323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,51200,0.44386132558186847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,256,0.01958311100800832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,7168,32,0.15099644660949707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,768,1536,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,10240,512,0.022408000297016565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,6144,65536,0.22038577662573922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2560,4096,0.018858666221300762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,51200,0.7069733407762316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,6144,0.016647110382715862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2560,2048,0.021928000781271193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,128,0.00314666661951277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,65536,8192,0.0714942216873169
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,3584,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,6144,10240,0.03743022349145677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,2048,0.0818106664551629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,3584,0.017881777551439073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,65536,3584,0.08270221948623657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,4096,0.027866668171352808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,8192,512,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,3584,0.10897066858079697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,2048,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,1536,0.09694666994942559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,12288,16384,0.04619377851486206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,51200,0.1396204498079088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,1536,7168,0.2886364459991455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,2048,0.042859554290771484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,10240,8192,0.049639112419552274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,512,16384,0.016255999604860943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,7168,768,0.047056890196270414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,1536,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,7168,2560,0.05051999953058031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1536,1536,0.02238222294383579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,8192,0.011351111034552256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,2048,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,6144,0.029354665014478896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,10240,2560,0.011368888947698804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,2048,65536,0.07897866434521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,3072,0.022520888182852004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,2048,0.052961776653925575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,3072,1024,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,512,0.020615110794703167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,12288,0.01588800052801768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,512,3072,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,6144,0.040229333771599665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,1024,0.04536266790495979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,256,2560,0.020630222227838304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3072,16384,0.2627306779225667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,8192,6144,0.016750221451123554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,256,256,0.014766222900814481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,10240,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,256,2560,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,5120,65536,0.12123110559251572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,3072,0.03055022160212199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2048,12288,0.1325546635521783
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,128,16384,0.03696977761056688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,51200,0.03843288951449924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,256,6144,0.008387555678685507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,768,12288,0.04461600051985847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,6144,0.09215644333097671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,8192,0.241920895046658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,10240,10240,0.0361057784822252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,5120,0.02914311157332526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2048,7168,0.0270471109284295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3584,5120,0.012362666428089142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,10240,0.030828442838456895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,4096,0.010478222535716163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,2560,0.015443555182880826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,4096,0.008723555339707268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,16384,2048,0.02994489007525974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,8192,768,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,768,512,0.014785778191354541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,2048,0.006937777830494775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,16384,256,0.0038924444880750445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,1024,0.0038968887594011095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,512,6144,0.07173066669040255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,51200,6144,0.32362400160895455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,51200,0.13174311319986978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,16384,51200,0.2493520047929552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,5120,65536,2.1590934329562717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,4096,0.04469866553942362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,256,0.005592000153329637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,10240,768,0.03148711058828566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,5120,0.013415999710559845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,1536,512,0.019892444213231403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,768,3584,0.1426746712790595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,128,3584,0.00738666703303655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,2560,0.012087999946541257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,3072,3072,0.008360889222886827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,6144,12288,0.03149155444569058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,10240,0.17332000202602812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,3072,0.005290666802061929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,512,256,0.005947555518812603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2048,4096,0.01889155473974016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3584,5120,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,2560,0.05219022101826138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,6144,32,0.012717333104875354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,128,0.014045332868893942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,256,12288,0.006617777877383762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,5120,10240,0.0184906671444575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,12288,0.3513208760155572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,16384,128,0.015827554795477126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,5120,0.06741778055826823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,2048,0.05990578068627251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,128,256,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,256,6144,0.03134399983617995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,2048,0.013267555170589022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,512,0.002499555548032125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,6144,3072,0.009211555951171452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,6144,256,0.00527022240890397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,256,0.01749955614407857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,51200,0.08086488644282024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1024,51200,0.19101244873470732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,1024,256,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1024,5120,0.0648017790582445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,6144,0.02810488806830512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3072,6144,0.009462222456932068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,65536,0.07891644371880426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,5120,10240,0.020301333732075162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,512,0.005934222290913264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,2048,3072,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,256,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,2048,51200,0.06354577673806085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,4096,0.0148053334818946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,512,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,4096,51200,0.08395200305514866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,768,0.0195768889453676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,4096,0.06592888964547051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,7168,0.009004444711738164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,5120,6144,0.011981333295504252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,7168,0.4201253255208333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,7168,0.047298666503694325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,3072,0.009938666390048133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,128,512,0.011363555159833698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,32,6144,0.009735999835862054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,4096,3072,0.02983199887805515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,3584,2560,0.03703555464744568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,768,7168,0.02725244396262699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,5120,0.006245333287451003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,6144,10240,0.19825866487291124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,51200,7168,0.10727555884255303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,2048,0.06694310903549194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,16384,2560,0.032905777295430504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,768,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,65536,12288,0.18971289528740776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,128,0.005229333208666907
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,64,0.026690666874249775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,768,12288,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,2048,0.013890667094124688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,1024,0.0063680000603199005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,2560,10240,0.033324443631702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,4096,0.02055022286044227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,16384,1536,0.05327199896176656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,12288,0.06000355879465739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,3584,16384,0.021357332666714985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,4096,0.07128089004092746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,6144,0.03720000055101182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,768,0.021594666772418555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,16384,0.02898933490117391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,512,0.013429332938459186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,12288,51200,0.11334578196207683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,2048,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,4096,0.021028444170951843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,6144,0.009306666751702627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,512,0.015840889679061044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,256,0.0021831111775504216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,12288,0.19215910964541963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,32,0.011830222275522021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,768,0.019002666076024372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,12288,0.13653688960605198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,5120,7168,0.016169778174824186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,1024,0.016763556334707472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,512,5120,0.017451556192504037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,128,512,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,10240,3072,0.054678221543629967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,51200,0.050592889388402305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,8192,5120,0.011853333148691388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,2048,0.004507555729813046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,3584,10240,0.0236151119073232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,16384,10240,0.03415822320514255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,2048,4096,0.010258666343159145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,1024,0.003169777699642711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,2048,512,0.017177777157889474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1024,65536,0.18693689505259195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2560,4096,0.021018667353524104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,5120,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1024,8192,0.01069777790043089
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2560,3584,0.01927288870016734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,4096,0.020301333732075162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,2048,64,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,12288,0.02509511179394192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,512,7168,0.01941244469748603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,768,4096,0.018607111440764535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,512,0.019160888261265226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,16384,4096,0.20230400562286377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,3072,0.02801333202256097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,128,0.014813333749771118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,256,10240,0.005597333527273602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,65536,10240,0.08676799800660874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,2560,3584,0.008007999923494127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,32,256,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,6144,0.02198311189810435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,64,0.002479111154874166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,10240,51200,0.11321155230204265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,8192,0.07549333572387695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,5120,0.04835022158092923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,1024,1536,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,10240,512,0.027061333258946735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,51200,768,0.016484444340070088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,2560,0.034761776526769005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,4096,0.03410755594571432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,3584,0.03828800055715773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,10240,10240,0.1422613329357571
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,8192,0.011719110939237805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2048,4096,0.04087111022737291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,64,0.006992888947327931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,3584,0.1252408954832289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,3072,12288,0.11961511770884196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,4096,0.6014809078640407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,7168,0.019274666905403137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,512,1536,0.0038817777401871155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,4096,6144,0.02368444369898902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,51200,2560,0.023195554812749226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,2560,0.16890399985843232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,128,0.0021537778278191886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,512,0.015288889408111572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1536,768,0.010424888796276517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,12288,5120,0.025044444534513686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,7168,0.03152799937460158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,10240,0.018928888771269057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,768,0.020228443874253165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,7168,16384,0.10531644688712226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,51200,0.3695671028561062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,3584,2048,0.024500444531440735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,5120,64,0.00832444429397583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,7168,3072,0.008820444345474243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,7168,0.07166577710045709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,768,0.025991110338105097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,2048,3584,0.009092444347010719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,128,0.009973333113723332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,128,256,0.012444444828563266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,32,8192,0.029373334513770208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,512,1024,0.00847999999920527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,1536,0.04316622349951002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,128,5120,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2560,3072,0.04821066723929512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,1024,12288,0.018592000007629395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,3072,0.007143999967310164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,65536,512,0.13270755608876547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,51200,1024,0.03312622176276313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,2560,2048,0.017900443739361234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,7168,0.010038221875826517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,51200,0.04462755719820658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,3584,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,512,0.0052595556610160405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,10240,0.4304204516940647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,512,12288,0.020319999919997323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,1536,4096,0.020759110649426777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,1024,0.00812177773978975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,2048,12288,0.048397332429885864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,2560,16384,0.021584888299306233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,768,0.005562666803598404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,128,7168,0.08183822366926405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,10240,0.01825066738658481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,6144,3584,0.023353776997990076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,51200,0.12111466460757786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,128,8192,0.04100088940726386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,512,0.00998577806684706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,2048,256,0.012050666742854647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,7168,7168,0.08138133419884576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,7168,16384,0.07717155747943454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,10240,0.07894311348597209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,2048,0.03257333238919576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,10240,2048,0.012772444221708508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,51200,0.2497048907809787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,5120,12288,0.03381422162055969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,5120,0.03319022059440613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,128,4096,0.009400000174840292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,65536,3072,0.21892444292704263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3584,8192,0.11819644769032796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,512,768,0.009698666632175446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,3584,0.009644444617960189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,32,0.05897688865661621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,2048,0.009685333404276106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,16384,2560,0.012667555775907306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2560,3072,0.025664889150195654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,51200,0.04237599836455452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2560,65536,0.0649813347392612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,7168,1024,0.011167111496130625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,512,0.008623999853928884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,256,0.002806222273243798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,12288,0.05090044604407417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,64,65536,0.03250666790538364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,5120,0.004606222112973531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,768,0.017151999804708693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,128,7168,0.013397333522637686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,12288,0.040716442796919085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,12288,2.0279208289252386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,10240,0.021454221672481958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,3072,0.014110222458839417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,1536,0.029002666473388672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,12288,0.03229066729545593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,1024,0.026591110560629103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,32,3584,0.004562666846646203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,16384,8192,0.10257511006461249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,7168,2560,0.04695555567741394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,7168,0.016735111673672993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,128,0.007625777688291337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,2048,0.007095110913117726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,6144,128,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,768,0.037477334340413414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,32,0.018716444571812946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,16384,0.15284621715545654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3584,2048,0.015843555331230164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,768,256,0.004179555508825514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,3072,10240,0.015826667348543804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,256,512,0.005663110978073544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,128,2560,0.004916444420814514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,51200,32,0.014074666632546319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,51200,65536,1.9244906107584636
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,4096,8192,0.06762844324111938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,2560,0.01910044418440925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,64,51200,0.049979554282294385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,10240,1536,0.008195555872387355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,16384,51200,2.8623600006103516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,7168,3072,0.07431644201278687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,51200,0.15230845080481634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,8192,7168,0.09211466709772746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,3584,51200,0.05326399869389004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,51200,0.24749954541524252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,10240,8192,0.2846497694651286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,32,0.004957333207130432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,2560,128,0.014477334088749357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,32,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,1536,0.0035253332720862497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,3584,0.043057776159710355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,65536,2560,0.030119998587502375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,1536,0.0069004446268081665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2048,2048,0.018758222460746765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,5120,0.012631999949614206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3584,51200,0.16502843962775335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,2048,0.005479111025730769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,1536,1024,0.017882666654056974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,5120,8192,0.028569777806599934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,7168,0.46397866143120664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,512,5120,0.02109866672092014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,512,65536,0.018452444010310702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,2048,0.011783999701340994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,3072,0.020290666156344943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,16384,3584,0.014337778091430664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,64,2048,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,8192,8192,0.08191288842095269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,3584,32,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2048,1536,0.024081778195169237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,12288,3072,0.031883554326163396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,2560,64,0.00554044461912579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,7168,3584,0.02369333306948344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,128,8192,0.019742222295867074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,16384,3584,0.01645599967903561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,8192,4096,0.0641804469956292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,7168,0.046448889705869884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,8192,0.02476177778508928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,5120,0.04929955469237434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,65536,0.06098488966623942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,51200,3072,0.37870489226447207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,10240,4096,0.014984000060293408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,1024,16384,0.18590933746761748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,65536,4096,0.07982222239176433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,2560,3072,0.0155395550860299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,64,7168,0.0052844443255000645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,10240,0.013343111508422427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,512,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,65536,0.9649733437432183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,5120,0.0383857786655426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,51200,0.06582844257354736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,4096,7168,0.013050666285885705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,1024,65536,0.20668978161282012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,768,0.02938577863905165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,65536,0.08089422517352633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,256,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,3584,0.5497360229492188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,128,0.025951999757024977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,128,0.0052533331844541765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,256,6144,0.018216000662909616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,256,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,51200,0.11275733841790093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,5120,0.21901067097981772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,512,3072,0.007861333588759104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,65536,0.01824977828396691
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3072,51200,0.9154098298814562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,7168,0.010022222167915767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,768,7168,0.011409777734014722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,4096,1536,0.035459554857677884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,51200,0.05066844489839342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,16384,4096,0.028347555134031508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,64,0.005239999956554837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,10240,65536,0.1389520035849677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,3584,2048,0.004939555707905027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,1024,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,128,0.009017777939637503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,128,0.009901333186361525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,12288,51200,1.170138676961263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,4096,8192,0.013731555806265937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,12288,4096,0.29293155670166016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,8192,12288,0.03219199842876858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,3584,8192,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,256,0.023697778582572937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,1024,768,0.0055440002017551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,65536,32,0.016953777935769822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,64,0.0038497779104444715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,64,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,128,6144,0.005973333285914526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,16384,0.14728089173634848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,51200,0.04356177647908529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,7168,8192,0.014496889379289416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,12288,65536,0.36836711565653485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,4096,0.027432888746261597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,768,0.07771199941635132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,7168,2560,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,10240,1024,0.0070266665683852295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,768,0.008961777720186446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,12288,51200,3.859108395046658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,3584,0.01074044406414032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,512,0.02809422214825948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,4096,0.09284800291061401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,128,0.011423110961914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,6144,3584,0.1854133341047499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,7168,0.024463999602529738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,7168,16384,0.02920888861020406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,768,10240,0.027023111780484516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2560,16384,0.02204088866710663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,32,2048,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,32,32,0.0052684446175893145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3072,16384,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,5120,0.009034666750166152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,7168,16384,0.20321155918969047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,4096,0.018869333797030978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,512,65536,0.08158933454089694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,51200,3072,0.5449271202087402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,8192,0.01331022216214074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,8192,0.110935111840566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,2560,0.03870044483078851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,8192,128,0.002500444443689452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,12288,0.06441422303517659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,6144,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,6144,0.040708445840411715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,51200,4096,0.07310844129986234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,12288,0.041032890478769936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,6144,16384,0.19571377171410453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,2560,51200,0.05163822240299649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,65536,5120,0.04682577649752299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,512,0.09659110837512547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,8192,0.07089333401785956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,51200,3072,0.1880871057510376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,128,0.0021751111166344746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,3072,0.07470577955245972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,2560,0.005221333354711533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,6144,0.2648693455590142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,10240,3072,0.01312800000111262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,51200,0.07104977634218004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,7168,768,0.03946311275164286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2560,2048,0.006069333189063602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,512,6144,0.005953777581453323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,5120,0.7200213538275825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,768,0.00350577798154619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,768,0.0184515549076928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,64,0.007305777735180325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,5120,256,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,1536,0.012072888513406118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,8192,768,0.15695199701521131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,7168,5120,0.1026631130112542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,65536,0.09826488627327813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,1024,0.022533333963818018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,3584,0.014564444621404013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,256,32,0.0018320000833935207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2048,128,0.0052720002002186244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,256,3072,0.04670666654904684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,3072,256,0.00831288927131229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,12288,1.3446666929456923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,6144,16384,0.024691555235120986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,3584,0.012700444294346703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,51200,4096,0.0326186650329166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,7168,0.03673422336578369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,12288,2560,0.6409351030985514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,32,3584,0.007607111500369177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3584,6144,0.6813733312818738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,512,7168,0.021538666552967493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,128,65536,0.15082400374942356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,4096,0.02730400032467312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,12288,0.05144355694452921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,6144,65536,0.17765511406792533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,16384,0.043471111191643604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,5120,0.04047911034690009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,256,2560,0.012267555627557965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,3584,0.01094044413831499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,8192,6144,0.1269057724210951
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,7168,0.43856000900268555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,12288,0.018235555953449674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,3584,0.008728889127572378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,3584,6144,0.052503999736573964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,10240,12288,0.2853075663248698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,12288,0.013806222213639153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2560,1024,0.00591822216908137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,10240,2560,0.12473689185248481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,512,51200,0.08565421899159749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,3072,0.011687999798191918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,2048,0.029057777590221826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,1536,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,1024,0.021136000752449036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,5120,768,0.018195556269751657
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,768,4096,0.020294222566816542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,7168,0.03903199897872077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,128,10240,0.012368000215954252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,6144,7168,0.015098666151364645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,10240,1536,0.03422755665249295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,512,0.026118222210142348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,2560,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,6144,1536,0.04618666569391886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,16384,2.9892311096191406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,8192,32,0.0038382220599386427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,768,0.017714665995703805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3072,65536,0.9482542673746744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1536,3584,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1536,10240,0.0509057773484124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,65536,0.18290311760372588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,2560,256,0.01723022262255351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,4096,0.08362044228447808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,3072,65536,0.09164266453848945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,12288,0.13708889484405518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,2560,0.05745244688457913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,4096,4096,0.009047110875447592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,3584,3072,0.04124888777732849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,8192,16384,0.10072533289591472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,10240,256,0.006248000181383557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,64,10240,0.005564444594913059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,6144,0.01908177799648709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,768,51200,0.0461733341217041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,3584,0.027172444595230952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,1536,0.011019555230935415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,4096,3072,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1024,12288,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,2560,0.10983822080824111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,768,0.1125226683086819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,256,1536,0.01219377749496036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,16384,10240,0.05522933271196154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,2048,3584,0.016537777251667447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,128,8192,0.005777777896987067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,8192,3584,0.07373778025309245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,16384,12288,0.07376711236106025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,51200,16384,0.11030844185087417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,10240,1024,0.02277333372169071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,64,0.09906222422917683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,128,5120,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,768,0.024338665935728285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,256,0.029149333635965984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,2048,0.009454222189055549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,1024,0.008643555144468943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,3072,0.006952889263629913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,65536,1024,0.02808711263868544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,5120,6144,0.018223999275101554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,65536,2048,0.14189067151811388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,2048,0.02104266650146908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,16384,0.024655999408827886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,6144,64,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,1024,0.009660444325870937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,51200,1536,0.029140445921156142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,7168,0.08079555299546984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,6144,0.024227556255128648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,2560,12288,0.0612000028292338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,5120,0.13638222217559814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,768,768,0.0058986664646201665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,10240,6144,1.4670435587565105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,3072,3584,0.032419555717044406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,3584,65536,9.287256876627604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,128,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,768,1024,0.004357333398527569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,5120,6144,0.014162666267818876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,2560,3584,0.022337777747048274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,51200,7168,0.11106133460998535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,7168,0.0206657780541314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2048,1024,0.006279110908508301
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,7168,128,0.027756444282001917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,8192,7168,0.055543998877207436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,128,32,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,128,1024,0.0062773335311147906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,1024,128,0.01588444411754608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,2560,3072,0.00756177802880605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,10240,0.02777066661251916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,16384,5120,0.01771022213829888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,2048,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,2048,0.023499555057949487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,1024,0.006597333484225803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,65536,2.9225457509358725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,12288,0.06782933076222737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,1024,0.2175911135143704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,12288,0.02716800073782603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,256,256,0.008185777399275038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,65536,0.5745831065707737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,3584,0.06275377670923869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,10240,7168,0.028432889117134943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,32,128,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,10240,12288,0.49870755937364364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,256,3072,0.007674666742483775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,12288,0.025052444802390203
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,4096,128,0.008388444781303406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1536,3584,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,3584,0.01464000013139513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,128,128,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,6144,0.01370133293999566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,2560,0.5455422401428223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3584,4096,0.013444444371594323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,10240,0.028753777345021565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,16384,0.2322311136457655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3584,8192,0.01716444392999013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,3072,0.14960266484154597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,3072,0.02586311101913452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,65536,0.1821582184897529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,64,0.19075022803412545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,16384,6144,0.025618667403856914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2048,6144,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,1024,2560,0.014427555931939019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,5120,6144,0.013507555756303998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,5120,0.008427555362383524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,3584,32,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3072,16384,0.06664178106519911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,2560,1024,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,16384,1536,0.07306044631534152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3072,51200,0.06548444430033366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,256,768,0.013432000246312885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,8192,0.019908444748984445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2560,2560,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,51200,3584,0.02909066610866123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,8192,0.012319111161761813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,2048,51200,0.05573599868350559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,768,2560,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,128,10240,0.011725333001878528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,1024,0.01332622187005149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,512,0.005938666562239329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,51200,0.00906666699382994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,256,0.005568000177542369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,128,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,65536,2560,0.178384886847602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,4096,0.017845332622528076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,64,6144,0.009415999882751042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,8192,0.2512133386400011
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,768,512,0.006217777729034424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,256,7168,0.02400355537732442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,51200,0.22919821739196777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,5120,2560,0.03416711091995239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,128,256,0.006230222268237009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,5120,256,0.008009778128729926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,3072,51200,0.05718400080998739
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,256,5120,0.019684443871180218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,65536,0.4276968903011746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,5120,0.04753244585461087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,128,0.006656888872385025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,51200,12288,0.21322222550710043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,65536,32,0.12003466818067764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,3072,0.018283555905024212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,256,16384,0.042055110136667885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,2048,0.019683554768562317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,128,64,0.0021333334346612296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,2048,32,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,768,2560,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,8192,64,0.014696000350846184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,64,768,0.0028159999185138275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,5120,3072,0.013682666752073499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,256,6144,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,7168,0.07451555464002821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,7168,12288,0.018769777483410306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,1024,16384,0.01660977800687154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,7168,0.08201777935028076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,1536,0.024336000283559162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2560,16384,0.0207324441936281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,16384,6144,0.04981244603792826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,128,0.0038728887836138406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,64,512,0.005941333456171884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,65536,0.07493511173460218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,51200,0.259481774436103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,64,0.010036444498433007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,64,128,0.007619555625650618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,10240,0.08681866857740615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,51200,256,0.13053422504001194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,1536,0.015483554866578845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,3072,0.007872888611422645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1536,5120,0.02160533269246419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,512,0.002444444431198968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,5120,10240,0.031031999323103163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,7168,768,0.016612443659040663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,256,2560,0.007269333634111617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,256,0.0151573336786694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,64,3072,0.00741155528359943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,8192,32,0.008329778081840938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,2560,4096,0.019193778435389202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,5120,0.009653333160612319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,5120,0.029951999584833782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,1024,2048,0.00963466697269016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,4096,12288,0.023016000787417095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,3584,512,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,51200,0.19209599494934082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,4096,0.00795999997191959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,8192,0.011520889070298938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,6144,1024,0.0041848888827694785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,64,3072,0.0052328887912962176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,16384,7168,0.14526400301191542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3584,10240,0.20703289243910047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,16384,0.06264444192250569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,1024,0.00794311116139094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,1024,512,0.013390222357379066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,5120,0.06444178024927776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,1024,6144,0.0192586663696501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,64,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,1024,5120,0.009444444543785518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2048,2560,0.02956977817747328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,16384,6144,0.04936800069279141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,5120,128,0.020412445068359375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,128,0.01350400017367469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,512,5120,0.020618667205174763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,512,65536,0.059912886884477407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,3584,0.011526222030321756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2560,1536,0.013751111096805997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,16384,0.017121778594122995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,4096,65536,0.07590222358703613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,32,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,6144,3584,0.009703111317422655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,51200,64,0.009473777479595607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,32,0.0018417777286635505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,5120,4096,0.07668355438444349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,5120,0.04328088959058126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,1536,2048,0.011690666278203329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,4096,7168,0.015283554792404175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,10240,16384,0.6352933247884115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,65536,0.11149244838290745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,12288,768,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,5120,1536,0.02169599963559045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,2560,0.004527111020353106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,12288,0.10961421993043687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,51200,51200,0.36127911673651797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2560,32,0.0069875551594628235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,256,1024,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,1536,0.018597332967652213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,2560,0.08399644162919785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,10240,0.035559111171298556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,12288,128,0.021203555994563635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,65536,0.05696800020005968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,8192,4096,0.013176889055305056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,2048,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,5120,512,0.015188443991872998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,7168,51200,1.7956852383083766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,65536,0.9930053287082248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,3584,4096,0.20692800150977242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,16384,512,0.005548444473081165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,32,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,3584,0.11309777365790473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,2560,0.039135111702813044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,12288,4096,0.7101982434590658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,51200,10240,0.06878133614857991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2048,2560,0.01867911054028405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,3072,128,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,51200,7168,0.16308089097340903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,5120,0.015828443898095023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,16384,8192,0.0247706671555837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,12288,10240,0.10767555236816406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,51200,10240,0.1759964492585924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,65536,0.06327288680606417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,3072,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,256,3584,0.007631999750932057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,256,6144,0.00960355583164427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,5120,7168,0.27378755145602757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,7168,2560,0.033111111985312566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,2048,768,0.017854221993022494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3584,12288,0.13836356004079184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,12288,2048,0.015809777710172866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,16384,0.23812622494167754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,1024,12288,0.032524443335003324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,8192,0.15267554918924967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,5120,16384,0.3114604420132107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,51200,5120,0.5796604686313206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,256,0.030505776405334473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,3072,51200,0.07458577553431193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,65536,12288,1.7731457816229925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,2560,0.009463110731707679
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,16384,0.030321776866912842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,65536,64,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,5120,3584,0.02551733288500044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3584,512,0.007683555285135905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,12288,16384,0.044792887237336904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,64,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,10240,768,0.009383999639087254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,1536,3584,0.0660346680217319
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,3072,3072,0.045788443750805326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,51200,768,0.015970667203267414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,10240,0.07671644290288289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,8192,0.03822133276197646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,1536,0.031459556685553655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,12288,768,0.006846222198671765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,10240,6144,0.07051999701393975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,128,6144,0.009647111097971598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,512,12288,0.04026044408480326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,1536,0.00999022192425198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,3072,0.09340888924068874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,5120,0.008225777910815345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,256,3072,0.01475288967291514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,8192,4096,0.48556264241536456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,32,7168,0.012412444584899478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,1536,0.013082666529549493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,3584,0.026445334156354267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,12288,2560,0.037131554550594754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,6144,0.01127911110719045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,1024,0.0162435554795795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,3072,0.007729777859316931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,128,10240,0.04208533300293816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,3072,512,0.014102222190962898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,32,0.00180177783800496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,7168,128,0.008980444735950893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,5120,256,0.01996888882584042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,16384,0.08296089039908515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,512,6144,0.013723555538389417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,10240,64,0.011379555695586734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2560,128,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,32,1536,0.003549333247873518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,10240,65536,0.21625422106848824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,51200,0.0328453348742591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,128,65536,0.052674667702780835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,128,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,6144,256,0.01891377733813392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,6144,128,0.005245333330498801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,65536,7168,0.5752204259236654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,6144,0.8841235372755262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,512,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,65536,1.182168854607476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,5120,65536,0.22625690036349824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,1024,0.11169511742062038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,6144,0.015219555960761176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,6144,3072,0.07753333118226793
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,51200,0.05567289061016507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,3072,0.020562666985723708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,65536,128,0.022057778305477563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2560,2048,0.018523555662896898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,3584,12288,0.04513422316975064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,51200,1.219365331861708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,1024,0.04855022165510389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,64,256,0.002143111079931259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,4096,6144,0.010317333042621613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2048,1536,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,10240,51200,1.23580445183648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,65536,0.07349422242906359
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,65536,1.453121821085612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,256,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,16384,0.12063466178046332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,2048,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,256,128,0.01683555543422699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,7168,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,32,2560,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,6144,3072,0.0315688881609175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,32,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,10240,4096,0.01477688882086012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,64,0.001806222316291597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,512,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,256,8192,0.022008889251285132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,16384,128,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,768,128,0.008275555239783393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,1024,0.2049084504445394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,16384,256,0.047028445535235934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,7168,16384,0.23104710049099395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,128,0.018254222141371835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,128,5120,0.009181333084901174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,12288,0.12410667207505967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,2048,0.006501333167155583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,2048,65536,3.2909147474500866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,7168,32,0.0035377778112888336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,128,0.011393778026103973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,8192,3072,0.07105777661005656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,3072,0.007501333124107785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,2048,5120,0.031217777066760596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,16384,0.019487111104859244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2048,32,0.0022604444788561925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,7168,768,0.006395555618736479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,65536,1536,0.0188044442070855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,768,0.03987377882003784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,256,0.03303555647532145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,4096,0.006940444310506185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,1536,0.00488800017370118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,8192,1536,0.0069475554757648045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,1536,4096,0.008012444608741337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,3584,6144,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,1536,2560,0.005214222189452913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,8192,0.012842666771676807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,3584,0.9079280429416232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,2560,5120,0.06343377961052789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,128,6144,0.009775111244784461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,768,0.005936000082227919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,51200,3072,0.02676800058947669
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,128,0.0035297779573334586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,10240,0.03510933452182346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,12288,12288,0.06645777490403917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,8192,10240,0.04358488983578152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,12288,7168,0.030857778257793848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,768,0.19318755467732748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,6144,16384,0.17497421635521782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,256,4096,0.007359111474619971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,8192,1024,0.006595555692911148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,1024,0.021555556191338435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,2048,10240,0.01236177815331353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,2048,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,51200,51200,1.3396586312188044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,512,0.007287999822033777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,51200,2048,0.029399110211266413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,6144,7168,0.017758222089873422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,7168,0.02869510981771681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,12288,0.0807466639412774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,768,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,32,5120,0.00563822231358952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,6144,0.009318222602208456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,512,51200,0.019675556156370375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,6144,0.034550223085615374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,256,128,0.002171555534005165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,2560,4096,0.00793333351612091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,65536,0.12658577495151097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,768,0.005620444400442972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,32,0.0021822222818930945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,10240,512,0.01957244508796268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3072,10240,0.0340666671593984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,8192,768,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,128,4096,0.01717688971095615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,8192,64,0.018750222192870248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,5120,51200,0.9113083945380317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,16384,0.044343110587861806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,7168,3072,0.048806223604414195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,128,5120,0.02508977717823452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,768,0.008350222474998897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,2048,0.022454222043355305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,4096,1536,0.004536889079544279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,1024,8192,0.04959466722276476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,12288,0.03423466616206699
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,7168,0.2065955532921685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,10240,7168,0.043440000878440015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,8192,256,0.016550223032633465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,64,1536,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,3584,768,0.018197332819302876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,4096,0.007612444460391998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,6144,12288,0.03023911184734768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,32,16384,0.015829333000712924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,16384,2048,0.035755554835001625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,8192,0.03516177667511834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,768,0.004522666749027041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,256,0.0064693333374129395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,768,3584,0.018901333212852478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,51200,0.06942222515741985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,32,0.008193777667151557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,768,0.026833777626355488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3584,65536,0.06992089086108737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,1024,51200,0.10595022307501899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,65536,0.06587111287646823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,4096,12288,0.11544888549380833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,512,256,0.026196445027987163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,64,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,1024,65536,0.027090667022599116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,8192,128,0.015116444892353482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,10240,0.09651288721296523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,5120,0.05055200060208639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,256,1536,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,2048,0.010214221974213919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,51200,256,0.009392889009581672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,16384,0.10905066463682388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,10240,0.1611537800894843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3584,3072,0.03544622328546312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,6144,64,0.016540444559521146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,2560,0.018086221483018663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,65536,0.42595733536614316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,7168,0.03358577688535055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3584,3072,0.016168889072206285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,32,7168,0.010004444254769219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,65536,32,0.09043466382556492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,4096,4096,0.060102224349975586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,1536,0.11126044061448838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1024,6144,0.01865688959757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,768,3584,0.01313866674900055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,6144,0.026802667313151892
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,4096,5120,0.09750755627950032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,6144,1024,0.02166755497455597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,768,5120,0.012087110843923358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,7168,51200,0.11127732859717475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,768,0.030117332935333252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,2048,768,0.0033244445092148254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,4096,16384,0.17980266941918266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,64,0.0018195555441909367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,51200,256,0.14842132727305093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,256,0.15523910522460938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,6144,4096,0.07288000318739149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,8192,1536,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,128,1536,0.012090666426552666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,4096,4096,0.008071111308203803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,128,5120,0.009149333669079674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,1536,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,128,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,1024,4096,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,10240,65536,0.19419466124640572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,2560,0.010103111465771994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,5120,6144,0.0123857781291008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,5120,6144,0.025702221526039973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,12288,2048,0.02568711174858941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,16384,3584,0.14122222529517278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,16384,0.12291288375854492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,256,0.002549333290921317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,2560,0.3717466725243463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,2560,65536,0.19842933283911812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,6144,512,0.03201599915822347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,128,0.025235555238193933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,16384,16384,0.07249155309465197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,768,0.020802666743596394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,4096,512,0.01478488908873664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,6144,0.024636444118287828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,256,0.006771555377377405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,1024,3072,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,2048,0.03704977697796292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,1024,0.006590222318967183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,64,6144,0.010044444766309526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,3072,0.02572711143228743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,2048,1024,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,10240,12288,0.10091200139787461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,8192,65536,0.1020488871468438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,10240,0.08493600289026897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,256,512,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,64,64,0.0020728889438841078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,768,2560,0.007255111303594377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,6144,768,0.006251555350091722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,2048,0.05255555444293552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,51200,1.1961796018812392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,768,3584,0.059671110577053495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1536,3584,0.21502933237287733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,3584,1024,0.013464888764752282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,512,65536,0.028154667880800035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,12288,16384,0.07220444414350721
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,12288,0.02791555391417609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,2048,768,0.003120888852410846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3584,6144,0.048057778014077075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,32,3072,0.004184000194072723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,768,0.024100444383091394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,65536,0.06524978081385295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3584,3584,0.023805333508385554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,512,0.0069511110583941145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,3584,0.007616000043021307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1536,1024,0.018153778380817838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,3072,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,6144,65536,0.08467200067308213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,51200,51200,6.004166920979817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,1536,0.003879111260175705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,8192,65536,0.09708622429105972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,3584,0.07226488987604777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,6144,0.7637733353508843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,10240,8192,0.02217955556180742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,256,65536,0.05156266689300537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,7168,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,12288,256,0.006280888699822956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,768,8192,0.1473919947942098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1024,5120,0.01720266706413693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,256,1536,0.0035333335399627686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,16384,0.9986631605360242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,5120,0.15174578295813665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,256,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,768,51200,0.05156711075041029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,2048,2048,0.025207999679777358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,4096,768,0.00350222239891688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,128,0.0226248883538776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,16384,3072,0.0173662222094006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,7168,1024,0.009117333425415887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,4096,256,0.005380444642570283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,5120,6144,0.12422933843400742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,2560,2560,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,5120,0.1462106704711914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,12288,3072,0.015152000718646579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,6144,8192,0.15120089054107666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,512,8192,0.02541333271397485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,7168,0.027998222245110407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,1536,0.006416888700591193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,10240,0.043086220820744835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3584,51200,0.07099911239412096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,256,10240,0.08157333400514391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,32,0.007639110916190677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1024,3584,0.01865688959757487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,51200,0.058134224679734975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,256,768,0.005976888868543837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,51200,0.029343999094433252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,3584,768,0.005879999862776862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,6144,16384,0.023807111713621352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,8192,1024,0.00490133340160052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,3072,0.10067555639478896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,64,0.005618666609128316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,128,0.01588444411754608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,5120,0.015321777926550971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3072,4096,0.04187555445565117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,16384,12288,0.08444711234834458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,64,0.0028595555987623003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,256,64,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,5120,0.11099111371570164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,7168,6144,0.1085653305053711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,3584,0.022163555026054382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,1536,0.016374222106403775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,3584,0.029323554701275293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2048,16384,0.019948444432682462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,768,8192,0.020633776982625324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,8192,0.04183822207980686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,12288,0.041908442974090576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,768,5120,0.021570665968788996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,1024,0.003478222423129611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,2560,8192,0.03224088748296102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,1024,0.006589333216349284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,16384,0.1347928841908773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,16384,3584,0.17603022522396514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,256,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,6144,0.018527110417683918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,10240,8192,0.2369333373175727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,7168,128,0.0021075554605987337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,768,65536,0.0915679997868008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,3584,64,0.005538666827811136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,5120,16384,1.2415599822998047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,12288,16384,0.5802488856845432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,2560,0.010824888944625854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,128,0.012084444363911947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,51200,0.22556887732611763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,1024,2560,0.0069617778062820435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,7168,0.013177778157922955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,3072,12288,0.016795555750528973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1024,3584,0.01611911091539595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,65536,6144,1.7142382727728949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,2048,512,0.005219555563396878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,32,6144,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,768,65536,0.01997066703107622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,51200,10240,1.0128658082750108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,16384,1536,0.15708977646297878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,8192,0.06085777944988675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,2560,0.01611555616060893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,3072,0.009341333475377824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,128,2560,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,2560,0.015591111448076038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,512,65536,0.0665102203687032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,7168,1536,0.03745244608985053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,768,0.004460444466935264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,6144,0.031390223238203264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,6144,2048,0.007976888782448238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,8192,512,0.00656088896923595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,2560,256,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,64,51200,0.0058853332367208265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,16384,0.2506595452626546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,1024,0.026138666603300307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,16384,0.01632355484697554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,10240,1.5886027018229167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,4096,10240,0.06285422378116183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,64,0.00758755538198683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,2048,2048,0.021937777598698933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,128,7168,0.032872888776991106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,3584,0.029209779368506536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3072,1024,0.010713777608341642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,6144,51200,0.10525600115458171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,5120,16384,0.0425164434644911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,12288,0.03524444500605265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,3584,0.07279555665122138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,64,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,51200,0.06327199935913086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,3584,65536,0.610124429066976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,16384,4096,0.1115680005815294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,8192,0.0052675555149714155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,1536,128,0.004197333421972063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,512,0.0366906656159295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,256,0.0027591110103660156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,5120,0.01829777823554145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,6144,0.012244444754388599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,16384,256,0.017504889104101393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,65536,0.1351848840713501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,51200,0.01921777758333418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,7168,32,0.006272888845867581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,3072,0.025780444343884785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3584,256,0.0024968888610601425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,51200,0.062104887432522245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,512,0.005416000054942236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,2560,8192,0.014450665977266101
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,64,512,0.00545866663257281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,4096,128,0.0021724444296624926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,6144,2048,0.008633777499198914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,7168,51200,1.0415973663330078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,256,0.01033155537313885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,2560,4096,0.01036711119943195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,4096,10240,0.09629155529869927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,128,0.006640000061856375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,768,7168,0.008390222158696916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,1536,0.021420444051424663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,12288,7168,0.07517777548895942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,512,0.014159111513031853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,256,8192,0.025619554850790236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,32,0.007635555333561367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,65536,32,0.017208000024159748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,3072,3584,0.009198221895429824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,2048,0.05079022381040785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,8192,0.15052533149719238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1536,8192,0.024837333295080397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,2560,0.05845599704318576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,5120,0.024077332682079736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,8192,3584,0.09636800156699286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,2048,0.009291555318567488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,65536,6144,0.3985457685258653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,51200,6144,0.044066667556762695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,2048,10240,0.05616888735029432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,6144,0.01791200041770935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,65536,512,0.014992000328169929
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,64,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,65536,128,0.025621333056026038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,5120,5120,0.012137778103351593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,10240,0.012088889049159156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2560,7168,0.01163200040658315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,7168,51200,0.11083733373218113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3584,5120,0.03678755627738105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,51200,0.11279466417100693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,7168,3584,0.011880888707107969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,10240,0.3279831144544813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,12288,0.018222222725550335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,8192,6144,0.17167645030551484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,3584,0.05360177821583218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,4096,7168,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,3072,0.06105955441792806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,51200,51200,0.40593955251905656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1536,5120,0.017083555459976196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,12288,1024,0.04528977804713779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,2560,0.20770577589670816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,768,10240,0.018276444739765592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,3072,0.1277164485719469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,256,0.006657777975002925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2560,8192,0.026176888081762526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,32,10240,0.049759109814961754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,6144,7168,0.321669340133667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,2560,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,1024,0.004010666575696734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,7168,0.07761155234442817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,16384,0.1374328931172689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,768,0.018212444252438016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,64,16384,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,12288,0.15154400136735705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,1536,16384,0.11387999852498372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,1536,1024,0.003528000166018804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,1024,3072,0.007273777491516537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3072,4096,0.009001777403884465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,12288,0.038134220573637224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,16384,0.09752621915605332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,7168,0.01074222226937612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,12288,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,128,6144,0.009702222214804756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,12288,1536,0.026377777258555096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1024,65536,0.028132445282406274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,256,65536,0.052368889252344765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,8192,2560,0.0127688886390792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,4096,0.011680888632933298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,32,10240,0.05032177766164144
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,6144,0.023543111152119104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1024,128,0.00565422202150027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,768,512,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,256,512,0.002518222149875429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,51200,0.43981244828965926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,2560,0.007041778001520369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,3072,0.00866844422287411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,10240,7168,0.04839288857248095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,16384,0.04043111205101013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,65536,0.03217511044608222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,51200,4096,0.07445066505008273
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,1536,0.006296000132958095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,512,0.009111999637550777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,128,0.018200000127156574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2560,128,0.0021262222694026097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,51200,0.19642578230963814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,8192,512,0.02165422174665663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,4096,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,512,5120,0.009359111388524374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,5120,5120,0.014867555763986377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,3072,1536,0.0069573331210348345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,256,5120,0.007982222570313348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,3072,3584,0.007244444555706448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,2048,0.007262222468852997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,768,0.009027555584907532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,3584,0.03404088815053304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,64,0.019263111882739596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2560,512,0.002463999928699599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,128,1536,0.012038222617573209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,512,0.005983110931184556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,512,65536,0.019276445110638935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,2048,3584,0.007609777980380588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,2048,128,0.00526577772365676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,3072,1536,0.014826666977670459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,10240,10240,0.18307822280459932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,10240,6144,0.04168088899718391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,51200,0.205878218015035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,4096,0.028318223026063707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,6144,1536,0.020020445187886555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,12288,0.29758045408460826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,12288,51200,2.396056916978624
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,8192,0.015790222419632804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,2560,5120,0.018218666315078735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,10240,8192,0.05206400156021118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,3072,0.006956444846259222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,51200,4096,1.011254204644097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1024,1024,0.022844443718592327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,12288,256,0.04290489024586148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,256,768,0.03505955470932855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,1024,0.040375110175874494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,6144,768,0.013816888961527081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,16384,2048,0.011319110790888468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,6144,0.02572622232966953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,512,6144,0.018572444717089336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,512,4096,0.008411555654472774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,16384,0.01992266707950168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,2048,0.016466667254765827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1024,12288,0.067285332414839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,51200,0.139101333088345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,5120,2048,0.006636444479227066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,768,12288,0.009707555174827576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,12288,3072,0.07059999969270495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,4096,3584,0.022058667408095464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,2048,1024,0.014483556151390076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,51200,5120,0.09952000114652847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,12288,0.21400977505577934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,1536,0.02147644427087572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,512,0.02470933397610982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,128,0.03517244590653314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,512,4096,0.044518222411473594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,5120,3584,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3072,1024,0.0035502223504914176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1024,6144,0.01921244462331136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,65536,0.06274844540490045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,3584,0.015087111128701104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,768,0.018000000052981906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,4096,256,0.0029164445069101122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,768,768,0.005576888720194499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,4096,768,0.01930488811598884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,256,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,1536,3072,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,1024,0.014831999937693277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,768,0.01588088936275906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1024,1536,0.006640888750553131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,7168,3072,0.010285333626800114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,768,256,0.0038942222793896994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,16384,0.03029511041111416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,6144,2048,0.0574186642964681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,4096,6144,0.03525066706869338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,64,3072,0.011414222419261932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1024,64,0.0018462222069501877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,7168,512,0.005561777700980504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,10240,51200,0.7195928891499838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,1536,2560,0.007650666766696506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2560,65536,0.05565600262747871
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,256,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,64,51200,0.21897155708736846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,32,3584,0.023405333360036213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,128,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,4096,0.17971021599239775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,3584,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,10240,128,0.010355555348926121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,8192,65536,9.450447930230036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,1536,0.0035386664999855887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,64,256,0.0031822222388452957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,2048,10240,0.01649599936273363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3072,512,0.003536000019974179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,5120,0.013054221868515015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,768,8192,0.015457777513398064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,128,51200,0.04559555649757385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,768,0.06747466988033719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,128,0.02129688858985901
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,1536,0.03637333379851447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2560,128,0.005579555614127054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,7168,0.03149688906139798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,256,1536,0.020639111598332722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,4096,0.005952000204059813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,32,0.0021102221475707158
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,1024,0.0038622220357259116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,12288,768,0.016120000018013846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,2560,64,0.012934221989578672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1536,3072,0.029273778200149536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,4096,0.020096888144810993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,8192,0.011385777758227454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,3584,0.007135999699433644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,7168,1024,0.006274666637182236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,1024,10240,0.2477191024356418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,1536,4096,0.012755555411179861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,8192,10240,0.029514667060640123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,51200,7168,0.1546880006790161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,1536,0.006278222219811545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,65536,3072,0.1004213359620836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,65536,0.08812710973951553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,16384,8192,0.5320364634195963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,10240,2048,0.01995644470055898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,128,4096,0.008339555727110969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,2560,8192,0.011333333121405708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,8192,65536,1.0064115524291992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,64,0.014127999544143677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,256,8192,0.0644399987326728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,65536,32,0.010571555958853828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,8192,0.02273955610063341
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,3072,10240,0.013993778162532382
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,4096,4096,0.06342577934265137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,51200,3584,0.21184889475504556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,8192,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,5120,0.03401599989997016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,3072,4096,0.01882222294807434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,10240,0.02754488918516371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3072,256,0.012258666257063547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,5120,51200,0.09160000085830688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,768,65536,0.6160080167982314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,10240,4096,0.01386044422785441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,256,0.009425777528021071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,65536,0.4157591130998399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,2560,0.0138568886452251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,5120,0.021923555268181696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,10240,16384,0.9411484400431315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,128,0.0041964443193541635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,64,3584,0.0269768883784612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,32,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,256,1024,0.007648888561460707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,4096,0.02607022225856781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,5120,256,0.013445333474212222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,64,0.005622222191757626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,10240,0.035219556755489774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,10240,0.050380445188946195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,1024,51200,0.16181955072614881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,64,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,3072,1536,0.0041306668685542215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,64,6144,0.009533333281675974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,64,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,7168,0.025073778298166063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,768,0.02699022160636054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,10240,0.07646577888064914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3072,1536,0.018208889497650992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,256,0.008976000050703684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,7168,0.13970755206214056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,12288,5120,0.24452532662285698
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,8192,0.016300444801648457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,2560,0.01852799952030182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,2048,32,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1024,1536,0.011758222348160215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,3072,0.06991910934448242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,12288,3072,0.015740444262822468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,128,0.026245334082179602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1536,1024,0.007657777931955125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3584,1024,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,8192,2560,0.009672889278994666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,16384,12288,0.5685457653469509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,2560,0.004219555606444676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,256,256,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,12288,0.26439200507269967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,10240,0.06968622075186835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,256,2048,0.013773333695199756
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,768,0.014402666025691561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,4096,0.05063466562165154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,1536,0.014160000615649752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,65536,3584,0.03496177660094367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,16384,10240,0.24615555339389375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,3072,6144,0.011381333072980246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,10240,0.09211733606126572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,1536,0.01682933337158627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,12288,0.20634222030639648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,2048,0.006963555183675554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,12288,0.00960799968904919
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,8192,7168,0.624742243025038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3072,65536,0.27471910582648384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,3072,0.005541333307822545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,2048,32,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,5120,51200,0.1279182169172499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,65536,0.08105511135525174
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,65536,0.42709512180752224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,2048,64,0.0019697778754764134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,3584,16384,0.027727110518349543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,768,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,6144,32,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,128,3072,0.014133334159851074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,256,0.0028888889484935333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,7168,1024,0.004545777622196409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,12288,1536,0.016936888297398884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,65536,768,0.24529067675272623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2048,1024,0.003907555507289039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,2560,1024,0.016177778442700703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,256,0.055536889367633395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,2560,0.004929777648713854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3584,16384,0.020615110794703167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,3584,0.00795199970404307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,65536,63.09754774305555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,256,3584,0.00833333366447025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,7168,16384,0.02377066678471035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,2048,0.033979554971059166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,128,128,0.003519999898142285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,2048,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,64,64,0.010748444332016839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,12288,0.043643554051717125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,7168,0.7807617717319064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,8192,0.026710222164789837
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,512,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,12288,51200,0.16566844781239828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,2048,3584,0.01758311192194621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2048,768,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,5120,4096,0.009376000199053023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,4096,7168,0.014870221416155497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,6144,0.14311822255452475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,512,0.022276444567574397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,2048,10240,0.014142221874660917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,12288,32,0.003132444495956103
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2560,6144,0.06979555553860135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,7168,0.06508621904585096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,3072,12288,0.07062577539020114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,16384,0.13588178157806396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,256,65536,0.05116711060206095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,16384,8192,0.2437573273976644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,2560,128,0.005225777626037598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,1024,0.014847111370828418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,128,10240,0.012040889097584618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,2048,51200,0.19691288471221924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,7168,0.09015644258923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,1536,64,0.008971555365456475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,256,0.0021866667601797315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1024,256,0.01718133356836107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,2560,0.0083137775460879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,64,32,0.007998222278224098
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,10240,64,0.026625777284304302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,2560,0.022666666242811415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,512,0.031296888987223305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,768,0.24363022380405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3584,3584,0.021913778450753953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,10240,768,0.04387999905480278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,6144,0.02405244443151686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,1536,16384,0.05844799677530924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,3584,4096,0.007996444072988298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,12288,16384,0.14073422220018175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,3584,7168,0.036552889479531184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,3584,0.027781334188249376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,32,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,16384,0.18681866592831084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1536,10240,0.016116445263226826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,2560,0.004979555391603046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,32,51200,0.007654222349325816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,1536,0.0038319999972979226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,32,0.005931555396980709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,3072,7168,0.12488622135586208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2560,2048,0.02815822097990248
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3072,6144,0.010390222072601318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,6144,1536,0.007079111205206976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,128,0.0028071111689011254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,5120,0.0835591091050042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,768,0.034986668162875705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1024,6144,0.018608000543382432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,1024,0.010067555639478896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,7168,2560,0.06785421901279025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,7168,0.01347733371787601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,1024,16384,0.012612444659074148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,12288,768,0.018808888064490426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,5120,128,0.004564444637960858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,16384,0.2606133355034722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3072,8192,0.012774222426944308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,12288,0.02914666798379686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,65536,0.17581333054436576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,51200,51200,0.40016889572143555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,3072,128,0.0059004442559348206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,8192,1024,0.005230222311284807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,32,8192,0.027450667487250432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,2048,0.006579555571079254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,5120,2560,0.06246755520502726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,16384,3584,0.11116888788011338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,7168,0.024263999528355066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,128,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2048,3072,0.031967110104031034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,2560,0.007603555917739868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,51200,0.02202933364444309
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,5120,1536,0.015990222493807476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,3072,65536,0.08312266402774386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,12288,3072,0.03189599845144484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,1024,6144,0.011659555964999728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1536,128,0.01479555500878228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,512,12288,0.027504889501465693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,3072,0.022636445032225713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,16384,0.2572302288479275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,7168,1024,0.03359288970629374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,2048,0.013218666944238873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,64,0.0052560000783867305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,7168,0.010015111002657149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,768,256,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,1536,768,0.005542222410440445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,64,32,0.005286222116814719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,10240,0.03132177723778619
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,1536,0.011485333244005838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,5120,16384,0.043622222211625844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,2048,128,0.010606221854686737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,128,65536,0.04703022374047173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,512,0.02029600077205234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,8192,3584,0.011580444044537015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,10240,0.012091555529170565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,10240,256,0.017488888568348356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,2560,16384,0.09444089068306817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,16384,0.017807111144065857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,256,8192,0.00629333323902554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,512,6144,0.021988444858127173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,10240,0.02277066641383701
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,512,12288,0.02718488872051239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,2048,16384,0.02054933375782437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,1024,1024,0.005939555664857228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,5120,0.01685688893000285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,512,0.011728888584507836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,1536,16384,0.01651733285850949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,51200,0.3609911070929633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,7168,65536,0.07919022109773424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,51200,2560,0.13086399767133924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,65536,0.8753031094868978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,7168,0.6181848843892416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,512,4096,0.02132622235351139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,65536,12288,0.12773688634236655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,2560,5120,0.016492444607946608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,768,0.003900444342030419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,2560,0.04433244466781616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,512,0.16645599736107722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,256,5120,0.012747555143303342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3584,3584,0.011384000380833944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,16384,5120,0.2667920059627957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,128,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,16384,16384,0.09759822156694199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,10240,1024,0.0070257774657673305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,7168,0.014451555079884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,256,0.012453333371215396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,5120,0.05520800087187025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,6144,10240,0.17174666457706025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2560,7168,0.02974933385848999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,4096,768,0.005928888916969299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,1024,768,0.011384000380833944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1024,4096,0.008595555192894405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,3584,0.02644444505373637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,256,0.0069324444565508105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,4096,128,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,51200,1024,0.02203111184967889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,512,6144,0.019872888922691345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,32,0.002256888896226883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,10240,2048,0.02952444553375244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,128,3584,0.10762577586703831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,6144,512,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,3072,3072,0.03572888837920295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,128,0.017080000705189176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,256,3584,0.007304888632562425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,65536,3584,0.03866044349140591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,2560,0.008000888758235509
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,8192,2048,0.04179644584655762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,4096,51200,1.2104693518744576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,6144,10240,0.03605777687496609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,2048,2560,0.006001777946949005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,16384,0.046554668082131274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1024,65536,0.4009386698404948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,32,3584,0.005897777775923411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,2048,3072,0.02803022331661648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,1024,0.005959110955397288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,2560,0.023039110832744177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,4096,16384,0.06442577971352471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,4096,2048,0.006953777538405524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,6144,2560,0.022980444961124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,64,4096,0.004588444613748127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,5120,2560,0.10603555705812241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,7168,0.051001777251561485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,8192,2560,0.008556444611814287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,4096,0.03215555681122674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,256,0.008290666672918532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,12288,0.057683553960588246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,1024,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,512,0.002512000087234709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,128,16384,0.033859554264280535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,32,256,0.0021991110924217436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,7168,0.08561333020528157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,2560,6144,0.18046755260891387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,768,0.018221333622932434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,10240,0.13188444243537054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,16384,0.025422222084469263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,256,0.02348711093266805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,5120,4096,0.04120799899101257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3072,2560,0.02093244426780277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2560,512,0.01367644468943278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,768,768,0.016192000773217943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,256,256,0.0052693333062860705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,7168,2048,0.16618043846554226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,1024,12288,0.0147706667582194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,3072,51200,0.049145778020222984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,7168,0.1240195565753513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,32,512,0.007963555554548899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,12288,256,0.014093332820468478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,6144,512,0.015824889143308003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,256,6144,0.007281777759393056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,5120,65536,1.8897554609510634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,8192,3072,0.1390213304095798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,3584,12288,0.01753777762254079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,51200,0.06117244561513265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,2560,16384,0.10056533416112264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,256,0.016543111867374845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,4096,0.04888000090916952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,1536,4096,0.007703999678293864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,10240,2048,0.022242666946517095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,1024,3584,0.007895111209816402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,2048,0.0041857775714662345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,10240,12288,0.16675110658009848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,12288,0.2746693293253581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,4096,16384,0.021144888467258875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,10240,1024,0.01461955573823717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,2560,0.09137688742743598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,32,6144,0.016531555189026725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,6144,0.029488000604841445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,5120,51200,0.869324419233534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,12288,51200,0.10751199722290039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,7168,256,0.0028355556229750314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,768,6144,0.010040000081062317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2048,128,0.01032266683048672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,7168,0.025928889711697895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,1024,0.004518222063779831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,16384,6144,0.04150222076310052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,6144,1024,0.018239999810854595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,128,32,0.002471111093958219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,8192,0.013553777502642738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,512,4096,0.03273244367705451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,3072,10240,0.0329075555006663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,2048,0.051923556460274585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,16384,10240,0.052941332260767617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,2048,0.011718221836619906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,256,0.0021271111650599376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,256,10240,0.022034666604465906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,2048,16384,0.08069066868888007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,2560,0.08119110928641425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,16384,64,0.02498311135503981
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,3072,12288,0.02868888775507609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,10240,0.028468443287743464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,65536,2048,0.022946667340066697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,12288,16384,0.35472443368699813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,5120,0.4966728952195909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,16384,0.030364443858464558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,1024,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,5120,128,0.005895111295912001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,256,7168,0.010840000377760993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,8192,0.10463733143276638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,256,3072,0.017501332693629794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,768,16384,0.03240177697605557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,1024,6144,0.02113866640461816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,1536,7168,0.010379555324713389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3584,65536,0.07716800106896295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1536,10240,0.0207004447778066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,256,0.006611555400821898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,10240,0.02475022276242574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,51200,32,0.0028577778074476454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,256,0.017982222967677645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,10240,51200,0.7816275490654839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,7168,0.07838133308622572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,6144,5120,0.01033866653839747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,16384,512,0.012357333468066322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2048,32,0.0042053332759274375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2560,7168,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,1024,8192,0.02128444446457757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,3072,2048,0.038506666819254555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,6144,65536,0.12133333418104385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,7168,0.01218311074707243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,65536,4096,0.049694223536385425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,7168,0.06498666604359944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,51200,0.33075110117594403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,10240,8192,0.20878844790988496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,64,128,0.005557333429654439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,51200,4096,0.04008177916208903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,7168,0.033739556868871055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,1024,0.010023111270533668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,5120,0.028194665908813477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,2560,5120,0.027408000495698717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,12288,0.08628977669609918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,4096,0.03195644418398539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,8192,0.1084337764316135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,7168,8192,0.03868800070550706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,6144,4096,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,4096,3584,0.03678933448261685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,7168,6144,0.050168000989490084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,512,10240,0.012395555774370829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,768,3584,0.01682488951418135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,512,7168,0.05548977851867676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,32,12288,0.016382222374280293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,768,128,0.008300444318188561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,16384,4096,0.2612888813018799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,8192,5120,0.2486328813764784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,16384,4096,0.0432586669921875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,1024,1024,0.015839111473825242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,1536,10240,0.025014221668243408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,10240,6144,0.021943110558721755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1536,3584,0.039134220944510564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,7168,0.04117600123087565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,10240,0.27434754371643066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,128,768,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,3072,1024,0.007718222008811102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,16384,0.02070844504568312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,65536,256,0.10165599981943767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,256,5120,0.008969777988062965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,256,768,0.005234666582610872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,8192,3584,0.08535733487870957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1536,768,0.01683022247420417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,768,51200,0.06885244448979695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,4096,4096,0.028031110763549805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3072,4096,0.02936799989806281
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,7168,0.13094577524397108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,768,0.03290844294759963
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,65536,0.08020889096789889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,128,10240,0.0481644438372718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,65536,0.0682800014813741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,256,0.008386666576067606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,2048,10240,0.015537778536478678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,65536,8192,2.7006087832980685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,7168,10240,0.14301155673133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,7168,256,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,768,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,6144,0.010016000105275048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,6144,0.009313777916961247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,4096,1024,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,128,7168,0.04837866624196371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,2048,51200,0.22334843211703828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,4096,0.024147555232048035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,3072,8192,0.051779554949866406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,10240,1024,0.007250666618347168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3072,12288,0.029005332125557795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,512,512,0.007272889216740926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,1024,3072,0.008302222523424361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,1024,1024,0.004589333302444882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,10240,128,0.006585777633719974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,10240,32,0.018191999859280057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,2048,0.004231111043029361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,256,768,0.002513777878549364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,10240,5120,0.6581102477179633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,8192,65536,0.08088355594211154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,1024,0.007961778177155389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3584,768,0.07503022087944879
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,7168,0.6544604301452637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,4096,12288,0.03798310955365499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1536,1536,0.022287110487620037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,16384,0.02026844521363576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,1024,0.01311555587583118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,4096,65536,0.12012978394826253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,768,1024,0.004504000147183736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,256,256,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,7168,256,0.018532445033391316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,1024,0.01499466598033905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,6144,1024,0.017895999881956313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,128,7168,0.024077332682079736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,64,16384,0.011698666546079846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,2048,0.5042906867133247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,3584,0.013447110851605734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,10240,2560,0.05093600021468269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,5120,0.01809866726398468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1536,5120,0.018200889229774475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,256,512,0.002584888910253843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,2048,3072,0.019310222731696237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1536,51200,0.3608720037672255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,5120,8192,0.018430221411916945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,32,51200,0.18264799647861055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,6144,0.019955555597941082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,3584,2.749815199110243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,51200,3584,0.03850133220354716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,64,8192,0.016520889268981088
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,512,3584,0.009679999616410997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,10240,1536,0.2555866771274143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,16384,64,0.008293333152929941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,7168,10240,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,12288,2560,0.01253600004646513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,6144,12288,0.01993777851263682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,1536,51200,0.5543395678202311
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,2048,0.004570666700601578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,256,0.0032071111102898917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,32,2560,0.007289778027269576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,256,7168,0.018561777141359117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,65536,2.1318933698866105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,7168,1024,0.024023110667864483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,3072,3072,0.04794666502210829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,16384,0.15248800648583308
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,6144,64,0.010501333408885531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,512,0.009365333451165093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,5120,0.011434666812419891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,768,0.006621333460013072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,1536,12288,0.016474665866957772
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,3584,0.03281599945492215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,8192,3072,0.12946311632792154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,10240,2560,0.02674755619631873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,65536,512,0.05660711394415962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,12288,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,768,5120,0.018592000007629395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,16384,6144,0.5129929118686253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,128,0.01962933275434706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,51200,12288,1.9962452782524958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,65536,3072,0.46849335564507383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,5120,1536,0.15424533685048422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,1536,0.03257688879966736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3072,4096,0.02962133288383484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,7168,3072,0.033260444800059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,10240,0.017158221867349412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,65536,5120,0.8200782140096029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,256,65536,0.0767715573310852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,8192,16384,0.02567555507024129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,5120,64,0.002177777803606457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,16384,0.024334222078323364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1024,256,0.005209777918126848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,10240,16384,0.0769617756207784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,8192,0.06507200002670288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,2048,768,0.005943111247486538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,5120,8192,1.297507603963216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1024,12288,0.013783111340469785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,12288,3072,0.031070222457249958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,7168,1536,0.035014222065607704
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,512,128,0.008941333327028487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,512,512,0.022325333621766832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,768,256,0.0052062223354975385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,128,10240,0.026722666290071275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,6144,65536,0.14066133234235975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,16384,8192,0.04408533374468485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,6144,3584,0.02843822373284234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,7168,2560,0.037194665935304426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,12288,3584,0.01847199930085076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,1536,7168,0.034471111165152654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,65536,4096,0.39299821853637695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,5120,0.03257866700490316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,51200,0.1798888842264811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,10240,0.01779911087618934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,51200,5120,0.7166613472832574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,2048,0.031151112582948472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1536,2560,0.011390222443474663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,768,3072,0.014303111367755465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,1024,0.028519110547171697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,64,7168,0.015169777803950839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,4096,0.00814222213294771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,16384,128,0.006315555423498154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,65536,0.04036888811323378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,32,8192,0.1380524502860175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,5120,0.008882666627566019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,1536,5120,0.009327111144860586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3584,3584,0.040885332557890154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,768,0.003519111209445529
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,512,0.01687022215790219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,512,12288,0.042111112011803516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3584,1536,0.01997511088848114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3584,128,0.012378666963842181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1024,768,0.005598222215970357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,8192,0.013975999421543546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,256,0.011746666497654386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,1536,0.021594666772418555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,5120,12288,0.057370669311947294
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,128,10240,0.07360977596706815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,12288,51200,0.17254400253295898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,10240,128,0.02363377809524536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,32,0.0032106666929192017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,2560,0.1079217791557312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,6144,0.013795555465751223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,10240,0.050408889849980675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,64,0.0018044445249769424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,256,16384,0.02751111156410641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,6144,51200,0.15168978108300105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,8192,768,0.006036444256703059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,6144,0.036624001132117376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,768,0.005623111294375525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,64,65536,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,2560,8192,0.029560890462663438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,3072,0.016186666157510545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,8192,1536,0.021780444516075984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,4096,0.018070222602950204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,8192,0.010684444672531553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,12288,0.10994311173756917
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,128,3072,0.007378666765160031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,1536,128,0.005612444546487596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,7168,0.013223110801643796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,5120,256,0.005496888938877318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,65536,128,0.09095822440253364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,768,768,0.005569777968857024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,16384,7168,0.10050400098164876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,1536,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,65536,4096,0.45998668670654297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,6144,768,0.01682222220632765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,10240,0.01242844429281023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,128,1024,0.015909334023793537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,2048,0.013752000199423896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,4096,0.025158221522967022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,12288,512,0.01867644488811493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,3072,0.007263999846246507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,128,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,5120,16384,0.03561333484119839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,7168,65536,1.6141964594523113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,5120,768,0.02886577778392368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,6144,128,0.019934222102165222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,2560,0.013712888790501488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,256,0.006835555450783835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,64,0.0018479999982648427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,6144,65536,0.6883253521389432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,1024,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,512,0.003128000017669466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,768,8192,0.07480533255471124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,1536,7168,0.01726222203837501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,4096,0.008974221845467886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,3072,0.015648888217078317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,8192,51200,0.07699466413921781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,51200,0.10327733225292629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,51200,0.014800000521871777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,51200,0.06640266709857516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,12288,2048,0.056468447049458824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,12288,2048,0.009367110828558603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,12288,0.08263199859195285
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,65536,64,0.01016977760526869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2048,1536,0.013776000175211163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,256,0.0024826667375034755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,8192,4096,0.026174222429593403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,3072,12288,0.06024977895948622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,65536,12288,0.48518933190239805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,51200,3584,0.33063822322421604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3584,32,0.003156444471743372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,6144,0.01184266640080346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,16384,65536,10.383399963378906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,10240,0.31969067785474986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,16384,4096,0.018432888719770644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,7168,6144,0.19769510957929823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,256,0.0028613333900769553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,7168,8192,0.024533333049880132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,3584,12288,0.03457333313094245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,8192,0.012606222596433429
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,3072,0.11472888787587483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,4096,1024,0.042824887567096286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,6144,768,0.005906666732496685
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,32,1536,0.017503110898865592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,32,0.0025271111064487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,5120,0.009354666703277165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,768,0.0026391111314296722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,3072,0.07659466399086846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,16384,3072,0.015320888823933072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,128,0.015843555331230164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,768,128,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,16384,2048,0.015267555912335714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,128,0.002020444514022933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,65536,3072,0.2979297902848986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,2560,0.0069146665434042616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,7168,0.05222310953670078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,4096,10240,0.06867288880878024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,8192,128,0.015475556254386902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,3072,512,0.014549333188268872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,1536,0.0158595558669832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,12288,0.01202755586968528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,51200,12288,0.15247821807861328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,1024,4096,0.03442666596836514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,51200,10240,1.3899208704630535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,10240,1536,0.05822044610977173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,512,3072,0.01346666696998808
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,512,0.025790222816997107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1024,10240,0.013279111021094851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,128,3072,0.004907555464241239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,8192,2048,0.05118311113781399
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,16384,512,0.03211199906137254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,768,0.017242666747834947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,128,8192,0.03398666779200236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,128,32,0.00828888929552502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,512,0.008289777570300633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,10240,0.01385600037044949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,256,2048,0.005656888915432825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,6144,3584,0.46212175157335067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,1024,0.019494222270117868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,512,10240,0.01274222218328052
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,3584,256,0.023331556055280898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,256,0.002477333363559511
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,5120,8192,0.04274400075276693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,7168,0.060954663488599986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,3072,0.007307555940416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,256,512,0.011052444577217102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,3584,0.01960444450378418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2560,3072,0.006220444622966979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,10240,1024,0.00591555568906996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,1536,0.0038657776183552216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,3584,4096,0.05856888824039035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,5120,128,0.0021475555582178966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,5120,2560,0.030058665408028498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,512,6144,0.020256888535287645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,512,16384,0.016545777519543965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,32,2560,0.005377777748637729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,3584,2048,0.020239111449983384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,768,512,0.013097777962684631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,7168,8192,0.20518578423394096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,1536,1536,0.004913777940803104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1536,16384,0.03214844399028354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,5120,0.03715822100639343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,6144,256,0.00526311124364535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,2560,8192,0.11187377240922715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,16384,2048,0.04124710957209269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,256,128,0.005244444641802046
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3584,32,0.0052728888889153796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,128,65536,0.011181332998805575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,65536,16384,2.9574578603108725
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,16384,5120,0.017791999710930716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,16384,1.3243288464016385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,32,1024,0.013056000073750814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,3584,1024,0.02875822120242649
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,3072,0.030755556292004053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,7168,6144,0.04351466563012865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,768,10240,0.0451528893576728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,12288,1536,0.10534311003155178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,256,1536,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,3584,0.009034666750166152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,6144,7168,0.054768890142440796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,10240,10240,0.026781333817376032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,768,3072,0.005599110904667113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,8192,512,0.017902221944597032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,768,6144,0.006263999889294307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,8192,0.0331857767370012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,2048,256,0.005342222336265776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,1024,10240,0.02745955520206028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,5120,2560,0.018914666440751817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,4096,1536,0.004950222041871813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,10240,1.8068915473090277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,51200,16384,0.1258044507768419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3584,3584,0.0331911113527086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,256,256,0.0103493332862854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,256,10240,0.02738222148683336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,4096,0.007991111112965478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2048,3072,0.006266666783226862
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,256,1024,0.008804444637563493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,32,0.0025111111915773815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,5120,0.0222080002228419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,16384,3584,0.01404355631934272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,3584,0.008685333033402761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,51200,768,0.0120666664507654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,6144,256,0.009687999884287516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,768,3072,0.011691555380821228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,4096,1536,0.025067556235525344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,1536,512,0.009276444713274637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,65536,256,0.057722667853037514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,768,1024,0.009582222335868413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,7168,2560,0.02216533323129018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,512,1024,0.003495111233658261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,128,768,0.006225777582989798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,12288,65536,1.1841244167751737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,12288,0.040808889600965716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,10240,0.016190222567982145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,5120,2048,0.05813866853713989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,2048,1536,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,512,7168,0.02057066725360023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,128,16384,0.15162132845984563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,8192,0.17502578099568686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3584,6144,0.035058667262395225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,51200,2048,0.029133333100212946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,8192,2048,0.09228889147440593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,16384,5120,0.083697775999705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,10240,1536,0.024375999967257183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,65536,0.0569795568784078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,1536,3072,0.028816001282797918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3584,65536,0.19663200113508436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2560,2560,0.0439955559041765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,4096,0.016144000821643405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,10240,6144,0.01902311046918233
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,4096,0.19936622513665092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,3584,128,0.009489778015348647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,6144,10240,0.03736622134844462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,64,12288,0.013767999907334646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,32,3584,0.009694221946928237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,12288,0.023970666858885024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,256,1024,0.006945778098371293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,12288,1536,0.025690666503376428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3072,3584,0.008815999660227034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,3584,51200,0.10244888729519314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,16384,51200,0.11950043837229411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,3072,64,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,128,65536,0.05304444498485989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,512,2560,0.00500177757607566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,128,65536,0.011520889070298938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,12288,32,0.007646222081449296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,16384,2560,0.033779554896884494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,8192,4096,0.010661332971519895
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3072,32,0.004523555437723796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1536,8192,0.01482488877243466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,5120,768,0.007974222302436829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1536,1536,0.006626666833957036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1536,128,0.012111999922328524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,8192,6144,0.014192000031471252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,2048,0.0047271110945277745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,256,64,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,1536,768,0.009726222190592024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,32,0.01683822274208069
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,5120,51200,0.14938133292728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,2048,51200,0.15335022078620061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,10240,512,0.04456088940302531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,1024,0.012574222352769641
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,10240,0.029674665795432195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,7168,3072,0.0195777780479855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,51200,5120,0.659896903567844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,3584,512,0.02887822190920512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,1536,0.0064995553758409284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,512,0.013086222112178802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,12288,1536,0.08396799696816339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,12288,10240,0.2545582188500298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,128,256,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,2048,0.03542933199140761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,2048,0.02879377868440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,128,8192,0.019926221834288705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,256,51200,0.028784887658225164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,512,1024,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,2560,0.47994576560126406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,1536,8192,0.043566223647859365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,12288,65536,0.13486933708190918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,10240,2560,0.09956977764765422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2560,1024,0.008299555215570662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,3584,6144,0.06674310896131727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,3072,0.020463110672103036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,2560,0.011596444580290051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,5120,3072,0.008595555192894405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,5120,0.027619555592536926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,6144,0.024039111203617517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,768,6144,0.00998222248421775
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,128,256,0.002124444478087955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,6144,0.009311111436949836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,65536,51200,0.40804799397786456
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,51200,0.04820800158712599
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,6144,3584,0.022673777408070032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,256,0.0024604443460702896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,4096,1024,0.006271111054552927
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,256,1536,0.0036071112586392295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,51200,8192,0.1511466635598077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,51200,8192,0.07281688849131267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,4096,32,0.006768888897365994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,256,0.0058702222175068325
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,5120,32,0.0021377777059872947
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,1024,51200,0.08172711398866442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,4096,0.14143999417622885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,2048,0.04658133453792996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1536,256,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,3584,8192,0.014647111296653748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,2560,0.07176178031497531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,768,65536,0.02570399973127577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,256,768,0.019904888338512845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,2048,256,0.01957155598534478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,4096,0.5465928713480631
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,16384,0.04631466666857401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,256,3072,0.018590221802393597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,4096,0.04148355457517836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3072,1024,0.02778488728735182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,32,0.0022693334354294673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,512,0.01553244392077128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,512,128,0.00249422217408816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,5120,2560,0.007024889190991719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,10240,8192,0.05581866370307075
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,1536,5120,0.008355555435021719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,6144,0.010062221851613786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,64,5120,0.01924800044960446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,7168,2048,0.0069288888739215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1536,8192,0.01367911116944419
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,768,0.01515111161602868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,32,0.00675999994079272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,64,0.015145777000321282
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,51200,51200,10.127360026041666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,16384,0.14644711547427708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,128,2560,0.009675555759006077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,3584,5120,0.10015644629796346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,3584,0.3792773352728949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,768,0.018897778458065458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,768,0.01889955500761668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1536,10240,0.04784799946679009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,768,0.006310222049554189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,2560,0.027401778433057997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,1536,5120,0.007019555403126611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,5120,2560,0.021597334080272253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1024,1024,0.0079644446571668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,128,2048,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1024,256,0.006646222124497096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,128,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,51200,10240,0.793142212761773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,10240,512,0.03380889031622145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,256,10240,0.030969778696695965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,512,768,0.02169955604606205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1024,768,0.01585422290696038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,4096,3584,0.029389331738154095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,1536,0.01684177749686771
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,8192,0.043896890348858304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,8192,16384,0.04957511027654012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,512,7168,0.009482666850090027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,64,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,128,3072,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,5120,1536,0.00452888881166776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,3072,5120,0.010671999719407825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,3072,3072,0.028464890188641016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,512,0.037964443365732826
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,768,10240,0.03696977761056688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,768,1536,0.00444444434510337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,65536,4096,0.24959021144443086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,2560,0.1755759980943468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3072,768,0.028892444239722356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,16384,51200,0.6361049016316732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3072,128,0.01959822244114346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,1024,16384,0.019984000258975558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,3584,768,0.012702222499582501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2560,768,0.01648533344268799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,51200,16384,0.6123421986897787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,256,0.006244444598754247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,768,4096,0.007295110987292395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,8192,3072,0.0650693310631646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,768,3072,0.015999111864301894
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,5120,0.011968889170222811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,4096,0.04313244422276815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,51200,8192,0.08012888828913371
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,1024,0.0228426671690411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,5120,7168,0.5377884440951878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,128,0.015463999576038785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,1024,0.022451554735501606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,3584,2048,0.026150221625963848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,6144,3584,0.048476444350348584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,7168,1536,0.02879377868440416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,65536,0.04882133338186476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,4096,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,5120,0.004457777573002709
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,3584,256,0.0028213332924577924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,12288,51200,2.765106625027127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,2048,0.015987555185953777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,512,7168,0.011384888655609555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,51200,0.056688891516791455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,8192,7168,0.06881511211395264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,2560,2048,0.04350044329961141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,7168,65536,0.08317689100901286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,32,32,0.0018737777653667664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,51200,256,0.04017511010169983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,128,16384,0.025394666526052687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,64,7168,0.010965333216720156
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,128,8192,0.04003377755482992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,51200,10240,1.3400808970133464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,65536,6144,0.29174666934543186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,1536,0.025397333833906386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,32,4096,0.019831111033757527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,768,4096,0.019908444748984445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,65536,5120,0.08799199925528632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1536,2048,0.025059555967648823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,16384,0.04612000121010674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,16384,2.009898715549045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,5120,65536,0.30375467406378853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,51200,65536,0.43126667870415586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,51200,3072,0.4838382403055827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,3072,10240,0.017124444246292114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,6144,0.039000001218583845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,51200,1536,0.10621244377560085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,6144,128,0.02812088860405816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,128,7168,0.011717333561844297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2560,65536,0.06655377811855741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,12288,8192,0.03168711066246033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,10240,65536,2.034348381890191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,10240,12288,0.18512711260053846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,5120,1024,0.006984000404675801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,5120,16384,0.11295110649532741
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,6144,768,0.011032889286677042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,5120,7168,0.01331022216214074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,256,0.0022213332768943575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,2048,2048,0.01573600040541755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,2560,65536,3.2195741865370007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,4096,2048,0.03803911142879062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,8192,51200,1.5513110690646703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,51200,3584,0.06138133340411716
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,7168,7168,0.15450133217705622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,768,7168,0.02653955585426754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,12288,6144,0.01902577777703603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,16384,7168,0.026889777845806543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,256,0.026353778110610113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,3584,1536,0.02162666618824005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,12288,128,0.0029022221763928733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,4096,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,10240,2048,0.06771110826068454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,65536,768,0.03758577836884393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,51200,1536,0.020767110917303298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1536,1536,0.021293333835071985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,128,16384,0.010346666806273991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,51200,8192,0.1410542196697659
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1536,1536,0.029940442906485662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,51200,6144,1.8701769510904949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3072,512,0.023315555519527856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,2048,5120,0.008096888661384583
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,128,16384,0.07064711385303073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,5120,0.03805333375930786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,256,0.018199111024538677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,65536,0.006604444649484422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,1024,10240,0.008064889245563084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,16384,10240,0.42768179045783145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,12288,51200,0.11373066902160645
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,2048,65536,0.050380445188946195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,12288,7168,0.14805244074927434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,16384,65536,0.15766310691833496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,7168,64,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,7168,2560,0.04094488753212823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,2048,0.015593777100245157
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,64,16384,0.012890666723251343
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,5120,0.023423110445340473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,10240,12288,0.13788178231981066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,8192,0.015419556034935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3584,4096,0.02054222259256575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,512,10240,0.053220444255405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,10240,4096,0.017972444494565327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,8192,256,0.051353779104020864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,51200,8192,0.11791466342078315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,5120,768,0.006233777850866318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,12288,1536,0.008889777792824639
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,2048,0.11282755268944634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,3072,5120,0.036064889695909284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,768,5120,0.014913777510325113
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,51200,1024,0.057152887185414634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,1024,0.005917333480384614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,2048,2560,0.04200622108247545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,3584,16384,0.25858667161729604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,51200,768,0.08964800172381931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,8192,7168,0.1285760005315145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,32,0.00574933323595259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,64,6144,0.02275466587808397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2048,1536,0.006583111153708563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,3072,0.02812622321976556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,4096,0.0675742228825887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,2048,6144,0.01015022231472863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,51200,65536,11.133910285101997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,2560,0.3430337905883789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,256,6144,0.023958222733603582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,3072,16384,0.022308443983395893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,5120,3072,0.03203199969397651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,7168,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,768,0.030574222405751545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,256,3584,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,7168,0.017535999417304993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,768,0.018224888377719455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,1024,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,2560,4096,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,12288,8192,0.027330666780471802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,10240,12288,0.035107556316587664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,65536,3072,0.093176888095008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,64,65536,1.0311892827351887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,6144,0.013110222087966071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,768,512,0.01406044430202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,256,2048,0.016833777228991192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,32,10240,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3072,6144,0.009364444348547194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,2048,256,0.015205333630243937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,1024,0.020563556088341605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,128,0.00977244476477305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1536,65536,0.07974044481913249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3072,7168,0.08674666616651748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,51200,3072,0.026290666725900438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,7168,2560,0.045935998360315956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,512,12288,0.023219555616378784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,128,0.1271253294414944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,6144,65536,0.12787911627027723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,64,0.004045333299371931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,512,4096,0.0069395556218094295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,5120,1536,0.006248888870080312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,3584,3584,0.06423911121156481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,768,768,0.005902222047249476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,6144,0.5034906599256728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,128,512,0.011950222154458364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,512,65536,0.3794106642405192
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,16384,1024,0.03412533468670315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,1536,768,0.005952888892756567
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,2560,128,0.014835556348164877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,1536,1536,0.013769778112570444
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,64,10240,0.01239377756913503
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,8192,0.005076444397370021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,65536,10240,0.21202399995591906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,512,7168,0.024373332659403484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2560,12288,0.01752444439464145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3584,2560,0.0075333333677715724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,5120,32,0.009467555416954888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,1024,768,0.0028373334142896864
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,3072,0.006790222393141852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,128,10240,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,8192,768,0.019475556082195707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,51200,16384,0.24914222293429902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,51200,0.24088711208767363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,51200,768,0.21093422836727568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,3584,0.007696888513035244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,10240,2560,0.07235644261042277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,3584,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,6144,0.01960177719593048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,51200,1024,0.015578667322794596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,2560,12288,0.019367110398080613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,4096,2048,0.0508284436331855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,6144,10240,0.03616266780429416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,4096,10240,0.01615111033121745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,1024,0.018545778261290658
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,256,512,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,7168,256,0.005873777800136142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1536,51200,0.36945245001051163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,12288,1024,0.06943111287222968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,51200,128,0.03744355506367154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,128,3584,0.01646133263905843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,1536,3584,0.014116444521480136
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,768,128,0.005565333283609814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3584,5120,0.026722666290071275
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,768,5120,0.025417778227064345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,32,3072,0.021533333592944678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,2560,8192,0.01038044442733129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,6144,10240,0.19738845030466715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,256,65536,0.0166631109184689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,512,16384,0.06564266814125909
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,128,768,0.017872000734011333
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,256,0.0028417776856157514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,6144,8192,0.19151200188530815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,64,0.002483555633160803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,10240,1024,0.02940000096956889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,8192,51200,0.10321955548392402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,65536,5120,0.3724844455718994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,12288,0.04267911116282145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,2048,10240,0.03948355714480082
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,1536,0.004864000197913912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,10240,1536,0.02217955556180742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,128,0.024403555525673762
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,1024,1024,0.00674933319290479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,32,16384,0.06465333037906222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,32,3072,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,3072,2048,0.012420444852775998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,64,8192,0.0069413334131240845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,3584,16384,0.020868443780475195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,51200,0.18641689088609484
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,7168,3072,0.00978133330742518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3072,512,0.0028239999794297745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,64,1536,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,128,0.0024951110697454875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,10240,2560,0.012067555553383298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,1024,0.01684977776474423
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,2048,0.007304000357786815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,6144,1024,0.004031110968854693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,2560,0.0529333319928911
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,51200,65536,0.40600000487433535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,128,2560,0.013077333569526672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,3584,0.013801777528391944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,64,2048,0.009355555805895064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,32,4096,0.004579555657174853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,65536,0.06706844435797797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,3072,256,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,5120,128,0.002805333377586471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,5120,10240,0.021233777205149334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,5120,0.0653724471728007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,12288,4096,0.08122399780485365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,8192,0.010718222293588849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,6144,768,0.02149066660139296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1024,1536,0.0069368887278768755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,65536,256,0.014073777529928418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,4096,0.022053332792388067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,5120,10240,0.018744000130229525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,64,0.009648000200589498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,256,3072,0.008388444781303406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,4096,12288,0.1099626686837938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,10240,10240,0.030766222212049697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,128,2048,0.009316444396972656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,10240,0.05360355642106798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,512,8192,0.023016000787417095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,4096,1024,0.03917422228389316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,3584,0.6807404624091254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,2560,3072,0.024676443801985845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,128,2048,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,51200,0.11973867151472305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,12288,51200,0.11153688695695664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,32,10240,0.005311111195219888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,1536,10240,0.04157600137922499
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,512,1024,0.014898666077189975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,64,7168,0.010088889135254754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,2560,4096,0.010280888941552904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3584,64,0.007903999752468532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,7168,2048,0.1191377772225274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,64,6144,0.006565333240562015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,7168,0.033688889609442815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,65536,3584,0.7228684425354004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,65536,768,0.04352622230847677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,256,0.0052622221410274506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,64,6144,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,5120,256,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,1536,0.006616000086069107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3072,65536,0.06565511226654053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,1536,128,0.014851555228233337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,65536,512,0.01440622243616316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,5120,1536,0.0158897770775689
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,8192,8192,0.023306666149033442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,4096,1024,0.020262221495310467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,1536,512,0.005570666657553778
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,2560,6144,0.01899377836121453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,65536,1536,0.023465777436892193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,5120,0.038395557138654925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,16384,10240,0.8452417585584852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,256,64,0.01071200023094813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,4096,10240,0.01685599982738495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,512,51200,0.08578666713502671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,7168,3072,0.03323555654949612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,5120,12288,0.056200000974867076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,10240,0.07994311385684542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,4096,5120,0.0898053314950731
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,10240,12288,0.03191733360290527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,128,0.019159111711714003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,16384,5120,0.020896888441509668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,5120,512,0.02199288871553209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,256,512,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,16384,0.06965777609083387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,128,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,16384,6.551981184217666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,12288,1536,0.2589893341064453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,1536,0.040928890307744346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,256,0.006249777972698212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,7168,0.010025777750545079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1536,7168,0.041145778364605375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1536,1024,0.015134221977657743
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,5120,3072,0.007292444507280986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,768,1536,0.012117333710193634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,128,8192,0.08941511313120525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,2048,65536,0.17704533206091988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,3072,0.014883556299739413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,51200,8192,0.9891431596544055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,12288,2560,0.10634399784935845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,64,2048,0.016607999801635742
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,2048,12288,0.11570310592651367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,32,0.013366222381591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,2048,256,0.0174666659699546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,6144,65536,0.08282844225565593
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,1024,512,0.02973866793844435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,51200,5120,0.4115280045403375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,65536,0.0697644419140286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2560,1536,0.018458666072951425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,16384,256,0.0069724445541699724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,128,10240,0.012048000262843238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,2560,2048,0.00424888895617591
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,2560,0.006232000059551663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,3072,6144,0.013406222065289816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,3584,0.005587555468082428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,5120,1024,0.028008888165156048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2048,16384,0.030582222673628066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,65536,0.1496115525563558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2560,65536,0.06655999686982897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,5120,7168,0.12724444601270887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,1024,0.015308444698651632
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,12288,512,0.005606222069925732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,512,32,0.001850666685236825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,5120,1024,0.017834666702482436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,8192,8192,0.19650667243533662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,64,65536,0.01647644407219357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,65536,0.06876089175542195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,2560,0.03142933381928338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,6144,3072,0.09675111373265584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,768,4096,0.012085333466529846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,16384,0.9941635131835938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,1024,0.006187555690606435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,64,0.011389333340856763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,2048,256,0.002135111019015312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,6144,512,0.003152888889114062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,16384,16384,0.9295164744059244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1536,6144,0.011423110961914062
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,16384,10240,1.3408880233764648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,10240,4096,0.5825511084662544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2560,6144,0.14348888397216797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,8192,10240,0.02868888775507609
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,65536,0.21724355220794678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,3584,256,0.005559999909665849
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,8192,16384,0.0646524429321289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,512,51200,0.1490666601392958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,3584,256,0.01479466590616438
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,128,512,0.01205244412024816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,7168,0.13694132698906794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3584,3072,0.047338667843076915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2560,7168,0.05134400063090854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,3072,0.07200977537367079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,128,10240,0.01426755554146237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,1024,0.023741333021057978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,32,32,0.006959999601046245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,10240,0.4109662108951145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,2048,0.021428444319301184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,128,64,0.007323555648326874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,2048,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,128,64,0.007919111185603673
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3072,256,0.002185777864522404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,5120,51200,1.0446613099839952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,3072,4096,0.026741334133678015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,7168,0.020238222347365487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,128,256,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2048,8192,0.01108000013563368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,7168,10240,0.055645333396063916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,10240,0.05589422252443102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,12288,0.03872444563441806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,768,3072,0.016460445192125108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,1536,768,0.003173333282272021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,7168,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,256,0.021553777986102637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,4096,16384,0.5722026824951172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,1024,1536,0.006619555668698416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,128,0.001847111102607515
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,3072,0.07303200165430705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,2048,64,0.005958222266700532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,5120,0.014085332552591959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,768,2048,0.015811555915408664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,12288,0.20052977403004965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,32,2560,0.005047111047638787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,8192,0.025974222355418738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,3584,0.008284444610277811
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,12288,51200,2.7207200792100696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,16384,64,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,6144,2560,0.008721777962313758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,1024,512,0.002883555574549569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,16384,0.03460622164938185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,3072,4096,0.010666666759385003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,512,8192,0.011043555206722684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,3072,64,0.014114666316244336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,1536,256,0.00630666646692488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,3584,0.02223644488387638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,51200,1024,0.2608106666141086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,512,0.0024862223201327855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,3584,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,10240,2048,0.026484444737434387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,51200,128,0.011571555501884885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,2048,8192,0.100545777214898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,2560,1024,0.0038480001191298165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,2560,64,0.01034311122364468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,768,16384,0.14599200089772543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,7168,2560,0.026906667484177485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,5120,0.03566311134232415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,5120,12288,0.21850844224294028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1536,16384,0.17237067222595215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,65536,0.3541831175486247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,12288,0.08466844426261054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,8192,16384,0.05253244439760844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,4096,2560,0.040384001202053495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,51200,32,0.012663111090660095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,6144,768,0.023057777020666335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,1536,0.037286221981048584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3584,3584,0.025352888637118872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,65536,14.111642625596788
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1536,1536,0.0159599996275372
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,16384,2048,0.018372444642914664
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,10240,2560,0.02779910961786906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,6144,65536,0.17659466796451143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,6144,256,0.03355822298261855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,65536,10240,0.18181067042880586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,4096,4096,0.030656887425316706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,1024,51200,0.08743377526601155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1536,768,0.0058808889653947614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,65536,768,0.022988445229000513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,65536,1024,0.03921688927544488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,256,16384,0.46915197372436523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,3584,0.013400889105266996
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,12288,51200,0.1366222169664171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,2560,1536,0.009767110976907942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,2560,65536,0.06458222203784518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,256,0.01551466683546702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,4096,5120,0.009886222581068674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,51200,2048,0.18759022818671334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,256,8192,0.00554933316177792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,7168,2560,0.03320711188846164
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,7168,0.010113777385817634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2048,768,0.005944889038801193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,768,0.002532444480392668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3584,16384,0.13934489091237387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,768,1536,0.004173333446184794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,4096,12288,0.018581334087583754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,7168,0.008644444247086843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,12288,0.054530666934119336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,1536,0.012510221865442065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,7168,1024,0.01256444470749961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,1024,512,0.013156444662147097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,64,0.004879999905824661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,12288,768,0.024726221958796184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,1536,0.01719999975628323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,768,3584,0.00796977761718962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,512,1024,0.0028684443483750024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,256,6144,0.005136000199450387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,3072,12288,0.05714311202367147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,768,0.746488889058431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,16384,0.01586577792962392
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,16384,0.027768888407283362
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,51200,16384,4.599405500623915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,128,10240,0.029448890023761328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,768,51200,0.21959377659691703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,512,12288,0.011626666618718041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,12288,256,0.027601778507232666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,65536,512,0.07907288604312472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,4096,1536,0.03235733177926805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2560,3072,0.04425422350565592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,3072,3072,0.022077333596017625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,768,512,0.0034888887570963967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,64,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,16384,7168,0.025666667355431452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,12288,2560,0.07043466965357463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,64,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,10240,0.0410844451851315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,2560,0.024130667249361675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,128,3584,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,256,3072,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,128,0.013829333086808523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,8192,256,0.002492444382773505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,128,7168,0.010017777482668558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,12288,6144,0.04450044367048475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,1024,0.014753777119848462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,8192,6144,0.013420444395807056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,8192,7168,0.12272977828979492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,256,4096,0.015496888094478183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,1536,256,0.0052488889131281115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,65536,128,0.02539199921819899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,12288,128,0.00906044410334693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,6144,0.012097777591811286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,768,2560,0.017635555730925668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,16384,6144,0.021425777011447485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,1024,0.025569778349664476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,8192,0.03916355636384752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,128,6144,0.028042667441897925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,7168,65536,0.08071288797590467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,768,0.026551999979548987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,3072,5120,0.009831111464235518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,5120,0.025158221522967022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,3072,0.02544533378548092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,2560,5120,0.009151111046473185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,1536,1536,0.015325332681337992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,768,16384,0.13591644499037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,16384,65536,3.9170985751681857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,4096,7168,0.013088000317414602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,51200,5120,0.10454933510886298
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,3584,16384,0.718474653032091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,5120,768,0.005799110978841782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,2560,2560,0.018927110566033255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,256,0.00867911097076204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,12288,128,0.006284444282452266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,64,10240,0.016055110428068373
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,8192,512,0.013070222404268054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,128,3072,0.023297778434223596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,2048,16384,0.021617778473430212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,16384,1536,0.01092888911565145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,1536,6144,0.008952889177534316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,16384,3.8224978976779513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,7168,12288,0.027279999521043565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2560,4096,0.010775110787815519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,2048,0.010794666906197866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,32,10240,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,65536,2560,0.06109155548943413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,64,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,65536,2560,0.0594959987534417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,16384,0.10688177744547527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,16384,256,0.003522666792074839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,7168,5120,0.014466666513019137
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,16384,1024,0.06311021910773383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,768,0.025919111238585577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,6144,0.011376889215575324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,1024,0.01570222278436025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,7168,0.04017422265476651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,6144,2560,0.023992000354660883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,2048,3584,0.00664444433318244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,8192,2048,0.05552800165282356
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,51200,64,0.009674666656388177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,12288,256,0.002489777695801523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,256,2560,0.004911111046870549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,128,16384,0.01757422255145179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,16384,16384,0.5942879782782661
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,51200,8192,0.08180089129341973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,6144,0.0672817760043674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,64,0.0025173332542181015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,10240,2048,0.04477599925465054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,2560,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,512,3584,0.00794044468137953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,64,7168,0.01036888857682546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,8192,256,0.005559111220969095
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,65536,0.13582311736212835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,65536,3072,0.0376853346824646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,1536,0.1773831049601237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,12288,4096,0.08122044139438205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,128,16384,0.01650044487582313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,1536,10240,0.01445777714252472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,1536,65536,0.03569333420859443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,768,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,128,1024,0.016135111451148987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,64,6144,0.009902222288979424
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,32,5120,0.01109333336353302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,1536,64,0.006269333263238271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,1536,12288,0.04357066750526428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,2048,2560,0.0520666672123803
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,6144,16384,0.11567821767595078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,2048,8192,0.1594453387790256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,6144,0.07876711421542697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,4096,0.017440888616773818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,5120,2560,0.022608889473809138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,128,65536,0.06291466951370239
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,12288,65536,0.14632533656226263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,12288,768,0.017143999536832173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,5120,2048,0.01443199978934394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,128,256,0.006622222148709827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,256,0.009700444837411245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,7168,0.07097066773308648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,512,65536,0.059228446748521596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,2048,12288,0.052876445982191295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,10240,0.034733331865734525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,10240,512,0.025435555312368605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,3584,4096,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,2048,256,0.005899555567238066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,10240,64,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,8192,0.03191377719243368
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,512,7168,0.01937866707642873
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,65536,0.07136178016662598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,3072,0.022322666313913133
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,6144,8192,0.05647822221120199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,3584,7168,0.1471031109491984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,3072,0.007637333538797166
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2560,65536,0.1605751117070516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,768,2560,0.015822221835454304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,768,1536,0.0046471113132105935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,4096,4096,0.02623644471168518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,3072,16384,0.03764977720048692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,2560,7168,0.03341777788268195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,64,128,0.0021235555824306277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,2560,0.016295111841625638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,256,6144,0.01333866682317522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,2560,128,0.005592888842026393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,12288,12288,0.08335200283262466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,2048,6144,0.055649777253468834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,8192,6144,0.1827075613869561
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,7168,32,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,32,12288,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,256,512,0.0069715554515520734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,6144,256,0.023680888944202002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,2048,10240,0.046554668082131274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1536,256,0.005596444424655702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,6144,6144,0.013447999954223633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,3584,12288,0.016720000240537856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,256,8192,0.04151644309361776
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,5120,0.010950222611427307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1536,1536,0.06338133414586385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,1024,32,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,256,1536,0.0035013332962989807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,2048,128,0.014742222097184924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,6144,256,0.00216977774269051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,65536,0.21827822261386445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,4096,256,0.0028533333291610083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,8192,2560,0.012822222378518848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,512,10240,0.08365955617692734
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,51200,65536,0.5537404484219021
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,2048,7168,0.07871377468109131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,5120,65536,1.1430586708916557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,2048,5120,0.03832622369130453
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,768,512,0.017709333035680983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,512,128,0.010294222169452244
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,7168,51200,0.06283022297753228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,6144,0.04923911227120293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,10240,1.8207804361979167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,5120,512,0.033364444971084595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,16384,1024,0.009484444227483537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,64,0.0028044444819291434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,6144,2048,0.06257155868742201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,4096,256,0.01070666644308302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,512,2560,0.007322666545708974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,10240,0.2393102116054959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,51200,128,0.011735111474990845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,7168,5120,0.026445334156354267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,12288,10240,0.917176882425944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,1536,2048,0.025736888249715168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,8192,10240,0.10689333412382339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3584,3584,0.0076453329788313965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,64,1536,0.012053333222866058
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,7168,0.023403555154800415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,128,65536,0.06714577807320489
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,1024,51200,0.41872088114420575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,512,64,0.0069653333889113525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,4096,51200,0.07261511352327135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,256,0.005930666708283954
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,51200,512,0.17827022075653076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,256,3584,0.00795377790927887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,8192,7168,0.05726755327648587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,10240,8192,0.042936000559065074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,2048,0.03490488727887472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,512,7168,0.02029244436158074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,8192,4096,0.029577778445349798
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,128,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,4096,65536,0.09388177924686009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,12288,51200,0.7044275601704916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,2560,0.06702666812472872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,4096,6144,0.040720889965693154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,256,256,0.011699555648697747
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3072,12288,0.09451288647121853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2048,1536,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1536,51200,0.8991208606296115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,7168,0.038443555434544883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,1024,0.014721777704026965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,6144,4096,0.03304533494843377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,64,1024,0.009677333136399588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,512,128,0.006610666712125142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,7168,16384,0.027952889601389568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,4096,128,0.005960000058015187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,6144,65536,0.0936773353152805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,2560,51200,0.05275644527541267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,2560,0.015592889653311836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,65536,0.5778337584601508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,51200,3072,0.08206044303046332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,3584,0.007984889050324758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,256,2048,0.005197333378924264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,4096,128,0.02650844388537937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,51200,2.2866488562689886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,51200,3072,0.0321760012043847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,65536,0.061089780595567494
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,3072,8192,0.04340622160169813
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,1536,32,0.004310222135649787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,4096,768,0.022487110561794702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,1536,6144,0.01920355525281694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,64,51200,0.014621333943472968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,2048,1536,0.005200888961553574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,51200,1024,0.026073777013354834
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,32,0.1640524466832479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,5120,2048,0.05891022417280409
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,768,2048,0.004209777961174647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,64,10240,0.0124177775449223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,3072,768,0.016559110747443307
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,2048,1024,0.0059057776298787855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2048,16384,0.016061334146393668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,512,64,0.0018453333112928604
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,6144,0.009466666314336989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,7168,3072,0.008975110948085785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,7168,512,0.016866667403115165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,65536,0.07777422004275851
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,1024,512,0.026112889250119526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1536,3072,0.005583111196756363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,7168,5120,0.13176443841722277
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,3584,0.004956444518433677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,4096,0.00960355583164427
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,1536,51200,0.1374115546544393
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,32,5120,0.009011555876996782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,256,3584,0.01510933372709486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,4096,2048,0.03942311141226027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,1536,3584,0.02772355576356252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,7168,128,0.020276443825827706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,7168,12288,0.03757244348526001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,7,256,6144,0.004925333377387789
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,12288,7168,0.06982933150397407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,768,64,0.0018364443547195857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,768,128,0.005623111294375525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2048,3072,0.03585066729121738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,65536,0.6531128883361816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,8192,12288,0.4242382314470079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,2048,10240,0.02629599968592326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,1024,51200,0.06873599688212077
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,6144,3584,0.0776160028245714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,4096,2048,0.02847733431392246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,3584,8192,0.015793777174419828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,512,5120,0.0092613332801395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,2048,0.007531555162535773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,1024,1536,0.024078221784697637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,256,4096,0.015478221906556023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,128,12288,0.05746933486726549
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,7168,2048,0.00793599999613232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,3072,0.010665777656767102
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,6144,51200,0.09603910975986057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,2560,1024,0.005903999838564131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,65536,2048,0.23418309953477648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,5120,65536,0.1396391126844618
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,1536,768,0.005552000055710475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,2048,2560,0.029143999020258587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,32,3072,0.009703999592198266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,1536,5120,0.008651555412345463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,8192,0.17581600613064238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,65536,128,0.10243111186557347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,65536,4096,0.1126213338640001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,64,0.0018257778137922287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,4096,0.008832000195980072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,5120,51200,0.0995546645588345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,3072,65536,0.3110097779168023
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2560,256,0.005581333405441708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,12288,512,0.007809778054555257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,128,64,0.001828444500764211
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,12288,128,0.02552977701028188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,5120,0.6084933280944824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,512,0.00629688882165485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,128,0.00924444446961085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,7168,12288,0.019160888261265226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,7168,0.014030221435758801
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,5120,512,0.0028515555378463534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,8192,5120,0.09825333621766831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,10240,512,0.0038533334930737815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,3584,8192,0.013393777940008374
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,512,3584,0.011761777930789523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,5120,1024,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,3072,16384,0.1467662254969279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,12288,32,0.022152000003390845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1536,1024,0.005889777921968036
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,10240,8192,0.030576000610987347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,3072,12288,0.016861332787407767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,4096,3072,0.00865955568022198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,768,16384,0.03519377774662442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,3072,7168,0.012531555361217923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3584,3584,0.007778666913509369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,16384,0.022755554980701868
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,3584,2560,0.01759288873937395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,1024,1024,0.011728000309732227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,64,4096,0.00867733359336853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3584,12288,51200,1.9018505944146051
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,128,768,0.006291555447710886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,51200,256,0.024078221784697637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,512,2560,0.006974221931563483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,512,128,0.006603555546866522
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,16384,6144,0.24938665495978463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,3072,0.09060000048743354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,7168,1536,0.0321048895517985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,256,6144,0.021355556117163763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,512,51200,0.06462399827109443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,4096,0.026014222039116755
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,8192,10240,0.022436444958051045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,7168,6144,0.03897777862019009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,5120,1024,0.005988444305128521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,4096,7168,0.14753688706292048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,512,3584,0.007995555798212687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,16384,1024,0.026143999563323125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,1024,6144,0.0672337810198466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,512,0.00253866674999396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,12288,4096,0.07003288798862033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,7168,51200,0.18342667155795625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,3072,0.007326222128338284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,3584,64,0.0018293333964215384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,512,5120,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,128,3584,0.008269333177142674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,1536,0.029740442832310993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,65536,768,0.034252444903055825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,4096,51200,0.4284942150115967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,128,3072,0.007628444168302748
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,51200,128,0.08146844307581584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,768,12288,0.0723031096988254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,1024,2560,0.014782221780882942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,4096,0.014339556296666464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,65536,6144,0.05997688902748955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,16384,0.031238221459918555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,512,16384,0.01795644395881229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,3072,0.02808533443344964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,65536,5120,3.9259117974175344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,5120,8192,0.013629333012633853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,65536,5120,0.1274133390850491
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,256,2560,0.004882666799757215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,4096,3584,0.008919110728634728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,3072,1536,0.01870666609870063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,5120,3584,0.05927822324964735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,51200,128,0.01145333300034205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,1024,3584,0.04469244347678291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,65536,0.06402489211824205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,512,1024,0.007613333563009898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,16384,1536,0.08931555350621541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,512,2560,0.035740445057551064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,4096,128,0.0021155555215146807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,512,768,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,4096,128,0.002481777841846148
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,65536,256,0.0206666671567493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,128,3072,0.006865777903132969
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,8192,2560,0.010370666782061258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,10240,5120,0.029122667180167303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,3072,0.01034400032626258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,128,1024,0.008325333396593729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,224,3072,65536,0.1327271064122518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,1536,256,0.0031555555760860443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,7168,128,0.016728889611032274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,1536,0.009210666848553551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,2560,2560,0.005362666729423735
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,256,8192,0.019545776976479422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,2048,0.43200355105929905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,7168,8192,0.04474844535191854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,8192,65536,0.21538133091396758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,2048,0.008282666405042013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,65536,1024,0.1149253315395779
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,51200,512,1.345097753736708
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,51200,128,0.020611556039916146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,65536,5120,0.40506400002373594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,128,0.01406044430202908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,1024,256,0.008967999782827165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,10240,51200,17.26928032769097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,512,0.027730666928821143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,1024,0.008399999803966945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,1024,7168,0.08544800016615127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,256,5120,0.010039110978444418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,3584,1024,0.016864889197879367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,65536,3584,0.3113102118174235
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,3072,4096,0.09204977750778198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,10240,128,0.020608888732062448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,1024,128,0.01887022289964888
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,5120,768,0.01536799967288971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,512,12288,0.020823111136754353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,4096,0.009992889232105678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,768,0.013992889059914483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,5120,3584,0.00776533368561003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,3584,1536,0.028036445379257202
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,12288,51200,2.965218649970161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,128,32,0.006549333532651265
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,1024,4096,0.020288000504175823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,256,65536,0.06701778040991889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,768,8192,0.08762666914198135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,768,1536,0.006054222169849608
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,512,32,0.007977777885066139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,128,4096,0.006605333338181178
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,16384,0.040940443674723305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,2560,6144,0.021663111117151048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,4096,3584,0.007642666498819987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,768,5120,0.02353688908947839
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,12288,3072,0.01442488862408532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,10240,128,0.008662222160233391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,5120,10240,0.05265600151485867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,3584,5120,0.024296000599861145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,3584,8192,0.12377422385745579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,12288,0.21276444858974883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,1536,51200,0.2394568920135498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,5120,2048,0.007631111476156447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3072,512,0.00282222218811512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,64,5120,0.013551111022631327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,256,51200,0.06495110856162177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,10240,16384,0.06092533138063219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,1536,768,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,32,2560,0.029410666889614526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,10240,0.009668444593747457
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3584,2560,0.020678222179412842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,65536,768,0.017779555585649278
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,64,16384,0.01616533266173469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,512,6144,0.009706666900051964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,65536,2560,0.5028488900926378
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,2048,0.00888177752494812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,65536,5120,0.8887475331624349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,1024,2560,0.021693333983421326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,128,768,0.0035457776652442086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,16384,4096,0.02251288957066006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,3072,7168,0.013072888884279462
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,4096,32,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,16384,10240,0.055679109361436635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,2048,12288,0.01963200006220076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3072,5120,0.13689600096808538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,128,5120,0.07929155561659071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,6144,10240,0.040876444843080305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,256,51200,0.013553777502642738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,32,0.007987555530336168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,2560,10240,0.02613777750068241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,16384,12288,0.13269866837395564
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,51200,256,0.009890666438473595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,64,3584,0.011602666642930774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,4096,10240,0.018232888645595975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,51200,0.22455644607543945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,16384,0.061675555176205106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,3072,0.008007110820876228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,768,3072,0.028033776415718928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,12288,64,0.007629333270920648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,65536,32,0.0041999999019834726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,128,16384,0.012678222523795234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,32,1024,0.006270222365856171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,3072,0.02404533326625824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,8192,2.1254906124538846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,12288,2048,0.011732444167137146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,512,5120,0.01294133315483729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,65536,1536,0.02093511157565647
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,4096,5120,0.06379377841949463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,4096,4096,0.009756444229020013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,16384,2048,0.03229688935809665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,8192,12288,0.024118221468395654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,2560,1024,0.02466933263672723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,10240,3072,0.012654222548007965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,256,65536,0.14362133873833552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,256,512,0.005201777650250329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2048,1536,0.00612088872326745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,4096,16384,0.30830221705966526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,10240,512,0.017063111066818237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,3072,32,0.012864000267452665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,7168,10240,0.018182223041852314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,5120,128,0.014479110638300577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1024,3072,0.03253688746028476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,8192,51200,0.7397493256462945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,2560,0.01715022159947289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,1536,512,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,4096,768,0.022311111291249592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,3072,0.4451342158847385
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,7168,128,0.018168000711335074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,1536,1024,0.022823111878501043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,128,0.0018382221460342407
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,3584,0.00794222205877304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,1024,4096,0.006566222343179915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,16384,1024,0.01017777787314521
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,512,6144,0.01854222185081906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,768,1536,0.006538666784763336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3072,6144,0.012347555822796293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,7168,4096,0.061067554685804576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2048,1536,0.006593777901596493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,2560,5120,0.01682044400109185
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,3072,16384,0.10323999987708198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,51200,2048,0.061439110173119434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,3584,0.011711999773979187
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,6144,0.009648888475365108
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,10240,64,0.006568000134494569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,512,0.008993777963850234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,3072,12288,0.017920888132519193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,5120,2560,0.03450577788882785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,7168,16384,0.6336151229010688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,10240,8192,0.21827022234598795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,6144,0.011727111207114326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,7168,0.022750222020679053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,16384,2048,0.02959377898110284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,10240,1024,0.007124444676770105
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,2048,3584,0.017821333474583097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,6144,3584,0.035049776236216225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,6144,8192,0.06610933277342054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,65536,10240,1.2405999501546223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,2048,7168,0.07292711072497897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,12288,0.15759200519985622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,5120,32,0.022953778505325317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,8192,0.04008444481425815
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,65536,6144,0.10569333367877537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,1536,16384,0.11772710747188991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,512,7168,0.022608889473809138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2048,32,0.010699555277824402
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,65536,1024,0.3709377712673611
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,1536,7168,0.028806222809685603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,3584,51200,0.05503111084302267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,1536,10240,0.015470221638679504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,5120,2048,0.0069431112044387394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,6144,256,0.023176888624827068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,32,1536,0.0035084444615576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,65536,3584,0.3644995424482558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,7168,7168,0.02144444485505422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,4096,8192,0.05617066886689928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,10240,10240,3.7730060153537326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,3584,65536,1.1035084194607205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,51200,4096,0.03218844532966614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,1536,3072,0.027052443888452318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,256,12288,0.0203804439968533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,6144,12288,0.01819022165404426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,10240,0.011736888852384357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,256,0.012791111237472959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,1024,6144,0.01682400041156345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,3584,2048,0.006615110983451207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3584,65536,0.23223288853963217
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,8192,7168,0.01498933302031623
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,8192,2560,0.04128266705407037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,4096,0.00868711123863856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,768,10240,0.03422844409942627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,64,65536,0.0570711096127828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,1024,256,0.005222222043408289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,51200,0.10716178019841512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,128,4096,0.02155911094612545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,5120,0.00834222220712238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,16384,65536,1.9885537889268663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,768,256,0.00658044425977601
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,4096,7168,0.1279040045208401
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,7168,64,0.024336889386177063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,2560,12288,0.11179910765753852
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,4096,512,0.005777777896987067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,768,256,0.0021600000974204806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,7168,0.019959999455346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,1024,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,10240,0.01792355544037289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,32,6144,0.012780444489585029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,65536,0.06557155317730375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,768,10240,0.0135057775510682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,256,5120,0.009193778038024902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,3584,2048,0.037137776613235474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,3072,6144,0.013793778088357715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,7168,0.01664177742269304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,65536,12288,0.13211733765072292
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,51200,256,0.021591110361946955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,256,65536,0.5149155722724067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,51200,0.9300311406453451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,8192,768,0.02375999920898014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,128,4096,0.007640000018808577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,768,65536,0.3629591200086806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,512,10240,0.06664622492260404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,1024,2560,0.00490044429898262
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,1536,1024,0.024520888924598694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,7168,1536,0.02624799973434872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,1024,6144,0.011594666375054253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,2560,65536,0.09931200080447727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,256,3584,0.017549332645204332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,16384,1024,0.05953333112928602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,2048,768,0.00592533333433999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,128,12288,0.23690578672620985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,512,512,0.0025191110455327565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,2048,16384,0.15447644392649332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,3584,65536,0.13747110631730822
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,128,1536,0.003528888854715559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,3072,12288,0.017871111631393433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,5120,6144,0.014120888378885059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,51200,3072,0.025417778227064345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,51200,0.5188479953342015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,2048,7168,0.03298399845759074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,1536,256,0.009654222263230218
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,16384,3584,0.024660444921917383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,512,8192,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,5120,7168,0.017091555727852713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,5120,51200,1.6001635657416449
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,12288,12288,0.030709332889980737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,5120,51200,0.6734684308369955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,128,1536,0.009020444419648912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,2560,512,0.003806222230195999
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,51200,16384,3.2848409016927085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,4096,1536,0.030966222286224365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,5120,3072,0.0348328881793552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,512,12288,0.028120001157124836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,128,65536,0.3516791131761339
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,16384,0.10817333062489827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,3584,65536,0.19693422317504883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,256,8192,0.010427555276287926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,128,0.023279110590616863
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,65536,0.01719555589887831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,65536,7168,1.2123457590738933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,128,6144,8192,0.023015111684799194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,51200,512,0.08498577939139472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,32,7168,0.01817599932352702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,128,0.0034924443397257063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,8192,0.01404711107412974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,6144,65536,0.705627547370063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,512,1536,0.014115555418862237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,32,16384,0.01391377713945177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,3584,0.0395066671901279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,256,3072,0.00758844448460473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,10240,128,0.0144177774588267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,5120,128,0.014936000108718872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,64,0.005961777849329843
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,2048,6144,0.00997511131895913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,2048,768,0.003485333174467087
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,10240,0.02400622268517812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,51200,1536,0.2802106539408366
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,256,2560,0.02125777800877889
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,1024,128,0.009993777506881291
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,10240,10240,0.029142220815022785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,2560,3072,0.016341333587964375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,7168,3072,0.06411733229955037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,64,2048,0.0052639999323421055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,768,0.05029155479537117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,768,3584,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,1536,7168,0.02884888980123732
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,16384,768,0.04259377717971802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,512,65536,0.0890071127149794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1024,51200,0.0532186660501692
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,10240,768,0.018030222919252183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,10240,0.06471200121773614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,256,0.012101333174440594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,5120,32,0.005896889087226655
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,2560,32,0.00795022232664956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,5120,0.02958933181232876
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,3584,0.02235200007756551
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,3072,4096,0.04225866662131416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,1536,0.030254221624798242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,5120,256,0.020917332834667627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,10240,3584,0.018225777480337355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,51200,0.06685155630111694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,65536,2048,0.42431823412577313
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,5120,3584,0.009394666386975182
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,1024,512,0.0069297779765393995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,4096,512,0.011250666446155973
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,12288,65536,1.021007113986545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,256,12288,0.0374035570356581
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,1536,0.19318577978346083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,8192,64,0.04654577705595228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16,128,7168,0.009351111120647853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,8192,0.08943377600775824
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,4096,128,0.01479022204875946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,256,768,0.00742133292886946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,32,1024,0.005923555543025334
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,6144,128,0.009925333162148794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,65536,0.1654933293660482
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,6144,51200,0.06467288732528687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,8192,6144,0.03382044368320041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,5120,0.01776977711253696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,7168,6144,0.023019555542204116
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,2048,0.0052355556852287715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,128,2048,0.010320888625250923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,257,12288,3584,0.018177777528762817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,6144,2048,0.07239733139673869
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,4096,5120,0.03204977843496535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,10240,12288,0.3174533314175076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,1024,256,0.03785955574777391
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,7168,7168,0.0373964442147149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,7168,2048,0.007311999797821045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,6144,0.014129777749379476
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,51200,51200,4.0646035936143665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,12288,0.06323999828762479
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,5120,1536,0.005249777601824866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,3072,16384,0.06960178083843656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,10240,10240,0.05686222182379829
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,5120,12288,0.0196142229768965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,65536,2048,0.02294933299223582
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,768,5120,0.01816177699300978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,7168,16384,0.19962222046322295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,10240,512,0.018926221463415358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,3584,768,0.018736888964970905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,768,65536,0.32995377646552193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,10240,256,0.017198221551047433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,1024,6144,0.010285333626800114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,16384,512,0.10982577668295966
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,32,256,0.0021466666625605687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1280,12288,5120,0.07375644312964545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2048,1536,0.015225778023401896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,6144,5120,0.01332800007528729
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,8192,64,0.005605333381228977
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,12288,0.027560000618298847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,51200,0.046000887950261436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,10240,12288,0.2586328983306885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,12288,10240,0.18464355998569063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,3584,3584,0.022330666581789654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,6144,1536,0.014985778265529208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,3584,1024,0.003181333343187968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,7168,8192,0.015831111205948722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,4096,5120,0.011740444435013665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,1024,7168,0.006911110960774952
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1024,8192,0.024586665961477492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,128,64,0.007988444632954067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,128,512,0.006246222390068903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,8192,16384,0.4534248775906033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,7168,1024,0.0058408888677755995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,3584,4096,0.014465777410401238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,13,6144,3584,0.012708444562223224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,65536,3072,0.08559466732872857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,128,3584,0.01551822159025404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,3072,768,0.00349866681628757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,3072,0.007669332954618666
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,6144,10240,0.01754399968518151
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,6144,65536,0.11842844221327041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,8192,0.0699404411845737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,768,16384,0.031331555710898504
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,51200,12288,0.1512826681137085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,1024,3072,0.007308444215191736
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,768,12288,0.018217777212460835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,3072,5120,0.012433778080675336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,51200,256,0.12301777468787299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,6144,65536,0.12986844115787083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,51200,3584,0.31528621249728733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,768,0.03872444563441806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,3072,1024,0.015535111228624979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,12288,5120,0.03286400106218126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2560,12288,0.03158222304450141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,8192,0.04913777775234646
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,7168,65536,0.07481777667999268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,51200,16384,15.481070624457466
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,6144,1536,0.01068800025516086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,10240,3584,0.03028800090154012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,128,10240,0.04500266578462389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,512,1024,0.005780444376998478
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,5120,8192,0.06015288829803467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,16384,10240,0.05230844351980421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,3072,51200,0.07364710834291247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,128,768,0.03765689002143012
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,6144,3584,0.04860888918240865
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,16384,4096,0.017653332816229928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,2560,3072,0.00813422186507119
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,16384,2048,0.13227644231584337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,5120,128,0.005877333382765452
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,3584,0.0711688862906562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,12288,8192,0.027116444375779893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,512,3584,0.020624000165197585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,768,512,0.01495733360449473
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,12288,8192,0.03607644306288825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,3072,0.06171022521124946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,3072,8192,0.030449777841567993
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,10240,0.0377146667904324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,512,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,512,32,0.0017964444640609955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,2560,128,0.0021315556433465746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,51200,16384,0.30700710084703237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,12288,512,0.003858666867017746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,2560,7168,0.011731555892361535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,5120,2048,0.030772444274690416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,5120,5120,0.014775110615624322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,10240,768,0.0068915556702348925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,65536,32,0.06389688783221774
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,768,0.0031911111954185697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,51200,0.7382497787475586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,7168,1536,0.019606222709019978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,32,512,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,2048,2048,0.025263999899228413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,512,1024,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,768,128,0.006286222073766921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,16384,1024,0.009280888570679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,10240,0.004584889031118817
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,4096,32,0.00387999994887246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,256,1536,0.016865778300497267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,1024,3584,0.018014222383499146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,7168,64,0.006234666539563074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,4096,128,0.014797333214018079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,3584,0.00833422193924586
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,12288,512,0.03528533379236857
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,3584,4096,0.00832711077398724
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,32,64,0.005929777605666055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,128,16384,0.08702399995591904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,2048,10240,0.01036177741156684
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,16384,5120,0.15340177218119302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,2560,0.14502755800882974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,1536,6144,0.009153777526484596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,256,4096,0.008369777765538957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,768,1536,0.006609777609507243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,6144,64,0.0021840000732077495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,3584,768,0.03270755542649163
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,2560,512,0.005933333188295364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2560,16384,0.0328693323665195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,16384,2560,0.016344888342751395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,4096,256,0.023940443992614746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,10240,12288,0.04277511106597053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,32,6144,0.004920888692140579
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,65536,256,0.12216088506910537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,64,64,0.0017928888814316855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,12288,4096,0.014108444253603617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,16384,1536,0.008983111215962304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,10240,12288,0.45530133777194554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,5120,128,0.01423733267519209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,128,51200,0.07129333416620891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,3584,3072,0.05697155661053128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,32,0.0037431112594074676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,5120,2048,0.015929778416951496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,1024,2048,0.00906666699382994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,1024,8192,0.0442755553457472
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,7168,512,0.027884443600972492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,16384,3584,512,0.03569688730769687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,5120,65536,1.409145779079861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,16384,0.0750017762184143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,6144,3072,0.007677333222495184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,6144,256,0.014902222487661572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,12288,3072,0.013080000049538083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,65536,128,0.010232888989978367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,32,1536,0.006613333192136552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,4096,65536,0.07427199681599934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,3072,768,0.005604444278611078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,768,0.005948444621430502
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,51200,51200,0.8497733540005155
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,7168,4096,0.11195554998185898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,2048,128,0.004193777839342753
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,16384,16384,1.631166246202257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,1024,5120,0.013608000344700284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,256,7168,0.05024622215165032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,1024,0.031166222360399034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,2048,512,0.02015022271209293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,65536,16384,1.2450035942925346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,7168,3584,0.009752888646390703
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,128,128,0.011742221812407175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,65536,512,0.012644444074895648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1024,1536,0.00850044439236323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,7168,0.6287137667338053
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,512,12288,0.007598222129874759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1536,7168,0.03728977839152018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,256,4096,0.008960888617568547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,10240,0.035818666219711304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,2048,0.04605688982539707
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,6144,2048,0.022405332989162866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,7168,512,0.010456889039940305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,3072,512,0.00555377784702513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1537,51200,1536,0.14514044920603433
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,768,3072,0.010384889112578498
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,1024,7168,0.04085955686039395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,7168,512,0.017305778132544625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,12288,2560,0.0637413329548306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,7168,32,0.017852443787786696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,3584,0.03187822302182516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,51200,16384,1.463380495707194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,7168,51200,0.5823520024617513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3073,3584,2560,0.022983999715911016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,51200,768,0.17279111014472115
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,512,0.0063288890653186375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,1024,32,0.0021395554973019496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,2048,0.006642666541867786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,512,7168,4096,0.015487111277050443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,8192,2560,0.02294755478700002
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,128,1536,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,32,10240,0.005904888941182031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,8192,256,0.007600000335110559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,5120,10240,0.052024000220828585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,6144,2560,0.02977777851952447
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,1536,3584,0.017172444197866652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,65536,4096,0.03918577896224128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,65536,6144,2.4146107567681208
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,8192,65536,0.14483466413285997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,7168,256,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1024,6144,0.010807111031479306
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,4096,3072,0.007666666474607255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3584,10240,0.22076800134446886
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3328,8192,2560,0.07204622030258179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,65536,128,0.056199113527933754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,7168,4096,0.012089778151777057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,4096,768,0.017180444465743173
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,512,65536,0.07458577553431193
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,1536,1024,0.006584888945023219
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,256,512,0.013183111117945777
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,8192,65536,1.1515146891276042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1792,32,65536,0.17731289068857828
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,51200,512,0.03159733282195197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,8192,128,0.03131555517514547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,7168,4096,0.14079911179012722
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,256,6144,0.004709333181381226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,8192,256,0.005629333357016246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8192,12288,6144,0.46407556533813477
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,64,0.008979555633332994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,512,6144,0.006288888967699475
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,5120,12288,0.1583120028177897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,2560,10240,0.01750666730933719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2560,6144,0.010249777800507015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,3072,4096,0.008834666675991481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,32,1536,0.006928000185224745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,16384,256,0.010337777435779572
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,512,768,0.016369778248998854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8,8192,6144,0.017456000049908955
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,16384,16384,0.07208266523149279
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,32,16384,0.011123555401961008
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,4096,0.008054222497675154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,12288,32,0.021976000732845728
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,16384,128,0.014192000031471252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,5120,16384,0.09324444664849176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,65536,4096,0.03887910975350274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,257,4096,7168,0.01891999940077464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,8192,2560,0.04177244504292806
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,4096,12288,0.048379557000266186
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,768,6144,0.01995111174053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,256,1536,0.01314222233162986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2048,256,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,3072,0.20901421705881754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,5120,1024,0.005954666684071223
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,7168,8192,0.018058665924602084
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,5120,4096,0.023974221613672044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,4096,8192,1.1329172982109916
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,10240,10240,0.029149333635965984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,128,768,0.0029306666304667792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,10240,128,0.007283555964628856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,10240,2048,0.026362667481104534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,256,1536,0.015099555253982544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,6144,768,0.02501511077086131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,1536,7168,0.010006222460005019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,256,512,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,7168,1024,0.005588444570700328
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,2048,10240,0.12535377343495688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,2048,51200,0.09315022495057847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,2560,6144,0.03911911116706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,512,5120,0.0096142225795322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,768,8192,0.05453333258628845
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,256,256,0.00867022242810991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,4096,10240,0.017610667480362784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,3584,0.09187289079030354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,3072,256,0.0022808888720141514
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,4096,1024,0.009693333672152625
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,7168,51200,0.3337093459235297
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,10240,5120,0.24117067125108507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,8192,64,0.012992888689041138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,32,3584,0.027775999572541978
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,16384,256,0.009920000202125972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,6144,3584,0.011952000359694162
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,1024,128,0.016162667009565566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,64,10240,0.012549333274364471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,64,3584,7168,0.011753777662913004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,3072,1536,0.01755911111831665
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,10240,256,0.023037332627508376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,65536,4096,0.09505955378214519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,1024,128,0.011017777853541903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,1536,5120,0.01682577696111467
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,768,6144,0.028055111567179363
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,112,2560,65536,0.06291111310323079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,5120,128,0.01367022262679206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,65536,10240,0.298196448220147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,10240,65536,1.4173591401841905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,1024,4096,0.008671999805503422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,128,65536,0.11768533123864068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,768,3584,0.007651555869314406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,5120,3072,0.012771555946932899
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,6144,512,0.02892177634769016
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,16384,0.04174400038189358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,5120,3072,0.049136890305413135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,8192,3584,0.12274577882554795
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,8192,1024,0.02921066681543986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,7168,0.13388533062405056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,2560,1024,0.010751999914646149
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,512,5120,0.015941333439615037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,8192,4096,0.02940444482697381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,16384,512,0.01565511193540361
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,1536,128,0.005892444401979446
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,65536,0.05966222286224365
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,12288,256,0.02199288871553209
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,1536,128,0.0026364444444576898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,3584,1536,0.0069822221994400024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,1024,51200,0.06232711341645983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,4096,5120,0.021898667017618816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,51200,512,0.0976311100853814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,3072,768,0.017687999539905123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,3072,65536,0.07817244529724121
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,513,4096,7168,0.019556444552209642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,5120,65536,0.07208533419503106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,8192,65536,0.12973866197797987
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,2560,12288,0.09778489006890191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,65536,16384,9.984939575195312
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,8192,128,0.005613333235184352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,5120,3072,0.008614222208658854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,2560,4096,0.008031110796663497
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,5120,51200,0.0558755530251397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,4096,512,0.014852444330851236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,2560,1024,0.014764444695578681
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,128,16384,0.027460444304678176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,256,2048,0.016538666354285348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3072,51200,0.1370017793443468
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,65536,768,0.017839110559887357
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,12288,2048,0.011366222467687396
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3584,256,0.01181688904762268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,16384,3072,0.036624001132117376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,512,0.015449777245521545
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,3584,128,0.0031759999692440033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,1024,5120,0.04129244552718268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,3584,8192,0.019313777486483257
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1536,128,0.00589155571328269
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,12288,512,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,1024,768,0.01607733302646213
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,3072,12288,0.017901332841979135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,2560,4096,0.028162668148676556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3072,3072,0.0075048887067370946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,128,6144,0.030608889129426744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,65536,5120,1.0086328718397353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,768,1536,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,51200,3584,0.8496781984965006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,128,7168,0.008962666822804345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,128,2048,0.006281777802440856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,768,512,0.008014221986134848
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,7168,128,0.005628444254398346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,96,1024,1024,0.006238222122192383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,64,2048,0.005250666704442766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,5120,1536,0.26935911178588867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,8192,12288,0.02106222179200914
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,512,1536,0.0035617777870761026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,1024,4096,0.008548444343937768
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,1536,2048,0.009747555686367882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,3072,10240,0.05896355708440145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,1024,2048,0.014135999812020196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,768,7168,0.011047110789351992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,4096,4096,0.07263911432690091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,65,16384,65536,0.23528711001078287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,768,1024,0.013088000317414602
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,512,1536,0.014799111419253878
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,5120,1536,0.04569866591029697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,512,1024,0.00555644432703654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,8192,3584,0.042703999413384326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,16384,64,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,5120,256,0.015170666906568738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,7168,4096,0.017910222212473553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,16384,6144,0.024697777297761705
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2560,10240,16384,0.3556160132090251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,768,1024,0.004127999974621667
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,4096,12288,0.2896168761783176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,512,12288,0.14872711234622532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3584,128,64,0.009302222066455418
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,10240,2560,0.010601777997281818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,2560,0.005935110979610019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,768,768,0.02015022271209293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,6144,128,0.005624889085690181
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,32,1536,0.0035128887328836652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,1536,2560,0.034597333934572004
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,33,512,16384,0.00794400026400884
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4,1024,3072,0.01888799998495314
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,512,16384,0.026898667216300964
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,768,0.01904000010755327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,128,51200,0.03586133321126302
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,65536,1536,0.06800622410244413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,7168,128,0.014893333117167154
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,7168,6144,0.021689777572949726
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,51200,256,0.066849774784512
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,10240,0.005568888866239124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,7168,6144,0.15488266944885254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,2560,512,0.008968888885445064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,12288,12288,0.04801422357559204
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,1536,5120,0.008638222184446123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,64,4096,0.0076044441925154785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,32,16384,0.005260444349712796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,4096,3072,0.056003557311164014
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,5120,0.0310204459561242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,32,7168,0.006280000011126201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,256,3584,0.01923111081123352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,32,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,512,5120,0.009301332963837517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,3072,2560,0.010790222220950656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,128,256,0.002480888946188821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2048,2048,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,51200,1536,0.05653599898020426
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,16384,1024,0.031153778235117596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,5120,16384,0.024304888314670984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,128,512,0.006919999917348226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,2048,65536,0.5027608871459961
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4,51200,51200,0.3499511082967122
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,7168,1024,0.025662221842341956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,7168,1024,0.005942222144868638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,128,32,0.00179822225537565
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,2560,10240,0.07880622148513794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,12288,1536,0.028621332512961492
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,6144,2048,0.030749334229363337
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,51200,0.13945421907636854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,1024,65536,0.064355558819241
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,128,2560,0.015446222490734525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,1024,3584,0.016151999433835346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,768,0.028889778587553237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,7168,128,0.002471999989615546
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,2560,5120,0.008616000413894653
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,16384,12288,0.03396888905101352
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,10240,8192,0.021697777840826247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,5120,2560,0.008930666579140557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,5120,0.11174044344160293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,8192,16384,0.13640444808536106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,256,6144,0.008725333544943068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,64,768,0.00286666676402092
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,768,64,0.0018497777895794974
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,4096,2048,0.023779556155204773
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,256,128,0.0063697778516345555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,6144,7168,0.05753688679801094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,3072,2048,0.006907555378145642
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,32,8192,0.012537777423858643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,1536,6144,0.01053422192732493
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,16384,65536,0.1427262226740519
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,256,4096,0.007617777420414819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,12288,2048,0.02306933369901445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,8192,3584,0.1009262204170227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,8192,65536,1.4652951558430989
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,65536,1536,4.911255306667752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,7168,256,0.018018666240904067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,7168,5120,0.020797333783573575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,3584,16384,0.02126577827665541
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,10240,16384,0.03768622212939792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,6144,3584,0.011402666568756104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,1536,2048,0.009033777647548253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,3584,2048,0.006712889091836081
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,7168,2048,0.022521777285469904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,3072,8192,0.047776887814203896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,2048,3584,0.03268266717592875
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,512,3584,0.015479111009173922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,768,16384,0.010674667027261523
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,768,0.019898666275872123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,256,10240,0.01237777786122428
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,4096,2048,0.005371555685997009
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,5120,768,0.005595555735958948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,10240,10240,0.18027022149827746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,128,3584,0.007664889097213745
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,7168,0.022064000368118286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,2048,7168,0.013788444300492605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,51200,5120,0.21374932924906412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,4096,0.021360889077186584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,1536,65536,0.03708088729116652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,16384,1536,1.0393493440416124
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,128,3584,0.025986666480700176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,7168,7168,0.3138773441314697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,2048,0.13026932875315347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,512,1024,0.006578666468461354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2048,16384,0.07433066765467326
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,512,2560,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,3584,0.04929600159327189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,16384,512,0.004176889028814104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,65536,64,0.012044444680213928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,5120,32,0.008266666697131263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1793,7168,3584,0.049351112710105054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,4096,12288,0.01939022209909227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,224,65536,51200,1.4545582665337458
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,1536,4096,0.0277128881878323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,7168,0.06273333231608073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,128,7168,128,0.002842666581273079
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,7168,16384,0.0524142218960656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,512,1024,0.014848000473446317
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,16384,512,0.0069155556460221606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,51200,10240,1.3661795722113714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,1024,8192,0.02882933285501268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3,3584,64,0.007656888829337225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,16384,32,0.008619555168681675
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,8192,1536,0.06726133161120944
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,5120,2048,0.03641866644223531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,16384,8192,0.08115022050009833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,2560,65536,0.7667182286580404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,51200,6144,0.18778844674428305
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,2560,0.006627555522653792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,2560,768,0.021236444513003033
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,512,1024,0.00629066675901413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,1536,3584,0.02161600026819441
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,128,65536,0.011380444798204633
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,3584,6144,0.012792000340090858
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,1536,0.009113777842786577
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,512,128,0.012709333664841123
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,1536,65536,0.0873875551753574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,51200,64,0.06368800004323323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1024,2560,0.010091555615266165
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,4096,1024,0.0031804444475306403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,512,768,0.015480000111791821
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,12288,3072,0.012073777616024017
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,10240,768,0.018559111489189994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,768,0.006256000035338932
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,5120,8192,0.10352000263002183
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,768,51200,65536,1.0416178173489041
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1025,2560,768,0.008309333688682979
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,6144,5120,0.1645875506930881
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,1024,64,0.033196445968416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,65536,8192,0.08913689189487034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,12288,0.035043554173575506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,1024,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,3072,256,0.011959999799728394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,64,768,0.005943999936183293
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,1536,8192,0.039572444227006696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,32,8192,0.010094222095277574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,6144,0.1330746677186754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,65536,768,0.013809777796268463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,65536,10240,0.08850577804777358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,512,4096,0.021034666233592566
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1536,10240,0.012423999607563019
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,3072,16384,0.140754673216078
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,768,768,0.0030951110853089225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,256,4096,0.027493332823117573
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,768,64,0.0018560000591807896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,5120,64,0.004632888982693355
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,65536,1024,0.2378915680779351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,16384,7168,0.03804533349143134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,3584,2048,0.0069226668112807805
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,512,4096,0.02125955621401469
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,512,6144,0.024728000164031982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,64,32,0.0017928888814316855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,8192,12288,0.019560888409614563
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,14,8192,16384,0.03050755461057027
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,2048,65536,0.0479928884241316
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,10240,7168,0.022281777527597215
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,6144,0.009734222458468543
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,2560,512,0.005560889012283749
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2560,3584,0.027786665492587622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,16384,256,0.017491555876202054
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,1536,5120,0.038346668084462486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,1536,2560,0.007169778148333232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,6144,0.009895111123720804
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3073,7168,128,0.011563555234008364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,512,1024,0.04657333427005344
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,4096,8192,0.05011555552482605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,2048,5120,0.03164711263444688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,3584,5120,0.029466665453381006
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,8192,12288,0.03491733471552531
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,4096,12288,0.018852444158660043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,3072,7168,0.011085333095656501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,16384,51200,0.13179022736019558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,8192,12288,0.034576889541414045
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,2560,4096,0.039177778694364764
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,51200,16384,10.23183356391059
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,768,2560,0.007624000310897827
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,16384,65536,0.3182426558600532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,10240,0.18514488803015816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,12288,7168,0.23417509926689994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,1536,3584,0.007602666815121968
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,6144,2560,0.008598222500748104
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1024,10240,12288,0.11571021874745686
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,512,0.0021520000365045336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1536,2048,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,2560,1536,0.004641777939266629
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,17,64,2048,0.006623999940024481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,8192,128,0.009463999834325578
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2048,4096,0.04824799961513943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,3072,1024,0.015105777316623263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,5120,0.04547022117508782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,1536,7168,0.010037333601050908
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,12288,4096,0.014319111903508505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,2560,6144,0.030945777893066406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,2560,0.023718222975730896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,12288,0.004935111022657818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,1536,0.006237333433495627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,12288,7168,0.031222220924165513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2560,512,0.009970666633711921
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,6144,65536,0.12099022335476345
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,51200,1024,0.01549600064754486
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,768,768,0.005582222094138463
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,128,51200,0.2148897780312432
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,4096,128,0.023366222778956097
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,128,1536,0.0038399998512532976
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,16384,7168,0.43428177303738064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,51200,10240,0.19073955217997232
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,8192,32,0.0029404444826973807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,4096,12288,0.0899511112107171
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,65536,6144,0.13131377432081434
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,8192,65536,0.08167466852400038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,1536,8192,0.13122844696044922
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,16384,3584,0.014294221997261047
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,51200,2048,0.3238568835788303
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,65,6144,7168,0.01995111174053616
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,2048,65536,0.06486666864818998
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,10240,1536,0.008780444661776224
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,3072,6144,0.011191999746693505
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,2560,4096,0.05687466594907972
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,1024,16384,0.4088897705078125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,65536,16384,3.284457736545139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,768,5120,0.014936000108718872
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,4096,12288,0.024280000064108107
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,32,1536,0.009298666483826106
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,8192,0.011720889144473605
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,768,512,3584,0.028037332826190527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,51200,2048,0.026465778549512226
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,7168,2560,0.03892533315552606
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,768,128,0.013467555244763693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,65536,0.7003057797749838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,4096,768,0.014099554883109199
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,7168,7168,0.04028800129890442
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,12288,3584,0.01385244478782018
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,768,3584,0.014180445008807711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,8192,512,0.03698488738801744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,3072,768,0.009396444592210982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,2048,12288,0.027095110880004034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,32,0.003536888708670934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,16384,768,0.02012355625629425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,8192,2560,0.5510293112860786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,256,51200,0.10273155238893296
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,768,7168,0.05764089028040568
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,12,256,7168,0.01853155593077342
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,3584,16384,0.024695111645592585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,8192,1024,0.019945777124828763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,2048,256,0.01202311118443807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,6144,12288,0.019492444064882066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,7168,6144,0.03376266691419814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,2560,64,0.004467555632193883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,6144,128,0.015125332607163323
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,12288,0.08870133426454331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,12288,2.385885238647461
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,16384,1024,0.07123021947013007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,7168,768,0.006231110956933763
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3584,51200,0.32444532712300617
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,128,512,0.014104000396198697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,128,1536,0.014440000057220459
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2817,2048,1536,0.013366222381591797
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,3072,2560,0.03449155555831061
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1,256,1024,0.00282577777074443
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,12288,3584,0.016556445095274184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,64,16384,0.07684266567230225
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,3072,1536,0.01958400011062622
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,16384,1536,0.011376000112957425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,8,16384,6144,0.049069331751929395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,2048,1024,0.024031110935741003
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,2560,12288,0.17238577206929526
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,32,16384,0.07429333527882893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,8192,8192,0.024264888630972967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,1536,1536,0.015102222561836243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,768,16384,0.016480889585283067
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2048,8192,10240,0.14190755950080022
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,10240,4096,0.030842665168974135
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,65536,128,0.011704888608720569
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,2048,12288,0.017941332525677152
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,8192,65536,1.8232623206244574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,14,1024,65536,0.06252621942096286
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,7168,5120,0.02643999954064687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,2560,3584,0.011008000208271874
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,64,256,0.0028204443968004654
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,6144,3584,0.014645333091417948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,2560,8192,0.9581759770711263
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,2560,0.1945635610156589
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,4096,3584,0.011983999775515662
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,8192,512,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,6144,1536,0.026484444737434387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,64,65536,0.01072177787621816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,7168,12288,0.020115555988417733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,3072,768,0.007632888853549957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,12288,1024,0.05050399899482727
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,1024,10240,0.09375644392437404
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1281,4096,2048,0.031160000297758315
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,12288,10240,0.26457156075371635
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,16384,5120,0.30733956231011283
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,10240,128,0.006287999865081575
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,128,2048,0.004530666602982415
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2048,3584,51200,0.22333243158128527
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,512,1024,0.005951111101441913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,6144,7168,0.03815911213556925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,768,16384,0.03475911087459988
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,512,2048,0.013518222504191928
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,128,64,0.001791999985774358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,4096,6144,0.10282578070958455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,512,3072,0.015166223049163818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,12288,7168,0.03702133231692844
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,2560,5120,0.010227555202113258
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,6144,65536,0.07212889194488525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,9,4096,3072,0.025368889172871906
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32768,1536,2560,0.21267378330230713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,51200,12288,0.26826222737630206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,3072,12288,0.05196266704135471
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,5120,8192,0.19911821683247885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,4096,51200,0.9170150756835938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,64,64,0.006817777951558431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,3584,5120,0.04866133464707268
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,64,3072,0.007293333609898885
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,5,65536,1536,0.07182311349444918
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,1024,8192,0.050817777713139854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,2048,0.029496000872717962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1024,3584,0.022513777017593384
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,1024,3584,0.014588443769348992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,7168,12288,0.25977688365512425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,49,10240,2560,0.027894222074084814
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,768,51200,0.053436444865332715
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,6144,7168,0.09978399674097697
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,4096,3072,0.010069333016872406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,12288,768,0.04896622233920627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,224,512,768,0.006136000156402588
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,32,64,0.005600888695981767
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,10240,7168,0.05174933208359612
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,512,768,0.005909333212508096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,512,3584,0.007303111255168915
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1,16384,10240,0.04269599914550781
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,12288,10240,0.12374222278594971
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,51200,64,0.14009777704874674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,3584,1536,0.013630222115251752
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,12288,1024,0.02132533325089349
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8192,64,16384,0.14117422368791369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,7168,64,0.002528000002106031
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,65536,5120,0.12397332986195882
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,64,2560,0.0042364444169733255
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,128,16384,0.02772355576356252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,6144,128,0.014149333039919535
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,129,6144,768,0.005903111149867375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,65536,512,0.05617333120769925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,12288,2048,0.011345778074529437
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,256,8192,0.01923199991385142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,10240,3072,0.026271111435360376
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,51200,64,0.146997332572937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,3584,51200,0.6423884497748481
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,512,65536,0.43619733386569554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,64,16384,0.026322666141721938
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,4096,0.008762666748629676
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,7168,10240,0.08629688951704238
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,5120,10240,0.015306666493415833
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,8192,4096,0.010446222292052375
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,12288,10240,0.02645333276854621
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,1536,1536,0.013062222136391534
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,7168,2048,0.00693866651919153
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,6144,1536,0.006870222174459034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,4096,65536,0.14680000146230063
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,5120,2560,0.008734222087595198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4353,1024,7168,0.08735466665691799
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,8192,8192,0.2425537904103597
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,4096,768,0.00590844452381134
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1793,7168,5120,0.09529866774876912
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,2560,12288,0.01735466718673706
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,10240,768,0.031700445546044245
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1536,51200,0.047684444321526416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,3072,12288,0.21811644236246744
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,65536,768,0.02454400062561035
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,32,12288,0.005203555441564984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,7168,51200,0.1088906659020318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,7168,0.2533573309580485
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,11,6144,2560,0.007660444411966536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,16384,6144,0.25722400347391766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,3584,2560,0.04524177643987867
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,12288,4096,0.044053332673178784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,65,51200,768,0.012533333566453723
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,12288,512,0.01958488921324412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,7168,8192,0.030398223135206435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,10240,5120,0.037054220835367836
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,8192,128,0.015421332584487068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3,16384,512,0.004204444587230682
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,1024,32,0.005910222315125995
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,4096,16384,0.08550222052468194
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2049,10240,12288,0.27398933304680717
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,5120,65536,0.07099466853671603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,512,65536,0.10934311151504517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16,6144,12288,0.035946667194366455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,8192,3584,0.011071110765139261
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,768,0.015124445160230001
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,12288,1024,0.04257066713439094
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,3584,0.013745778136783175
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,65536,8192,0.7646337615119086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,12288,5120,0.07635644409391615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,6144,3072,0.11224267217848037
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2305,1536,256,0.01720177796151903
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,768,512,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,7,51200,10240,0.15966044531928167
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,256,2560,0.028505778974956934
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,768,12288,0.03142755561404758
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,256,51200,512,0.014463111758232117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,12288,5120,1.9332656860351562
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1024,10240,10240,0.17811644077301025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2561,3072,256,0.011277333729796939
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,6144,3072,0.08316622177759807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,33,32,1024,0.008272888759771982
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,512,3072,0.02195377813445197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,128,2560,0.045243554645114474
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,51200,6144,0.11252800623575847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,8192,10240,0.018192888961897958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,10240,128,0.022103110949198406
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,65536,51200,0.6599760055541992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,4096,2048,0.006218666831652324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4353,1024,32,0.004558222161398994
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,8192,512,0.026549332671695288
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,256,768,0.014544889330863953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,2560,1536,0.003839111162556542
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,51200,65536,7.193149142795139
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,65536,256,0.01129155523247189
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,128,3584,0.020942222740915086
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,2560,1536,0.0690551135275099
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,6,2560,1536,0.009662222531106737
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,32,16384,0.007931555310885111
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,3072,512,0.018613333503405254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,128,0.006066666709052191
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,1536,1536,0.015626667274369132
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3073,2048,6144,0.08032000064849854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,6144,6144,0.07455644342634413
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,5120,768,0.03071199854214986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,513,6144,51200,0.2151119973924425
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,256,0.01475822263293796
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,12288,3072,0.013064000341627332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,10,65536,51200,0.5176382064819336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,16384,3072,0.013223999904261695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,16384,3072,0.1698897812101576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,2560,0.0496346652507782
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3072,8192,65536,1.466567145453559
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,10240,3584,0.08851733472612168
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,3584,65536,0.3357697857750787
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,16384,6144,0.05048622356520759
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,3584,2048,0.06663111183378431
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,15,3584,8192,0.012755555411179861
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,65536,3584,0.03508177730772231
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,16384,1024,0.025701334079106648
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,3072,1536,0.004538666870858935
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,65536,6144,0.12044621838463677
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,128,7168,0.03122044603029887
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,12288,1536,0.009152000149091085
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,12288,64,0.0021484444538752237
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,1536,768,0.0035173334181308746
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,16384,512,0.02937955657641093
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,7168,2560,0.04059111078580221
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,3072,3584,0.015803555647532146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,4096,32,0.00797333319981893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,64,2048,0.004521777646409141
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,7168,256,0.014484445254007975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,65536,2048,0.10036444664001465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,51200,51200,5.11769061618381
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,65536,2560,0.5313982433742946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,10240,7168,0.2670346630944146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1024,8192,10240,0.12359289328257243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16,7168,6144,0.020451555649439495
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,12288,51200,0.2883217864566379
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,12288,12288,0.6104506916469997
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,257,16384,2560,0.023046221998002794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,512,12288,0.014974221587181091
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1024,65536,51200,1.4581670761108398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,3072,256,0.005223111146026188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,51200,10240,0.9903501934475369
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3585,512,3584,0.029782222376929388
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,49,1024,1024,0.005609777652555042
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,160,4096,51200,0.057518223921457924
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,5,3584,768,0.005744888964626525
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,128,6144,0.04520710971620348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,81,1536,128,0.014155555102560254
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,1536,2560,0.030491554074817236
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,4096,0.05928533607059055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,256,51200,3584,0.09252178006702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,14,65536,51200,0.6075466473897299
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,1536,5120,0.022678222921159532
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,11,256,5120,0.00869422240389718
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1792,3072,2048,0.017800889081425138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,8192,2560,0.047648890150917896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,10240,7168,0.020054222808943856
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,6144,128,0.026825777358478967
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,256,768,0.005608888963858287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,2560,3072,0.007595555649863348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,16384,51200,16.101050482855904
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4097,8192,512,0.02937777837117513
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,6144,1536,0.04277777671813965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,1024,8192,0.021963554951879714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,51200,3.8677084181043835
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,8192,512,0.005927111125654644
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,65536,10240,2.409672843085395
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,1024,65536,0.06142933501137627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,3072,256,0.03885511226124234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4352,1024,64,0.004896000027656555
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,3072,8192,0.8709884219699435
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,16384,16384,0.9593093660142687
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1537,16384,2560,0.10770222213533188
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,3072,3072,0.011022222538789114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,5120,256,0.010046222143703038
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,32,1536,0.005882666756709416
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,512,768,0.006695111178689533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,768,3072,2048,0.018811555372344125
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,1024,1024,0.018957333432303537
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,81,128,16384,0.018578666779730055
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,96,7168,16384,0.06343733602099948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,2048,2048,0.023481777972645227
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,1536,2048,0.01570222278436025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,64,64,32,0.006265777680608962
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3072,128,128,0.002507555608948072
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,12288,4096,0.14961866537729898
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,16384,3072,0.023248000277413264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,48,8192,65536,0.17889066537221274
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,51200,3072,0.055600000752343066
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,3584,0.04295911060439216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3,2048,512,0.014168888330459595
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,4096,7168,0.03640710976388719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,224,4096,2048,0.007302222152551015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,512,6144,0.008992888861232335
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,8192,3584,0.012758221891191272
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,1536,2048,0.01333688861793942
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,12288,6144,0.018630221486091614
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,192,512,128,0.0053324442770746015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,6144,3584,0.009474666582213508
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,8192,1536,0.028027554353078205
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,193,10240,5120,0.016536888149049547
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,81,65536,3584,0.043262223402659096
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3585,3072,5120,0.08656533559163411
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,128,65536,0.0506897767384847
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,2560,8192,0.015517334143320719
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,3584,256,0.0021848889688650766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2,1536,51200,0.04791288905673557
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3585,6144,8192,0.0874168872833252
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,2048,6144,0.009008888569143083
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,80,65536,1024,0.015795555379655626
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,5120,16384,0.04147555430730184
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4096,6144,1536,0.04362488786379496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,512,128,0.007937777373525832
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,6144,2048,0.007615110940403408
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,160,12288,6144,0.042486220598220825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,8192,51200,0.07684977849324544
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4097,12288,65536,1.616596433851454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,256,256,0.008310221963458592
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,768,1536,0.006724444528420766
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2048,512,256,0.008336889247099558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,15,128,128,0.0018444444156355327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4352,7168,4096,0.17889422840542266
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,13,768,768,0.015840000576443143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1,10240,3072,0.0477111107773251
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1536,64,256,0.003539555602603488
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,5,65536,768,0.023207111491097346
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,7168,2560,0.099781334400177
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,16384,4096,4096,0.3189200030432807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,7168,4096,0.018245332770877414
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,256,16384,6144,0.020385776956876118
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,3584,2048,0.015899555550681222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,7168,12288,0.019152000546455383
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,65536,10240,0.9686880111694336
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,8192,8192,0.045936889118618436
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,5120,2560,0.013570666313171387
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,6,3072,5120,0.00926577796538671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,512,256,0.002199999988079071
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,192,51200,1536,0.036228444841172956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,256,65536,0.10819822549819946
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32768,65536,128,1.2284960216946073
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1281,768,3584,0.011714666253990598
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3329,6144,3072,0.08266133069992065
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32,2048,512,0.005544888890451855
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,113,8192,8192,0.019526221685939364
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,1024,10240,0.026279111703236897
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,3584,51200,0.08142044146855672
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,768,5120,0.022220444348123338
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2048,12288,1024,0.03865955604447259
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,12,3072,6144,0.009359999663299983
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,113,51200,3072,0.02590844366285536
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,1536,5120,0.02606933315594991
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3841,128,51200,0.12807733482784694
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,112,4096,1024,0.006243555496136348
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,7,512,51200,0.043476445807351
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,225,1024,10240,0.01643288963370853
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3328,65536,16384,3.126103083292643
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3073,8192,512,0.036064889695909284
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,10240,12288,0.07294311126073201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,65536,8192,0.5438533359103733
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,49,3584,65536,0.07763022184371948
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,2048,768,0.0062871111763848206
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,256,6144,0.009711999860074785
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,2560,65536,0.5146204630533854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,513,1024,65536,0.13155999448564318
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,1024,512,0.01073155552148819
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,2560,32,0.0058755555914507965
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,768,7168,4096,0.04443733228577507
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2049,512,65536,0.11965333090888129
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3840,7168,16384,0.5776951048109267
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,97,2048,10240,0.03647733396954007
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,17,16384,6144,0.03567644291453891
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,5120,0.017832888497246634
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,768,1536,0.008273777862389883
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,193,256,128,0.002528888897763358
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,9,1024,2048,0.008667555948098501
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,2048,768,0.003896000070704354
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3840,3584,768,0.02442933287885454
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,3072,2048,0.007954667011896769
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,80,3072,3072,0.009024000002278222
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,512,64,0.009279111193286048
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2,10240,3072,0.013374221821626028
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4,65536,2560,0.03461866577466329
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,64,5120,0.004563555535342959
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2,12288,7168,0.032687111033333674
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,1536,6144,0.03432000014517043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,16384,7168,0.05311377843221029
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,7168,1536,0.020630222227838304
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,512,768,5120,0.02196888956758711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,8192,256,0.016178665889634024
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1280,2560,4096,0.03289777702755398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4352,1024,12288,0.09437777598698933
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,128,7168,1536,0.018918222851223417
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,769,6144,768,0.00850311087237464
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,80,3072,256,0.014800889624489678
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16,512,32,0.005255110975768831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,256,1536,0.004960000101062986
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1537,8192,512,0.01277955538696713
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,112,2048,16384,0.033013333876927696
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3329,6144,3072,0.08870133426454331
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,11,32,6144,0.004578666554556953
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,6144,3584,0.02223199937078688
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32,12288,6144,0.029643555482228596
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,3584,3072,0.031097776359981958
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,256,1536,16384,0.022150221798155043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,225,256,16384,0.03241511185963949
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,1024,256,0.011069333387745751
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2816,7168,8192,0.14526577790578207
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1536,256,256,0.006999111009968652
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1025,1536,8192,0.02334044377009074
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,6144,2560,0.021664889322386846
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2817,1024,3072,0.03682311044798957
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,8,1536,32,0.005558222118351195
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,129,4096,10240,0.018170666363504197
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,3584,1536,0.01808088852299584
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,512,3072,0.00962577760219574
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,32768,8192,16384,6.393745846218533
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,161,1536,32,0.009030222064918941
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,3072,65536,0.5140808953179253
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2560,7168,256,0.011872000164455838
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,128,0.02090222140153249
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2305,1536,6144,0.04311466548177931
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,2560,32,0.006575999988449945
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3840,12288,16384,0.3015146785312229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3584,8192,256,0.02075111038155026
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,64,12288,0.013440888788965015
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,1024,3584,0.0058746664888328975
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1025,5120,5120,0.0466231107711792
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,768,16384,0.01997422178586324
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,1536,768,0.0031831111345026228
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,768,512,0.005252444495757421
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,512,16384,65536,0.46988619698418516
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,33,6144,3584,0.008473777936564552
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,6144,8192,0.017663111289342243
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,192,51200,51200,0.7239502271016439
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,160,12288,51200,0.22527734438578287
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4,51200,6144,0.09814755784140693
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1793,5120,51200,0.6822515593634711
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,7,512,256,0.0059075554211934405
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,1536,8192,0.0483822226524353
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,256,512,0.0031928889867332247
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1025,512,65536,0.06164533562130398
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,128,7168,0.03478044602606032
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,512,10240,0.012735111018021902
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1792,51200,512,0.04801333281728956
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,257,7168,1536,0.02174844510025448
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2561,65536,4096,1.0160462061564128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1793,4096,8192,0.04171999957826403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,17,51200,128,0.018191999859280057
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,2048,256,0.0052897776994440295
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,12,2048,10240,0.015573332707087198
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8192,5120,16384,0.3038960032992893
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,256,7168,0.005603555589914322
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,7168,1024,0.03045511245727539
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,161,5120,16384,0.04663999875386556
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,16384,512,0.007817777494589487
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3072,1536,0.006607999818192587
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,6,768,65536,0.07565777831607394
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,3072,51200,0.051424887445237905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1792,16384,51200,1.1402533849080403
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,51200,8192,0.7557431326972114
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,6144,6144,0.03715199894375271
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4352,12288,7168,0.15731199582417807
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,48,12288,768,0.018360889620251127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,193,1536,3584,0.008664888640244802
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,225,16384,65536,0.15070577462514242
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,16384,3584,0.014140443669425117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,113,8192,3072,0.024319999747806128
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,1024,51200,0.15946577654944524
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,7168,4096,0.11436711417304145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,768,2048,0.004278222305907143
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,5,2560,2560,0.007287110719415877
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,12288,12288,0.04939377639028761
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2305,512,2048,0.006632888896597757
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,768,768,0.0028657778683635923
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2817,3072,64,0.005247999810510212
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1536,5120,5120,0.04235022266705831
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1281,3584,1536,0.026752889156341553
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,9,51200,1536,0.01628888812330034
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,10240,16384,0.10041955444547866
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,13,32,16384,0.0052151112920708126
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,97,10240,64,0.0027715555495686005
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,1536,3072,0.008205333517657386
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2304,2048,768,0.012788444757461548
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2560,10240,65536,0.7979626655578613
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3328,12288,65536,4.093749152289496
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3841,3584,2048,0.04682844546106127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,97,12288,512,0.006913777854707506
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2561,7168,7168,0.1939368910259671
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4097,128,3072,0.022649778260125056
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,96,12288,10240,0.026603556341595117
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2049,3584,2560,0.026643556025293138
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,8192,12288,0.06538489129808214
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2816,128,2048,0.010401777923107147
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,81,16384,256,0.0035217776894569397
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,4353,5120,3072,0.06944444444444445
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,11,3072,1536,0.016862221890025668
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,64,6144,4096,0.011583111352390714
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,10,5120,51200,0.05598666932847765
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1,10240,6144,0.022987556126382615
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,12288,2560,0.1320319970448812
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,160,256,128,0.0059164443777667145
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,9,768,1536,0.004465777840879229
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,256,6144,0.03398311138153076
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,113,64,6144,0.005924444645643234
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,112,1536,64,0.005920889063013925
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2049,768,3584,0.014781332678265043
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,161,65536,4096,0.10548088947931926
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1281,16384,7168,0.07631288634406196
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,2048,10240,0.012413333687517377
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4097,256,65536,0.5010186831156412
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,17,4096,7168,0.011706666813956367
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3585,8192,512,0.02683111031850179
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2,7168,3584,0.03050666716363695
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,10,65536,2048,0.04087466663784451
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,6,3072,12288,0.0166284441947937
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3072,7168,51200,0.5294497807820638
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,15,64,128,0.005543111099137201
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,512,16384,128,0.005919999960396025
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,256,51200,0.04928177926275465
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,3072,1024,0.013776889277829064
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,12,64,128,0.0018355554590622585
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,1536,1536,10240,0.05423466695679558
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,12288,64,0.019941333267423842
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,1280,10240,256,0.010088889135254754
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,4096,256,768,0.02068622244728936
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2305,2048,768,0.010955555571450127
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3841,2560,3072,0.05846755372153389
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,32,1024,5120,0.01720355616675483
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,193,512,768,0.014821334017647637
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2817,32,64,0.006574222197135289
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,48,10240,1536,0.007905777957704332
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2304,2048,2048,0.01475644442770216
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,65536,8192,0.7273413340250651
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,1537,8192,64,0.015114666687117683
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,13,51200,6144,0.045528001255459256
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,65,256,5120,0.008970666262838576
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,15,256,4096,0.014843554960356818
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3329,256,12288,0.05355466736687554
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,769,5120,3584,0.04776888754632738
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,10,2560,3584,0.03046222196684943
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4353,10240,512,0.02715911136733161
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,32,5120,51200,0.0592177775171068
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,129,768,3072,0.016146666473812528
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,513,6144,32,0.003170666595300039
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3329,3072,2048,0.01958666741847992
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,14,12288,51200,0.20692977640363905
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3584,6144,0.018014222383499146
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,64,16384,5120,0.04782133301099142
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2304,1536,16384,0.05348266826735603
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,96,7168,16384,0.04804355568355984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,1280,16384,1536,0.022780444886949327
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,49,10240,7168,0.022688888841205176
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,3072,768,7168,0.040058665805392794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,16384,64,128,0.002855111120475663
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,128,1024,5120,0.008707555631796518
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,225,3072,2560,0.00665244460105896
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,161,1536,768,0.005940444353553984
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,32768,3584,12288,0.6616800096299913
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,5120,65536,1.5158221986558702
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,4096,64,3584,0.010053333308961656
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,2816,8192,10240,0.3186302185058594
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,769,6144,5120,0.03195377853181627
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,6144,2048,0.13922488689422607
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,129,256,4096,0.006620444357395172
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,8192,10240,3072,0.32766932911343044
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,192,3072,65536,0.07944711049397786
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,65536,768,0.03321511215633816
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,80,8192,64,0.0022008888837363985
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3841,2560,32,0.005881777654091517
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,48,1024,10240,0.013942221800486246
SGLang,0.5.10,NVIDIA B200,gemm,sglang,float16,3584,6144,768,0.02462844385041131
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2816,2560,6144,0.07168177763621013
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,2304,4096,10240,0.112225784195794
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,2560,10240,12288,0.2675226794348823
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,33,16384,12288,0.060156444708506264
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,4096,4096,1536,0.021019554800457422
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,97,3584,768,0.006229333165619109
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3,16384,2048,0.013046222428480784
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,769,12288,2048,0.019271110494931538
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8,16384,256,51200,0.8046959771050347
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,2561,7168,51200,0.4898844295077854
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,8,7168,12288,0.018998222218619455
SGLang,0.5.10,NVIDIA B200,gemm,sglang,fp8_block,3840,8192,768,0.04194400045606825
SGLang,0.5.10,NVIDIA B200,gemm,sglang,nvfp4,3328,51200,512,0.08249511321385701
